Human Factors in Incident Investigation and Analysis

Human Factors in Incident Investigation and Analysis
WORKING GROUP 1
2ND INTERNATIONAL WORKSHOP ON HUMAN FACTORS IN OFFSHORE OPERATIONS (HFW2002)
HUMAN FACTORS IN INCIDENT INVESTIGATION AND ANALYSIS
Dr. Anita M. Rothblum
U.S. Coast Guard Research & Development Center
Groton, CT 06340
Capt. David Wheal and Mr. Stuart Withington

Marine Accident Investigation Branch
U.K. Department for Transport
Southampton SO15 2DZ
Dr. Scott A. Shappell

FAA Civil Aeromedical Institute
Oklahoma City, OK 73125
Dr. Douglas A. Wiegmann

University of Illinois at Urbana-Champaign
Institute of Aviation
Savoy, IL 61874
Mr. William Boehm

Stolt-Nielsen Transportation Group
Houston, TX 77015
Dr. Marc Chaderjian

California State Lands Commission
Marine Facilities Division
Long Beach, CA 90802
April 8 10, 2002

Houston, Texas
Working Group 1 - HFW2002
Houston, Texas
Form Approved
OMB No. 0704-0188
Report Documentation Page
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it
does not display a currently valid OMB control number.
1. REPORT DATE
3. DATES COVERED
2. REPORT TYPE
APR 2002
00-04-2002 to 00-04-2002
4. TITLE AND SUBTITLE
5a. CONTRACT NUMBER
5b. GRANT NUMBER

5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S)
5d. PROJECT NUMBER

5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
U.S. Coast guard Research & Development Center,Groton,CT,06340

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
8. PERFORMING ORGANIZATION
REPORT NUMBER
10. SPONSOR/MONITORS ACRONYM(S)
11. SPONSOR/MONITORS REPORT
NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release; distribution unlimited

13. SUPPLEMENTARY NOTES
The original document contains color images.

14. ABSTRACT
15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF:
17. LIMITATION OF
ABSTRACT
a. REPORT
b. ABSTRACT
c. THIS PAGE
unclassified
unclassified
unclassified
18. NUMBER
OF PAGES
19a. NAME OF
RESPONSIBLE PERSON
149
Standard Form 298 (Rev. 8-98)

Prescribed by ANSI Std Z39-18
ABSTRACT
Studies of offshore and maritime incidents (accidents and near-misses) show that 80% or more
involve human error. By investigating incidents, we can identify safety problems and take
corrective actions to prevent future such events. While many offshore and maritime companies
have incident investigation programs in place, most fall short in identifying and dealing with
human errors. This paper discusses how to incorporate human factors into an incident
investigation program. Topics include data collection and analysis and how to determine the
types of safety interventions appropriate to safeguard against the identified risks. Examples are
provided from three organizations that have established their own human factors investigation
programs.
DISCLAIMER
This document is disseminated under the sponsorship of the Department of Transportation in
the interest of information exchange. The United States Government assumes no liability for its
contents or use thereof. The United States Government does not endorse products or
manufacturers. Trade or manufacturers names appear herein solely because they are
considered essential to the object of this report. This report does not constitute a standard,
specification, or regulation.
ACKNOWLEDGEMENTS
The authors wish to thank their organizations for granting the permission both to contribute their
time to the development of this paper and to share their organizations data and lessons learned
so that others may profit from our experiences. Thanks to Rodger Holdsworth and Christy
Franklyn of Risk, Reliability and Safety Engineering for providing guidance and reviews of this
paper and for supplying information, references, and a figure used herein. Thanks to Gerry
Miller (G.E. Miller and Associates) for the clarifying discussion of his Triangle of Effectiveness,
help with locating references, and for his editing suggestions. Rachael Gordon contributed
material on error detection and recovery, as well as information about her Human Factors
Investigation Tool (HFIT). Jan Erik Vinnem shared a presentation on Norways offshore nearmiss database. This paper benefited from lively discussions at the April workshop and from
amplifying information contributed by Scott Bergeron (McDermott), Melissa Roberts
(ChevronTexaco), Sally Caruana (BOMEL Ltd.), and Cyril Arney (ExxonMobil). Thanks also to
the following people who kindly took the time to critique a draft of this document: Capt. Dan
Maurino (International Civil Aviation Organization), Trond S. Eskedal (Norwegian Petroleum
Directorate), Jeffery Thomas (ExxonMobil), and James Reason (University of Manchester).
Page i
Houston, Texas
TABLE OF CONTENTS
PAGE
1.0
2.0
3.0
INTRODUCTION ....................................................................................................................1
1.1
Why Study Incidents? ............................................................................................1
1.2
Background from the 1996 International Workshop on Human Factors in

Offshore Operations .............................................................................................2
1.3
The Typical Offshore Incident System and How Human Factors Data
Can
Enrich It ................................................................................................................2
HUMAN ERROR ...................................................................................................................6

2.1
Types of Human Error ...........................................................................................7
2.2
The Maritime System: People, Technology, Environment, and

Organizational
Factors .................................................................................................................8
2.3
Human Factors Issues in the Maritime Industry ..................................................13
2.3.1 Fatigue .....................................................................................................13

2.3.2 Inadequate Communications ...................................................................14
2.3.3 Inadequate General Technical Knowledge ..............................................14
2.3.4 Inadequate Knowledge of Own Ship Systems .........................................14
2.3.5 Poor Design of Automation......................................................................14
2.3.6 Decisions Based on Inadequate Information ...........................................14
2.3.7 Poor Judgement.......................................................................................15
2.3.8 Faulty Standards, Policies, or Practices ..................................................15
2.3.9 Poor Maintenance....................................................................................15
2.3.10 Hazardous Natural Environment..............................................................15
BUILDING A HUMAN FACTORS INCIDENT INVESTIGATION PROGRAM ....................................15
3.1
Keys to a Successful Human Factors Incident Investigation Program ................16

3.1.1
3.1.2
3.2
An Open, Fair, Improvement-Seeking Culture.........................................17

Common Understanding of the Purpose and Scope of the
Incident
Investigation Program ..............................................................................17
3.1.3 Appropriate Training for Incident Investigators ........................................18
3.1.4 Incident Database Classification Scheme ...............................................19
3.1.5 Simple Data Entry ....................................................................................20
3.1.6 Feedback on Results of the Incident Investigation Program ...................20
Investigating for Human Factors Causes ............................................................20
3.3
How an Incident Evolves .....................................................................................21
3.4
Improving Safety through Incident Investigation and Analysis ............................23
Page ii
Houston, Texas
3.5
4.0
5.0
6.0
Error Recovery.....................................................................................................25
AN EXAMPLE OF A HUMAN FACTORS TAXONOMY: HFACS................................................26

4.1
HFACS: Unsafe Acts ..........................................................................................28
4.2
4.1.1 Decision Errors ........................................................................................28

4.1.2 Skill-based Errors ....................................................................................29
4.1.3 Perceptual Errors .....................................................................................30
4.1.4 Routine Violations....................................................................................31
4.1.5 Exceptional Violations..............................................................................31
HFACS: Preconditions for Unsafe Acts ..............................................................31
4.3
4.2.1 Adverse Mental States.............................................................................32

4.2.2 Adverse Physiological States ..................................................................33
4.2.3 Physical/Mental Limitations .....................................................................33
4.2.4 Crew Resource Mismanagement.............................................................34
4.2.5 Personal Readiness.................................................................................34
HFACS: Unsafe Supervision ..............................................................................35
4.4
4.3.1 Inadequate Supervision ...........................................................................35

4.3.2 Planned Inappropriate Operations ...........................................................35
4.3.3 Failure to Correct a Known Problem........................................................37
4.3.4 Supervisory Violations .............................................................................37
HFACS: Organizational Influences .....................................................................37
4.5
4.4.1 Resource Management............................................................................38

4.4.2 Organizational Climate.............................................................................39
4.4.3 Organizational Process............................................................................39
The Benefits of Using HFACS .............................................................................40
PUTTING IT ALL TOGETHER: INVESTIGATING AN INCIDENT FOR HUMAN FACTORS CAUSES .41
5.1
Introduction ..........................................................................................................41
5.2
Incident Investigation: Going Beyond the Obvious .............................................41
5.3
A Tool for Investigation: Events and Causal Factors Charting ...........................42
5.4
5.3.1 TORREY CANYON Synopsis ..................................................................43

5.3.2 An Event and Causal Factors Chart of the TORREY CANYON Incident 44
The Art of Incident Investigation ..........................................................................49
ANALYZING INCIDENT DATA ...............................................................................................52

6.1
Learning from Individual Incidents .......................................................................52
6.2
Identifying High-Risk Activities or Facilities Using Simple Frequency Analysis...52
6.3
Looking for Trends...............................................................................................53
Page iii
Houston, Texas
6.4
Looking for Similar Incidents ...............................................................................55
6.5
Determining Under What Conditions a Given Error Happens .............................55
6.6
Looking for Underlying Causes Using Meta-Analysis..........................................56
6.7
Identifying Relationships Among Incident Attributes by Statistical Analysis........57
6.8
After the Analysis.................................................................................................58

6.8.1
6.9
7.0
8.0
Data-Driven Research: the Link Between Data Analysis and Solving

Safety Problems.......................................................................................59
6.8.2 It May Be Data, But Its Not Necessarily Telling You Anything ................59
Summary of Analysis Techniques .......................................................................60
FINDING SAFETY SOLUTIONS .............................................................................................61

7.1
The Triangle of Effectiveness: A Guide to Safety Interventions .........................61
7.2
Relationship Between Reasons Slices of Cheese and the Triangle of

Effectiveness......................................................................................................64
7.2.1 Policies & Culture ....................................................................................65

7.2.2 Workplace Design....................................................................................67
7.2.3 Environmental Control .............................................................................67
7.2.4 Personnel Selection.................................................................................68
7.2.5 Training and Standard Operating Procedures (SOPs) ............................69
7.2.6 Interpersonal Relationships (Communication).........................................70
7.2.7 Job Aids ...................................................................................................71
7.2.8 Fitness for Duty........................................................................................71
USER EXPERIENCES WITH STARTING AN INCIDENT INVESTIGATION PROGRAM .....................72
8.1
U.K. Marine Accident Investigation Branch .........................................................72
8.2
8.1.1 Background..............................................................................................72
8.1.2 Events Leading to Human Factors Investigations ...................................73
8.1.3 MAIBs Classification of Human Factors Causes ....................................73
8.1.4 Human Factors Training for MAIB Inspectors..........................................74
8.1.5 Benefits of MAIBs Database ...................................................................75
California State Lands Commission, Marine Facilities Division...........................75
8.3
8.2.1 Background..............................................................................................75
8.2.2 Initiation of Human Factors Investigations ...............................................76
8.2.3 Adaptations to HFACS.............................................................................76
8.2.4 Training and Job Aids ..............................................................................77
8.2.5 Benefits of Human Factors Investigations ...............................................77
Stolt-Nielsen Transportation Group .....................................................................78
8.3.1
Background..............................................................................................78
Page iv
Houston, Texas
9.0
8.3.2 Human Factors Incident Investigation......................................................79

8.3.3 Benefits of the Incident Database............................................................79
SUMMARY .........................................................................................................................82
10.0
REFERENCES ....................................................................................................................84
Appendix A
Sample Incident Reporting Form ...................................................................90
Appendix B
Database Items on General Incident Information ..........................................93
Appendix C
Examples of Human Factors Taxonomies...................................................101
Appendix D
Human Factors Taxonomy Used by the UK Marine Accident

Investigation Branch ....................................................................................103
Appendix E
Adaptation of HFACS by the California Sate Lands Commission,

Marine Facilities Division .............................................................................105
Appendix F
Stolt-Nielsen Accident/Incident Investigation...............................................113
Appendix G
Sample Taxonomies for Investigation for Fatigue, Communications

Problems, and Inadequate Skills and Knowledge .......................................115
Appendix H
Human Factors Investigation Questions......................................................134
Appendix I
Norwegian Petroleum Directorate Near-Miss System .................................139
Page v
Houston, Texas
TABLE OF FIGURES
PAGE
Figure 1.
Incident Investigation Process...............................................................................5
Figure 2.
The Maritime System Is A People System ............................................................9
Figure 3.
The Maritime System: People ............................................................................10
Figure 4.
The Maritime System: Effect of Technology on People......................................11
Figure 5.
The Maritime System: Effect of Environment on People ....................................12
Figure 6.
The Maritime System: Effect of Organization on People....................................12
Figure 7.
An Accident in the Making ...................................................................................22
Figure 8.
How An Effective Human Factors Incident Investigation Program Can Improve

Safety...................................................................................................................24
Figure 9.
The Human Factors Classification and Analysis System (HFACS).....................27
Figure 10.
Classification of Unsafe Acts ...............................................................................29
Figure 11.
Classification of Preconditions for Unsafe Acts...................................................32
Figure 12.
Categories of Unsafe Supervision. ......................................................................36
Figure 13.
Organizational Factors Which Contribute to Incidents ........................................38
Figure 14.
Events and Causal Factors Chart for TORREY CANYON Grounding ................45
Figure 15.
The Incident Investigation Process......................................................................51
Figure 16.
Example of a Frequency Analysis .......................................................................52
Figure 17.
Example of a Trend Analysis ...............................................................................54
Figure 18.
Result of Statistical Analysis of Incident Database Items....................................58
Figure 19.
The Triangle of Effectiveness for Safety Interventions.....................................63
Figure 20.
Types of Interventions Needed to Plug the Holes in the Layers of Defenses. ..66
Figure 21.
Comparison of Lost Time Injury Rates ................................................................80
Figure 22.
Number of Incidents as a Function of Body Part Injured. ....................................81
Figure 23.
Tools for an Effective Incident Investigation Program .........................................82
Page vi
Houston, Texas
HUMAN FACTORS IN INCIDENT INVESTIGATION AND ANALYSIS
1.0
INTRODUCTION
Traditionally, incident investigation has focused on hardware issues, such as material failures
and equipment malfunctions. In the last fifteen years or so, it has become increasingly evident
that human factors, rather than hardware factors, are responsible for most of the precursors to
incidents. While many offshore and maritime companies have incident investigation programs
in place, most consider human contributions to incidents only in a superficial way, if at all. The
purpose of this paper is to help offshore and maritime companies incorporate human factors
into their incident investigation programs so that they can identify human causes of incidents
and determine effective safety interventions to prevent such incidents in the future.
1.1
Why Study Incidents?
An accident is defined as an unplanned event or sequence of events that results in

undesirable consequences (Center for Chemical Process Safety, 1992, p.327). Accidents
represent the proverbial tip of the iceberg. It has been estimated that for every accident, there
are about 600 near-misses1 (Det Norske Veritas, 1995; Ferguson & Landsburg, 1998; Bea,
Holdsworth, & Smith, 1997). Essentially, a near-miss is an accident that almost happened.
Near-misses and accidents have the same causes, so studying near-misses can help us
understand safety problems and make corrective changes before an accident takes place. In
addition, since near-misses do not result in full-blown casualties, studying near-misses can help
us learn how to develop early-warning systems to detect when conditions have become nonnormal and also show us what steps were taken that avoided the accident.
Incident2 investigation and analysis that is, the study of accidents and near-misses is
squarely in line with the intent of the International Safety Management (ISM) Code. ISM
requires that a company provide for a safe work environment and safe practices in maritime
operations and establish safeguards against all identified risks. Incident investigation helps the
company to identify its risks and to understand the underlying causes of incidents. This in turn
helps the company develop safe work practices.
A near-miss is defined as an extraordinary event that could reasonably have resulted in a negative
consequence under slightly different circumstances, but actually did not (Center for Chemical Process
Safety, 1992, p. 329).
2
An incident is defined as including all accidents and all near-miss events that did or could cause
injury, or loss of or damage to property or the environment" (Center for Chemical Process Safety, 1992,
p. 1).
Page 1 of 141
Houston, Texas
This paper will help you to learn about human error and how it contributes to virtually every
incident. This paper will also show you how to establish a human factors incident investigation
program in your company and how to analyze the data collected so that you can learn from
incidents and identify how to improve your policies and work practices to achieve a higher level
of safety. Examples are provided from three different organizations that have established their
own human factors investigation programs, sharing lessons learned and how the program has
benefited them.
1.2
Background from the 1996 International Workshop on Human Factors in Offshore

Operations
Studies of accidents and other incidents on offshore platforms have indicated that the vast
majority of these accidents involve human error. In fact, about 80% do, and a further 80% of
these occur during operations (Bea, Holdsworth, & Smith, 1997). The need to understand and
control these human errors led to the assemblage of the 1996 International Workshop on
Human Factors in Offshore Operations. The 1996 Workshop took a broad look at how Human
Factors often called Human and Organizational Factors to underscore the fact that most of
these errors occur not within the span of control of the frontline operator, but are caused instead
by decisions, policies, and operating procedures handed down by higher levels of the
organization affect every aspect of the offshore industry, from the design and fabrication of
offshore production facilities, to field operations and maintenance, to management systems for
improving safety and productivity. That Workshop laid the groundwork for the current
workshop, which is delving into more detail on a number of human factors issues, including
incident investigation and analysis.
The 1996 Workshop provided some good background material on human error (Bea,
Holdsworth, & Smith, 1997; Card, 1997; Wenk, 1997), and even provided some tools that can
be employed for incident investigation (Bea, 1997; Howard, et al., 1997; Kirwan, 1997; Moore,
et al., 1997; Scient, Gordon, et al., 1997); these papers are heartily commended to the
interested reader. The present paper goes into more detail on these topics and focuses the
discussions on: the understanding of how human errors arise and contribute to incidents; a
specific set of tools for representing the events and causes of an incident; dissecting out the
different levels of human error; analyzing incident data; and using the human error model to
select the most effective safety interventions. In short, this paper attempts to provide the reader
with a soup to nuts examination of how to build a successful human factors incident
investigation program.
1.3
The Typical Offshore Incident System and How Human Factors Data Can Enrich It
Thanks, in part, to OSHA and EPA regulations on Process Safety and Risk Management (e.g.,
29 CFR 1910.119 and 40 CFR Part 68) and to the International Safety Management Code,
many offshore and maritime companies already have an incident investigation program in place.
These programs often follow well-grounded investigative practices, providing investigation team
members with training in the basics of incident investigation, gathering and documenting
evidence, and interviewing techniques. Many of these companies also keep an incident
Page 2 of 141
Houston, Texas
database and may do frequency and trending analysis. In short, they have many of the
elements of a good incident investigation program already in place. However, where most of
these programs fall short is in the areas of identifying human factors causes and determining
how best to correct these problems.
While a number of companies attempt to consider operator errors during incident
investigations, these operator errors represent only the tip of the human factors iceberg. As
described in more detail in Section 2, most human factors causes originate further up the
organizational chain, taking the form of poor management decisions, inadequate staffing,
inadequate training, poor workplace design, etc. Simply identifying the mistake an operator
made, and not drilling down to identify the underlying, organizational causes of that mistake,
will not help to prevent reoccurrences of the incident. Because most offshore incident
investigation programs do not have a thorough process for identifying the many types of human
error, and the various levels of the organizations from which such errors originate, they lack the
tools with which to make effective, human error-reducing, and thus incident-reducing, changes.
The remainder of this paper will provide the tools to understand, investigate for, and
productively solve human error causes of incidents:
Section 2 will describe what human error is and how it causes incidents; it will also
discuss some of the most pervasive types of human error in the maritime and offshore
industries.
Section 3 outlines the keys to building a successful human factors incident investigation
and analysis program. It will describe in detail the concept of an organizations layers of
system defenses against catastrophic events, and how a weakening of these system
defenses can result in incidents.
Section 4 presents the Human Factors Analysis and Classification System (HFACS) a
simple to understand and use system for categorizing the types of human errors at each
layer of system defense. HFACS has been used successfully in military and other
industrial applications, and is compatible with maritime and offshore needs.
Section 5 walks you through the analysis of an incident. It introduces Events and
Causal Factors Charting, a method which first determines the events which occurred in
the evolving incident (similar to a timeline analysis), and then considers the contributing
causes to each event. The combination of Events and Causal Factors Charting,
followed by an HFACS analysis of the causes, provides a powerful tools for ferreting out
the underlying human error contributions to an incident.
Data analysis is the topic of Section 6. Several different approaches are introduced,
allowing companies to go well beyond the simple frequency and trend analysis in
common usage today. Proactive, thoughtful data analysis is key to a companys
awareness of safety issues and their probable underlying causes.
Finding effective safety solutions is discussed in Section 7. This section takes you from
the results of your HFACS and data analyses, through focused information-gathering on
safety problems, to crafting effective interventions. The triangle of effectiveness is a
Page 3 of 141
Houston, Texas
tool that will help you to find the most effective human factors interventions and safety
solutions.
Section 8 presents the lessons learned and success stories from three organizations
now involved in human factors incident investigation and analysis. Their experiences
can give you a head start with your own program.
Section 9 wraps up the paper with a summary of the most important points discussed.
The Appendices provide you with sample human factors-related questions to ask during
an investigation, additional examples of human factors incident investigation
classification schemes, database elements to capture the non-human factors relevant to
an investigation, and specific data elements that can help to identify and understand
three of the most prevalent types of human error in maritime/offshore accidents: fatigue,
inadequate communications, and limitations in skill and knowledge.
When a focus on human error is incorporated into your existing incident investigation, analysis,
and intervention program (as in Fig. 1), it can produce great benefits for your company,
including fewer incidents, fewer lost-time accidents, improved employee morale, greater
productivity, and an overall improvement to the bottom line.
Page 4 of 141
Houston, Texas
,QFLGHQW,QYHVWLJDWLRQ3URFHVV
Investigative
Identify Team Leader

& Team Members
:KR"
:KDW"
:KHQ"
:KHUH"
+RZ"
Develop Incident
Statement
Gather
Facts
3K\VLFDO'DWD
,QWHUYLHZLQJ
Develop Timeline
$FKURQRORJLFDOVHTXHQFHRI
HYHQWVWKDWVWDUWVRQ
GHYLDWLRQIURPQRUPDO
FRQGLWLRQVDQGHQGVRQ
UHWXUQWRQRUPDOFRQGLWLRQV
Identify Causes
Improvement
INVESTIGATION PHASES
Preliminary
7KH7HDP/HDGHULVUHVSRQVLEOHIRUHYHU\WKLQJ6FKHGXOLQJ
$VVLJQLQJ7HDP0HPEHU5HVSRQVLELOLWLHV/HDGLQJWKH
%UDLQVWRUPLQJ$FWLYLWLHV&RPSOHWLQJWKH,QYHVWLJDWLRQ5HSRUW
DQG,QIRUPLQJ$IIHFWHG
(PSOR\HHV
'HYHORS
(YHQW
&DXVDO
)DFWRUV
&KDUW
Events
Causal
Factors
Identify
Human Errors
l
l l l
l
ll
l
9HULI\UHFRPPHQGDWLRQV
HOLPLQDWHKXPDQDQGRWKHUFDXVDO
IDFWRUVDQGWKDWWKH\FDQEHFORVHG
Report
Make
Recommendations
9HULI\
UHSRUWLV
FRPSOHWH
DQG
DFFXUDWH
Complete
Investigation
Report
:K\"
Track
Recommendations
Until Closed
Figure 1.
Incident Investigation Process.
(modified from Franklyn, 2000)
Page 5 of 141
Houston, Texas
2.0
HUMAN ERROR3
Over the last 40 years or so, the shipping industry has focused on improving the structure of
ships and platforms and improving the reliability of equipment systems in order to reduce
casualties and increase efficiency and productivity. Todays maritime and offshore systems are
technologically advanced and highly reliable. Yet, the maritime casualty rate is still high. The
reason for this is because ship/offshore structure and system reliability are a relatively small
part of the safety equation. The maritime system is a people system, and human errors figure
prominently in casualty situations. About 75-96% of marine casualties are caused, at least in
part, by some form of human error. Studies have shown that human error contributes to:
84-88% of tanker accidents (Transportation Safety Board of Canada, 1994)
79% of towing vessel groundings (Cormier, 1994)
89-96% of collisions (Bryant, 1991; U.K. P&I Club, 1992)
75% of allisions (Bryant, 1991)
75% of fires and explosions (Bryant, 1991)
Therefore, if we want to make greater strides towards reducing marine casualties, we must
begin to focus on the types of human errors that cause casualties.
One way to identify the types of human errors relevant to the maritime and offshore industries is
to study incidents and determine how they happen. Chairman Jim Hall of the National
Transportation Safety Board (NTSB) has said that accidents can be viewed as very successful
events. What Chairman Hall means by successful is that it is actually difficult to create an
accident (thank goodness!). Accidents are not usually caused by a single failure or mistake, but
by the confluence of a whole series, or chain, of errors. In looking at how accidents happen, it
is usually possible to trace the development of an accident through a number of discrete events.
A Dutch study of 100 marine casualties (Wagenaar & Groeneweg, 1987) found that the number
of causes per accident ranged from 7 to 58, with a median of 234. Minor things go wrong or little
mistakes are made which, in and of themselves, may seem innocuous. However, sometimes
when these seemingly minor events converge, the result is a casualty. In the study, human
error was found to contribute to 96 of the 100 accidents. In 93 of the accidents, multiple human
errors were made, usually by two or more people, each of whom made about two errors apiece.
But here is the most important point: every human error that was made was determined to be a
necessary condition for the accident. That means that if just one of those human errors had not
occurred, the chain of events would have been broken, and the accident would not have
happened. Therefore, if we can find ways to prevent some of these human errors, or at least
increase the probability that such errors will be noticed and corrected, we can achieve greater
marine safety and fewer casualties.
This section is taken from Rothblum (2000).

This means that half the accidents had 7-23 causes and the other half of the accidents had 23-58
causes.
Page 6 of 141
Houston, Texas
2.1
Types of Human Error
What do we mean by human error? Human error is sometimes described as being one of the
following: an incorrect decision, an improperly performed action, or an improper lack of action
(inaction). Probably a better way to explain human error is to provide examples from two real
marine casualties.
The first example is the collision of the M/V SANTA CRUZ II and the USCG Cutter
CUYAHOGA, which occurred on a clear, calm night on the Chesapeake Bay (Perrow, 1984).
Both vessels saw each other visually and on radar. So what could possibly go wrong? Well,
the CUYAHOGA turned in front of the SANTA CRUZ II. In the collision that ensued, 11 Coast
Guardsmen lost their lives.
What could have caused such a tragedy?
Equipment
malfunctions? Severe currents? A buoy off-station? No, the sole cause was human error.
There were two primary errors that were made. The first was on the part of the CUYAHOGAs
captain: he misinterpreted the configuration of the running lights on the SANTA CRUZ II, and
thus misperceived its size and heading. When he ordered that fateful turn, he thought he was
well clear of the other vessel. The second error was on the part of the crew: they realized what
was happening, but failed to inform or question the captain. They figured the captains
perception of the situation was the same as their own, and that the captain must have had a
good reason to order the turn. So they just stood there and let it happen. Another type of
human error that may have contributed to the casualty was insufficient manning (notice that this
is not an error on the part of the captain or crew; rather, it is an error on the part of a
management decision-maker who determined the cutters minimum crew size). The vessel
was undermanned, and the crew was overworked. Fatigue and excessive workload may have
contributed to the captains perceptual error and the crews unresponsiveness.
The second example is the grounding of the TORREY CANYON (Perrow, 1984). Again we
have clear, calm weather--this time it was a daylight transit of the English Channel. While
proceeding through the Scilly Islands, the ship ran aground, spilling 100,000 tons of oil.
At least four different human errors contributed to this incident. The first was economic
pressure, that is, the pressure to keep to schedule (pressure exerted on the master by
management). The TORREY CANYON was loaded with cargo and headed for its deep-water
terminal in Wales. The shipping agent had contacted the captain to warn him of decreasing
tides at Milford Haven, the entrance to the terminal. The captain knew that if he didnt make the
next high tide, he might have to wait as much as five days before the water depth would be
sufficient for the ship to enter.
Page 7 of 141
Houston, Texas
This pressure to keep to schedule was exacerbated by a second factor: the captains vanity
about his ships appearance. He needed to transfer cargo in order to even out the ships draft.
He could have performed the transfer while underway, but that would have increased the
probability that he might spill a little oil on the decks and come into port with a sloppy ship. So
instead, he opted to rush to get past the Scillies and into Milford Haven in order to make the
transfer, thus increasing the pressure to make good time.
The third human error in this chain was another poor decision by the master. He decided, in
order to save time, to go through the Scilly Islands, instead of around them as originally
planned. He made this decision even though he did not have a copy of the Channel Pilot for
that area, and even though he was not very familiar with the area.
The final human error was an equipment design error (made by the equipment manufacturer).
The steering selector switch was in the wrong position: it had been left on autopilot.
Unfortunately, the design of the steering selector unit did not give any indication of its setting at
the helm. So when the captain ordered a turn into the western channel through the Scillies, the
helmsman dutifully turned the wheel, but nothing happened. By the time they figured out the
problem and got the steering selector back on manual, it was too late to make the turn, and
the TORREY CANYON ran aground.
As these two examples show, there are many different kinds of human error. It is important to
recognize that human error encompasses much more than what is commonly called operator
error. In order to understand what causes human error, we need to consider how humans work
within the maritime system.
2.2
The Maritime System: People, Technology, Environment, and Organizational

Factors
As was stated earlier, the maritime system is a people system (Fig. 2). People interact with
technology, the environment, and organizational factors. Sometimes the weak link is with the
people themselves; but more often the weak link is the way that technological, environmental, or
organizational factors influence the way people perform. Lets look at each of these factors.
Page 8 of 141
Houston, Texas
E nvironment
T echnology
Organization
Figure 2.
The Maritime System Is A People System
First, the people. In the maritime system this could include the ships crew, pilots, dock
workers, Vessel Traffic Service operators, and others. The performance of these people will be
dependent on many traits, both innate and learned (Fig. 3). As human beings, we all have
certain abilities and limitations. For example, human beings are great at pattern discrimination
and recognition. There isnt a machine in the world that can interpret a radar screen as well as
a trained human being can. On the other hand, we are fairly limited in our memory capacity and
in our ability to calculate numbers quickly and accurately--machines can do a much better job.
In addition to these inborn characteristics, human performance is also influenced by the
knowledge and skills we have acquired, as well as by internal regulators such as motivation and
alertness.
Page 9 of 141
Houston, Texas
Figure 3.
Knowledge
Skills
Abilities
Memory
Motivation
Alertness
The Maritime System: People
The design of technology can have a big impact on how people perform (Fig. 4). For example,
people come in certain sizes and have limited strength. So when a piece of equipment meant
to be used outdoors is designed with data entry keys that are too small and too close together
to be operated by a gloved hand, or if a cutoff valve is positioned out of easy reach, these
designs will have a detrimental effect on performance. Automation is often designed without
much thought to the information that the user needs to access. Critical information is
sometimes either not displayed at all or else displayed in a manner which is not easy to
interpret. Such designs can lead to inadequate comprehension of the state of the system and
to poor decision making.
Page 10 of 141
Houston, Texas
Anthropometry
Reach, strength, agility
Equipment layout
Perception &
Information display
Maintenance
comprehension
Decision-making
Safety & performance
Figure 4.
The Maritime System: Effect of Technology on People
The environment affects performance, too (Fig. 5). By environment we are including not only
weather and other aspects of the physical work environment (such as lighting, noise, and
temperature), but also the regulatory and economic climates. The physical work environment
directly affects ones ability to perform. For example, the human body performs best within a
fairly restricted temperature range. Performance will be degraded at temperatures outside that
range, and fail altogether in extreme temperatures. High sea states and ship vibrations can
affect locomotion and manual dexterity, as well as cause stress and fatigue. Tight economic
conditions can increase the probability of risk-taking (e.g., making schedule at all costs).
Finally, organizational factors, both crew organization and company policies, affect human
performance (Fig. 6). Crew size and training decisions directly affect crew workload and their
capabilities to perform safely and effectively. A strict hierarchical command structure can inhibit
effective teamwork, whereas free, interactive communications can enhance it. Work schedules
which do not provide the individual with regular and sufficient sleep time produce fatigue.
Company policies with respect to meeting schedules and working safely will directly influence
the degree of risk-taking behavior and operational safety.
Page 11 of 141
Houston, Texas
Temperature, noise
Sea state, vibration
Regulations
Economics
Figure 5.
Physical and mental

performance
Fatigue
Risk-taking
The Maritime System: Effect of Environment on People
Work schedules
Fatigue
Crew complement
Knowledge & skills
Training
Work practices
Communication
Teamwork
Safety culture
Risk-taking
Figure 6.
The Maritime System: Effect of Organization on People
Page 12 of 141
Houston, Texas
As you can see, while human errors are all too often blamed on inattention or mistakes on
the part of the operator, more often than not they are symptomatic of deeper and more
complicated problems in the total maritime system. Human errors are generally caused by
technologies, environments, and organizations which are incompatible in some way with optimal
human performance. These incompatible factors set up the human operator to make
mistakes. So what is to be done to solve this problem? Traditionally, management has tried
either to cajole or threaten its personnel into not making errors, as though proper motivation
could somehow overcome poorly designed management and equipment systems and inborn
human limitations. In other words, the human has been expected to adapt to the system. This
does not work. Instead, what needs to be done is to adapt the system to the human.
The discipline of human factors is devoted to understanding human capabilities and limitations,
and to applying this information to design equipment, work environments, procedures, and
policies that are compatible with human abilities.
In this way we can design technology,
environments, and organizations which will work with people to enhance their performance,
instead of working against people and degrading their performance. This kind of humancentered approach (that is, adapting the system to the human) has many benefits, including
increased efficiency and effectiveness, decreased errors and incidents, decreased training
costs, decreased personnel injuries and lost time, and increased morale.
2.3
Human Factors Issues in the Maritime Industry
What are some of the most important human factors challenges facing the maritime industry
today? A study by the U.S. Coast Guard (1995) found many areas where the industry can
improve safety and performance through the application of human factors principles. Below are
summaries of the top ten human factors areas that need to be improved in order to prevent
casualties.
2.3.1
Fatigue
The NTSB has identified fatigue to be an important cross-modal issue, being just as pertinent
and in need of improvement in the maritime industry as it is in the aviation, rail, and automotive
industries. Fatigue has been cited as the number one concern of mariners in two different
studies (Marine Transportation Research Board, 1976; National Research Council, 1990). It
was also the most frequently mentioned problem in a recent Coast Guard survey (U.S. Coast
Guard, 1995). A recent study has objectively substantiated these anecdotal fears: in a study of
critical vessel casualties5 and personnel injuries, it was found that fatigue contributed to 16% of
the vessel casualties and to 33% of the injuries (McCallum, Raby, & Rothblum, 1996).
A critical vessel casualty was defined as a vessel casualty in which there was significant damage to
the vessel or property, or in which the safety of the crew was at risk.
Page 13 of 141
Houston, Texas
2.3.2
Inadequate Communications
Another area for improvement is communications--between shipmates, between masters and

pilots, ship-to-ship, and ship-to-VTS. An NTSB report (National Transportation Safety Board,
1981) stated that 70% of major marine collisions and allisions occurred while a State or federal
pilot was directing one or both vessels. Better procedures and training can be designed to
promote better communications and coordination on and between vessels. Bridge Resource
Management (BRM) is a first step towards improvement.
2.3.3
Inadequate General Technical Knowledge
In one study, this problem was responsible for 35% of casualties (Wagenaar & Groeneweg,
1987). The main contributor to this category was a lack of knowledge of the proper use of
technology, such as radar. Mariners often do not understand how the automation works or
under what set of operating conditions it was designed to work effectively. The unfortunate
result is that mariners sometimes make errors in using the equipment or depend on a piece of
equipment when they should be getting information from alternate sources.
2.3.4
Inadequate Knowledge of Own Ship Systems
A frequent contributing factor to marine casualties is inadequate knowledge of own ship

operations and equipment. Several studies and casualty reports have warned of the difficulties
encountered by crews and pilots who are constantly working on ships of different sizes, with
different equipment, and carrying different cargoes. The lack of ship-specific knowledge was
cited as a problem by 78% of the mariners surveyed (National Research Council, 1990). A
combination of better training, standardized equipment design, and an overhaul of the present
method of assigning crew to ships can help solve this problem.
2.3.5
Poor Design of Automation
One challenge is to improve the design of shipboard automation. Poor design pervades almost
all shipboard automation, leading to collisions from misinterpretation of radar displays, oil spills
from poorly designed overfill devices, and allisions due to poor design of bow thrusters. Poor
equipment design was cited as a causal factor in one-third of major marine casualties
(Wagenaar & Groeneweg, 1987). The fix is relatively simple: equipment designers need to
consider how a given piece of equipment will support the mariners task and how that piece of
equipment will fit into the entire equipment suite used by the mariner. Human factors
engineering methods and principles are in routine use in other industries to ensure humancentered equipment design and evaluation. The maritime industry needs to follow suit.
2.3.6
Decisions Based on Inadequate Information
Mariners are charged with making navigation decisions based on all available information. Too
often, we have a tendency to rely on either a favored piece of equipment or our memory. Many
casualties result from the failure to consult available information (such as that from a radar or an
echo-sounder). In other cases, critical information may be lacking or incorrect, leading to
navigation errors (for example, bridge supports often are not marked, or buoys may be offstation).
Page 14 of 141
Houston, Texas
2.3.7
Poor Judgement
Risky decisions can lead to accidents. This category contained actions that were not consistent
with prudent seamanship, such as passing too closely, excessive speed, and ignoring potential
risks.
2.3.8
Faulty Standards, Policies, or Practices
This is an oft-cited category and covers a variety of problems. Included in this category is the
lack of available, precise, written, and comprehensible operational procedures aboard ship (if
something goes wrong, and if a well-written manual is not immediately available, a correct and
timely response is much less likely). Other problems in this category include management
policies which encourage risk-taking (like pressure to meet schedules at all costs) and the lack
of consistent traffic rules from port to port.
2.3.9
Poor Maintenance
Published reports (Bryant, 1991; National Research Council, 1990) and survey results (US
Coast Guard, 1995) expressed concern regarding the poor maintenance of ships. Poor
maintenance can result in a dangerous work environment, lack of working backup systems, and
crew fatigue from the need to make emergency repairs. Poor maintenance is also a leading
cause of fires and explosions (Bryant, 1991).
2.3.10 Hazardous Natural Environment
The marine environment is not a forgiving one. Currents, winds, ice, and fog make for
treacherous working conditions. When we fail to incorporate these factors into the design of our
ships, platforms, and equipment, and when we fail to adjust our operations based on hazardous
environmental conditions, we are at greater risk for casualties.
These and other human errors underlie almost every maritime incident. By studying incidents to
understand their contributing causes, we can learn how to redesign our policies, procedures,
work environments, and equipment to be more compatible with our human users and, thus,
bring about improved safety and productivity. In the next sections we will discuss how to
develop a human factors incident investigation program for your company.
3.0
BUILDING A HUMAN FACTORS INCIDENT INVESTIGATION PROGRAM
A well-designed company safety program is multi-faceted. Health, Safety, and Environment

(HSE) management, risk assessment and management, behavior based safety management
(BBSM), quality programs, and project management all play a role in improving safety. A
careful Job Hazards Analysis (JHA) or Job Safety Analysis (JSA) can identify work hazards and
recommend redesigns of equipment and work procedures, as well as associated precautions
that can prevent accidents.
Page 15 of 141
Houston, Texas
Behavior based safety management and other related processes work hand-in-hand with
incident investigation to identify potential problems. Behavior based safety management is a
proactive process which examines the workplace to identify problems before an incident occurs,
while incident investigation is a reactive process which identifies workplace and procedural
hazards that caused an accident or near-miss. The data from both types of processes should
be used together to gain the most complete understanding of potential hazards. The fact that
both these processes can be used to prevent incidents was underscored by a fatal fall that
occurred at a construction site. Just prior to the accident, the companys BBSM data had
shown that personnel were not hooking up or using fall protection properly. The data also
identified several barriers to the safe behavior, including: lack of available hook up points; lack
of available fall protection in high hazard areas; lack of training on proper use of protective
equipment; unclear procedures; and discomfort associated with wearing harnesses. By
collecting these kinds of upstream indicators, a company can correct the situation before an
incident happens.
Unfortunately, were not always able to foresee and prevent every type of incident that might
occur. This is what makes incident investigation an important part of the companys overall
safety strategy. An incident investigation and analysis program is essential to understanding
the underlying, and sometimes hidden, causes of workplace incidents. Proper identification of
the true contributors to accidents allows a company to establish workable preventive measures.
This section discusses how to build a human factors incident investigation program.
Additional information and publications on human factors and incident investigation in the
offshore industry are offered on several web sites. The American Institute of Chemical
Engineers (http://www.aiche.org) offers two documents, Guidelines for Investigating Chemical
Process Incidents and International Conference and Workshop of Process Industry Incidents.
The U.K.s Health and Safety Executive (http://www.open.gov.uk/hse/hsehome.htm) has the
publication Human and Organisation Factors in Offshore Safety. The International Association
of Oil and Gas Producers (http://www.ogp.org.uk) has a safety incident reporting system and
incident statistics.
3.1
Keys to a Successful Human Factors Incident Investigation Program
Before we jump into the details of how to investigate for human factors causes of incidents, it is
important to mention a few key factors which will encourage cooperation in incident
investigations and will promote good data quality. These key factors are: an open, fair,
improvement-seeking culture; an understanding of the purpose and scope of the incident
investigation program; training for investigators on human factors; a database classification
scheme (taxonomy) that supports the goals of the incident investigation program; a simple,
user-friendly way of entering incident data; and feedback to show how incident data have been
used to improve safety (Hill, Byers, & Rothblum, 1994).
Page 16 of 141
Houston, Texas
3.1.1
An Open, Fair, Improvement-Seeking Culture
The fundamental purpose of an incident investigation is to understand the circumstances and

causes of the incident with the aim of improving safety. We want to understand: what
happened; how it happened; why it happened; and, most importantly, what steps can be taken
to prevent it from happening again. Only by dispassionately analyzing the incident evolution in
detail and determining its underlying contributing factors can we design and implement effective
remedial actions. It is important to remember that we are not out to attribute blame: actions
taken solely to blame and shame generally do little to prevent similar incidents from occurring
in the future. This is because, as discussed in the previous section, most incidents are not the
fault of a given person; rather, they are indicative of deficiencies within the system.
Companies whose incident investigations focus on finger-pointing (i.e., identifying the person
who is supposedly to blame for the incident) short-circuit their ability to find and understand
the real causes of the incident. Only by analyzing and addressing the contributing factors the
system deficiencies that underpin the actions of those directly involved, can we make real
progress in reducing the frequency of incidents. Therefore, it is necessary to foster an open
and trusting environment where personnel feel free to discuss the evolution of an incident
without fear of unjust reprisal. If personnel know that the purpose of the investigation is to
identify how to improve safety, and that the investigation will lead to a fair and objective analysis
of the incident, they will be much more likely to participate in a candid interview. Without such a
supportive environment, involved individuals will be reluctant to cooperate in a full disclosure of
the events leading to an incident.
3.1.2
Common Understanding of the Purpose and Scope of the Incident Investigation

Program
The incident investigation program, and the database which supports it, should be constructed
to accomplish a well-defined purpose. Program managers need to agree on specific questions
the program and, therefore, the incident database will be expected to answer. For example,
a company might wish to focus on reducing maintenance incidents which result in lost time for
the employee. Such a program, and its database, would need information on the type of
maintenance activity being performed, the type of injury sustained (accident) or narrowly
avoided (near-miss), damage to equipment or workplace, lost time and money due to injury (or
potential loss, in the case of a near-miss), and causes of the of the (near-) injury (such as poor
standard operating procedures, insufficient lighting, undermanning, equipment defects,
inadequate task design, lack of safety policies, etc.). In contrast, a program focused on
preventing hazardous material spills/emissions could have a significantly different set of factors
of interest (such as type of hazmat, regulations violated, location and size of spill, fines and
clean-up costs; operational activities at time of spill; events and underlying causes leading to
the spill). The point here is that the goals of your incident investigation program must drive the
types of questions you will want to answer, which in turn dictate the types of data you will collect
during the investigations.
Page 17 of 141
Houston, Texas
A knowledge of the purpose of the database will guide the form the investigation takes and will
help in determining the appropriate resources to devote to the investigation. If certain causal
areas are known to be particularly important, effort will be concentrated in those areas.
Conversely, if the investigators do not understand the purpose of the program, they will shape
their investigations around their own biases and areas of expertise, rather than around the goals
of the program.
Clear guidance is needed for investigators to know what level of detail is sufficient, and what
resources are needed to properly fulfill the purpose of their investigations and thoroughly report
their findings. In the offshore industry this is often formalized through a charter which is
developed at the beginning of an incident investigation. The charter identifies the investigation
team, states the responsibilities of the team, its goals (e.g., to identify causes or to develop
recommendations), and a timeline for the investigation. The incident investigators, and all
personnel, must understand the program goals and how their input will help promote safety
improvements. Only then will the investigators know what types of data are important to collect,
and only then will employees understand why their active cooperation is important.
3.1.3
Appropriate Training for Incident Investigators
An incident investigation program rests on the abilities of its investigators. Incident investigation
does not come naturally: it must be trained. Investigators need background on how incidents
evolve and the myriad events and attributes which can cause or contribute to the severity of an
incident. They need to know how to ask appropriate questions, how to work with uncooperative
witnesses, how to build an events and causal tree (or other tool to help guide the investigation).
And, of course, they need to understand the specific goals of the companys incident
investigation program.
Human factors-related information is often overlooked even by seasoned investigators if they
have not been specifically trained to identify such data. While it is both natural and expected
that investigators will use their individual experiences and unique areas of expertise (e.g.,
engineering, navigation, drilling) when conducting investigations, some individuals may not
have an adequate perspective to search for or recognize human-related causes. A related
problem is that if a human factors element is not overlooked entirely, it is often oversimplified. A
single obvious human-related contributing factor may be identified, such as inattention,
without looking for the root cause (perhaps information overload, as a result of a poor display
design). As described earlier, many external factors (technology, organization, environment)
affect human performance, and it takes training for investigators to understand and recognize
these underlying contributors.
In the offshore industry, it is fairly common for a company to charter an incident investigation
team when an incident occurs. These ad hoc team members may include a combination of
workers, line supervisors, and managers. It is important to choose a team that will be fair,
unbiased, and objective. While team members are usually chosen because of their experience
Page 18 of 141
Houston, Texas
in the area where the incident occurred, they may have little or no background in either incident
investigation or human factors. If the team is going to be successful at identifying the
underlying causes of the incident, then at least some of the team members must have training
and experience in human factors incident investigation.
3.1.4
Incident Database Classification Scheme
The database classification scheme (taxonomy) must be directly linked to the purpose and
scope of the incident investigation program. The database elements must match the level of
detail that is needed to answer the safety-related questions upon which the program goals are
based. Too often an incident database is constructed in a haphazard way, with the program
managers trying to think up data elements without first determining the questions the database
is meant to answer. The sad result is a database of little value, which falls far short of
supporting safety improvements.
When it comes to human factors information, the database must be compatible with both the
program goals and the level of knowledge of the investigators. The terminology used in the
classification scheme must be well-defined and understood by the investigators. In some
cases, tools may be needed to help the investigator determine whether a given human factor is
related to the incident.
For example, the term fatigue is very hard to define many of us carry our own beliefs (correct
or incorrect) as to what fatigue is and how it relates to safety. In order to obtain reliable and
valid data on fatigue, it may be useful to determine specific pieces of data the investigator would
collect and to provide an algorithm that would use these data to determine whether fatigue
played a role in the incident. An example of this is the Fatigue Index Score being used by the
U.S. Coast Guard (McCallum, Raby, & Rothblum, 1996; see App. G): investigators collect the
number of hours worked and slept in the twenty-four hours preceding the casualty and also
collect information on fatigue symptoms (e.g., difficulty concentrating, heavy eyelids, desire to
sit down). These data are put into an equation which tells the investigator whether fatigue is a
likely cause, and therefore whether a more extensive investigation needs to be done to
determine what contributed to the fatigue. When the classification scheme is based on welldefined, quantifiable data, it increases the reliability and validity of the human factors causes
identified (e.g., fatigue), and, more importantly, it keeps the investigator focused on why the
human factors cause was present (e.g., insufficient sleep due to extended port operations).
A good database should also be adaptable to the changing needs of the organization. As the
organization learns lessons from the incident data, it is probable that additional items or levels
of detail will be desired from investigations, requiring a modification of the classification scheme
and database. One final note: while a classification scheme is extremely helpful for data
analysis, it can never capture the flavor of the incident. Narrative sections are crucial for a full
understanding of the evolution of the incidents and for capturing important information that just
does not fit into the taxonomy.
Page 19 of 141
Houston, Texas
3.1.5
Simple Data Entry
An incident database should reside on a computer system so that data analyses can be
performed. It is best to have the investigators enter their own incident data, as a clerk may
easily misread or misunderstand the investigators notes. The user interface of the database
needs to be efficient and user-friendly in order to promote data validity and completeness.
Unfortunately, examples of poor user interfaces abound. Just as the classification scheme will
determine the data collected and reported, the computer interface will determine the quality of
the data entered. If a certain data field is required to be filled out, it will always be filled out,
even if the data entered are of questionable quality. When the computer interface is poorly
designed the system becomes an obstacle to be overcome, and effort will be focused on just
getting something into the system, rather than spending effort on the veracity and
completeness of the data entered (Hill, Byers, & Rothblum, 1994). A good incident database
must be simple to use, allowing investigators to enter all relevant data easily and completely,
and allowing them to skip data fields that do not pertain to the case.
3.1.6
Feedback on Results of the Incident Investigation Program
Nothing dulls an investigators enthusiasm more than to be working hard to capture useful data,
only to get the feeling that its all going down some deep, dark hole. Feedback is crucial to a
successful incident investigation program. Investigators need to see the results of their work.
And all personnel need to know that the program is not just another flash in the pan, but
something to which management has an on-going commitment. Publish results of incident
analyses, make specific incidents the topic of safety meetings, use the results to start
discussions on how to improve safety, and let personnel know that the new policies going into
effect were based on lessons learned from incident investigations. When the use of the
incident database is made public, investigators will redouble their efforts to collect complete
data, and personnel will be more likely to cooperate in investigations.
3.2
Investigating for Human Factors Causes
Historically, companies and agencies that investigate incidents have overlooked human factors
causes almost entirely. Material deficiencies in incidents (for example, equipment malfunction
or a deficiency in the structural integrity of the vessel or platform) can normally be readily
identified (e.g., a shaft is broken or theres a hole in the hull). However, the real difficulty in
incident investigation is to answer why these deficiencies occurred, and the answer is usually
related to human behavior. For instance, the shaft may have broken because of company
management decisions, such as cutting back on maintenance, purchasing a less costly (and
less well-made) piece of equipment, or selecting less-experienced engineers. Or, the shaft may
have broken due to poor supervision of operations or maintenance, or due to an error made
during maintenance, or to someone who used the equipment outside its safe operating range.
Each of these underlying factors needs to be probed for why it happened, as well (e.g., did the
company cut back on maintenance to save money or to offload its minimally-manned crew?
was the equipment operated outside its range due to inexperience or willful violation by the
Page 20 of 141
Houston, Texas
operator?). Only after the investigator understands the true underlying cause(s) can meaningful
solutions be developed. In typical investigations, however, the why is often ignored.
Another problem with the way most investigations unfold is that individuals are usually targeted
for either incompetence or for criminal negligence. This is particularly true when the
investigator discovers that a given individual appeared to be responsible for the incident
because the individual: had fallen asleep on duty; was under the influence of alcohol or drugs
on duty; violated a regulation or standard operating procedure; appeared to be inattentive; or
made an inappropriate decision. While it is sometimes the case that an individual is
incompetent or negligent, the investigator should always look for contributing causes or other
factors underpinning such behavior. It is often the case that work policies, standard operating
procedures, and poorly designed jobs or equipment are at the core of the problem. Sanctioning
the individual will not solve the problem and only creates a culture of fear and secrecy.
Discovering the real reasons which underlie a given incident and working to solve the core
issues will engender trust and openness in the work culture and lead to real improvements in
safety.
3.3
How an Incident Evolves
There are usually multiple causes of an incident, with multiple people and events contributing to
its evolution. As mentioned in the Introduction, the accidents studied in detail by Wagenaar and
Groeneweg (1987) had anywhere from 7 to 58 distinct causes, with 50% of the cases having at
least 23 causes. We are often very good at identifying the error most immediately linked to an
incident. This is usually an error made by one of the people at the scene of the incident, such
as that made by the helmsman of the TORREY CANYON when he failed to take the ship off
automatic pilot in time to make the turn. We call these active failures because they represent
an action, inaction, or decision that is directly related to the incident. However, we are often not
as good at identifying other contributing causes, because many of these contributing causes
may have occurred days, months, or even years before the incident in question. We call these
latent conditions, because they are error-inducing states or situations that are lying dormant
until the proper set of conditions arise which expose their unsafe attributes. One of the latent
conditions in the TORREY CANYON incident was the poor design of the steering selector
switch: it gave no indication at the helm as to whether steering was set to manual or
automatic. An even more important latent error was management pressure on the master to
keep to schedule, for that sense of urgency underlay his poor decisions. In this way the human
operator is set up to make errors because the latent conditions make the system in which he
works error-inducing rather than error-avoiding.
James Reason (1990) offered a useful paradigm, often referred to as the Swiss cheese model,
that explains how the many types of contributing factors can converge, resulting in an incident6
6
Reasons work underscores the fact that human errors and human factors relate to the entire
system, not just to an individual operator. The International Ergonomics Association defines human
factors as being concerned with the understanding of interactions among humans and other elements of
a system. Further, it states that the area of human factors considers the design of tasks, jobs,
Page 21 of 141
Houston, Texas
(Fig. 7). A company tries to promote safety and prevent catastrophic accidents by putting into
place layers of system defenses, depicted in the figure below as slices of Swiss cheese.
Essentially, system defenses refers to the safety-related decisions and actions of the entire
company: top management, the line supervisors, and the workers. The Organizational Factors
layer (slice) represents the defenses put into place by top management. This level of system
defenses might include a company culture which puts safety first, and management decisions
which reinforce safety by providing well-trained employees and well-designed equipment to do
the job. The second layer of defenses is the Supervision layer. This refers to the first-line
supervisor and his or her safety-consciousness as displayed by the operational decisions he or
she makes.
Trajectory of
Events
Organizational
Factors
Unsafe
Supervision
Preconditions
for Unsafe
Unsafe
Acts
Acts
Marine
Casualty
Figure 7.
An Accident in the Making
(after Reason, 1990, as adapted by Wiegmann & Shappell, 1999)
For example, a good supervisor will ensure that personnel receive the proper training and
mentoring, that work crews have the necessary skills and work well together, and that safetyrelated procedures are used routinely. The actions and fitness for duty of the worker make up
the third layer of system defenses. In a safe system, the operator is physically and mentally
ready to perform and routinely adheres to safe operating practices and procedures.
products, environments and systems in order to make them compatible with the needs, abilities and
limitations of people. It is crucial to understand that in a human factors incident investigation we are not
looking to identify a person at fault; instead, we are looking primarily for weaknesses in the links
between the human workers and other parts of the system, such as management policies, equipment
design, and work environment.
Page 22 of 141
Houston, Texas
These system defenses can slowly erode over time in response to economic pressures,
increasing demand for products and services, diminishing attention to promoting a safety
culture, and others. Each time safety is sacrificed (e.g., by cutting back on preventive
maintenance or by taking unsafe shortcuts in operational tasks), it puts another hole into that
slice of cheese. If synergistic reductions in safety occur at all three levels of the system (that is,
when the holes in the Swiss cheese line up), then the system no longer has any inherent
protections, and it becomes an accident waiting to happen. All it takes is one mistake (unsafe
act).
Heres an example of how chipping away at system defenses can result in a casualty. Lets say
as a cost-cutting measure, a company decides to decrease the inventory of spare parts on its
ships (hole in the Organizational Factors slice). One day the ship develops engine problems
from clogged fuel injectors and doesnt have sufficient spare parts (this would be analogous to
an equipment precondition). The captain, knowing that the company would penalize him if he
spent money to be towed into port (hole in Supervision, since the captain reports to the
company), decides to take a risk and transit on only one engine (Unsafe Act). That engine
fails, and the vessel drifts and grounds.
3.4
Improving Safety through Incident Investigation and Analysis
Maritime and offshore operations are inherently risky. Company managers have to weigh oftencompeting interests in safety, productivity, profitability, and customer expectations in order to be
viable. Sometimes, well-meaning decisions back-fire and cause unanticipated safety problems.
One way management can keep its finger on the safety pulse of the company is through
incident investigation and analysis. As Fig. 8 shows, by thoroughly investigating incidents and
the human errors that cause them, one can identify the holes in the system defenses and
develop workable solutions.
An incident investigation program consists of five components (Fig. 8). First, the company must
support the investigation of incidents. This requires objective investigators with at least a
minimal amount of training in investigation techniques and a firm understanding of the purpose
of the investigation and the types of data which must be collected to support the companys
objectives. Second, the company must develop and maintain an incident database. As
mentioned earlier, such a database should be computerized for easier analysis. The database
must be composed of a set of taxonomies (classification schemes) which will capture the
incident elements of interest to the company. The database should also incorporate narrative
fields so that investigators can explain events and causes in more detail. Third, the company
must then support regular analysis of the incidents in the database. As will be discussed in
more detail later, analysis allows the company to find patterns common to a group of incidents,
and allows the determination of how frequently different types of incidents occur and, in the
case of near-misses, the potential severity of the accident that was avoided. Such data are very
helpful in targeting the types of safety problems that the company will want to spend time and
money to solve.
Page 23 of 141
Houston, Texas
Effective
Safety
Programs
Data-Driven
Research
Results of analysis
used to identify
processes to study
Research is needsbased and interventions are effective
Human
Error
Incident
Investigation
Must look
beyond
obvious
cause
Errors occur
frequently
Errors are a
major cause
of accidents
Data are
often qualitative and
subjective
Incident
Database
Requires
new, human
factors lexicon
Terms must
be welldefined,
understood by
investigators
Database
Analysis
Analysis helps
identify safety
problems
Different levels
of analysis
possible
Feedback
Figure 8.
How An Effective Human Factors Incident Investigation Program

Can Improve Safety
(Modified from Wiegmann and Shappell, 1999)
The fourth component of a successful incident investigation program is data-driven research.

Incident investigation will frequently just skim the surface of a safety problem. The value of
incident investigation and analysis is that it identifies areas of concern. In most cases, incident
investigation and analysis will not be sufficient to solve the problem. Solving the problem will
require getting more information on the policies, standard operating procedures, common work
practices, equipment and job design, and employee attributes (like training, preparedness,
physical and mental condition) linked to the activities or situations in which the incidents have
occurred. The research might take the form of a risk assessment, or it may require the
collection of additional, detailed information in subsequent investigations of related incidents, or
perhaps a comparison of current company policies and practices with those employed by other
companies (benchmarking). Through research, the company gains a more complete
understanding of all the various contributing factors which drive the incidents of interest. The
research might then extend to a comparison of the effectiveness of different prevention
methods.
Page 24 of 141
Houston, Texas
Finally, the result of the research step is an addition to or a revision of the companys safety
program. Using the concept of barrier analysis (Hollnagel, 2000), the company wants to
understand how safety failures arise and implement barriers (such as equipment shields to
protect workers from exposure to potential harm, or procedures which prevent activities known
to be hazardous) to prevent incidents. Successful prevention can eliminate certain hazards.
Other incident causes may not be easily prevented, but there may be ways to mitigate (reduce)
their consequences. When a safety program acts on the incident data which contains
underlying causes, it will be effective.
3.5
Error Recovery
Error recovery is an important supplementary safety goal, and will be mentioned briefly here.
Many offshore companies have a zero accidents policy which, while the ultimate safety goal,
may be difficult to fully attain (Kontogiannis, 1999). In some industries, systems are being
developed which focus on preventing the consequences of human error by providing
opportunities for error recovery (Helmreich, et al., 2000; Sasou & Reason, 1999).
A framework developed by Kontogiannis (1999) categorizes error recovery according to the
process used, the outcome, and the stage of performance. The process is usually either
detection, explanation, or correction. For example, if you are using a word processor and
misspell a word, one option for the spell check function is to merely detect the misspelled word
and allow the user to decide what, if anything, to do about it. A different option has the spell
checker both detect and explain or suggest options for correcting the word. Some word
processors are even capable of correction, by automatically detecting and correcting
misspellings as you type. Similar detection, explanation, and correction features can be built
into offshore and maritime systems, either through automation or through procedures.
Outcomes of the recovery process refer to the state of the system after recovery. For example,
an error-detecting system might block the error from happening and return the system to its
original state prior to the error (this is called backward recovery). The stage of performance
relates to the stage in which the error detection was made. For example, the outcome stage
would be where an error is detected based on a mismatch between the expected outcomes of a
process and the outcomes actually observed. For instance, if there is a low pressure reading
and the operator erroneously turns the valve in the wrong direction, the outcome will be an even
lower pressure. An error detector in the outcome stage would note the discrepancy between
the present pressure reading and the higher pressure that was intended, and signal the error.
A simplified version of this framework is being used in a U.K. oil industry research project
developing a human factors investigation tool (HFIT see App. C; Gordon, Flin & Mearns,
2001). Investigators are asked three questions regarding the possible recovery process of the
error: was the error detected (realized or suspected), was it understood why the error occurred,
and was it corrected (e.g., by modifying an existing plan or developing a new plan)? How the
error was detected (e.g., via system feedback, external communications, etc.) is also discussed.
By including these types of questions into an incident investigation, it may illuminate changes to
equipment or procedures which may act to prevent such errors in the future.
Page 25 of 141
Houston, Texas
4.0
AN EXAMPLE OF A HUMAN FACTORS TAXONOMY: HFACS7
An effective incident database has at its heart a set of classification schemes or taxonomies:
schemes to classify the type of incident, the type of people involved in the incident, the type of
platform or vessel involved, the geographical area and weather conditions, the type of
equipment that failed, the activities occurring at the time of the incident, and of course, the
human factors causes. A maritime example with all these components and more is the
International Maritime Incident Safety System (IMISS; Rothblum, Chaderjian, & Mercier, 2000;
see App. B). It is important to understand all of the different types of factors (equipment,
human, weather, etc.) involved in an incident. Since the offshore and maritime industries
already have adequate ways of identifying equipment and other non-human contributions to
incidents, this section will focus only on the identification and classification of human factors
causes.
There are many, many human factors taxonomies that are in use by NASA, the Nuclear
Regulatory Commission, the Transportation Safety Board of Canada, the U.K. Marine Accident
Investigation Branch, and others. Some of these are listed in Appendix C; they are all useful
taxonomies. These taxonomies vary with respect to how they chose to group human factors
elements, the level of detail they provide, and the level of expertise required on the part of the
investigator.
In order to provide an example of a human factors taxonomy and to show how it would be used
during an investigation, we selected the Human Factors Analysis and Classification System
(HFACS; Shappell & Wiegmann, 1997a, 2000; Wiegmann & Shappell, 1999) because it is
relatively easy to learn and use, and because its effectiveness has been demonstrated through
its use by the U.S. Navy and U.S. Marine Corps for aviation accident investigation and analysis.
This classification scheme, as well as several of the others mentioned, is based on the wellestablished human error frameworks of the SHEL model (Software-Hardware-EnvironmentLiveware; Edwards (1972) and Hawkins (1984, 1987) as cited in TSB, 1998), Rasmussens
taxonomy of errors (1987, as cited in TSB, 1998), and Reasons (1990) Swiss cheese model
of accident causation.
HFACS seeks to understand all the human-related contributing causes to an incident by
considering the holes in the four layers of system defenses: unsafe acts, preconditions for
unsafe acts (unsafe conditions), unsafe supervision, and organizational factors (see Fig. 9).
The discussion below summarizes some of the types of latent conditions and active failures
associated with these layers of system defenses. For more information, please see Shappell &
Wiegmann (1997a, 2000) or Wiegmann & Shappell (1999).
7
The majority of this section has been taken from Shappell & Wiegmann (2000). For additional
information about HFACS, or to take a seminar on HFACS, please contact either Dr. Scott Shappell at
scott_shappell@mmacmail.jccbi.gov and (405) 954-4082, or Dr. Doug Wiegmann at
dwiegman@uiuc.edu and (217) 244-8637.
Page 26 of 141
Houston, Texas
ORGANIZATIONAL
INFLUENCES
Resource
Management
Organizational
Climate
Organizational
Process
UNSAFE
SUPERVISION
Inadequate
Supervision
Planned
Inappropriate
Operations
Failed to
Correct
Problem
Supervisory
Violations
PRECONDITIONS
FOR
FOR
UNSAFE
UNSAFE ACTS
ACTS
Substandard
Conditions of
Operators
Adverse
Mental States
Adverse
Physiological
States
Substandard
Practices of
Operators
Physical/
Mental
Limitations
Crew Resource
Mismanagement
Personal
Readiness
UNSAFE
ACTS
Errors
Decision
Errors
Figure 9.
Skill-Based
Errors
Violations
Perceptual
Errors
Routine
Exceptional
The Human Factors Classification and Analysis System (HFACS)
Page 27 of 141
Houston, Texas
4.1
HFACS: Unsafe Acts
In an incident investigation, the investigator starts with the immediate actions and events
surrounding the incident and then works backwards to uncover contributing causes. In terms of
human errors, those immediately linked to the incident are typically unsafe acts. There are
two types of unsafe acts: errors and violations.
Errors represent the mental and physical activities of individuals that fail to achieve their
intended outcome; that is, the result of the persons action was not as expected. For example, if
the captain orders right 20 degrees when he meant to order left 20 degrees, that would be an
error. A violation, on the other hand, is when the persons action reflects a willful disregard for
standard operating procedures or regulations (even though they probably did not intend to
cause an incident). For example, an engineer doing maintenance might decide to cut corners
and not perform a maintenance procedure the way it should be done. His performance is an
intentional violation of the correct procedure. Errors and violations can be further subdivided,
as shown in Figure 10. Errors can be decision errors, skill-based errors, or perceptual errors.
Violations can be routine or exceptional.
4.1.1
Decision Errors
The decision error represents an activity or behavior that proceeds as intended, yet the plan
proves inadequate or inappropriate for the situation. Often referred to as honest mistakes,
these unsafe acts represent the actions or inactions of individuals whose hearts are in the right
place, but they either did not have the appropriate knowledge or just simply made a poor
choice. These types of knowledge-based and rule-based errors have been referred to in
Reasons taxonomy as mistakes (TSB, 1998).
Decision errors can result from multiple causes. For example, a wrong decision can be made if
the person does not fully understand the situation at hand, misdiagnoses the problem, and
proceeds to apply the wrong solution (because hes solving the wrong problem).
Troubleshooting an electrical fault could lead to this type of procedural decision error. Decision
errors can also occur if the person does not have sufficient experience to guide his decision, or
if there is not enough time to fully work through the problem properly before a choice must be
made (called a choice decision error). Problem-solving errors can occur when the problem is a
novel one, requiring the person to reason through it.
Page 28 of 141
Houston, Texas
Examples:
Decision
Errors
misdiagnose problem
wrong response to problem
inadequate information to make
decision
Examples:
Errors
Skill-Based
Errors
do procedural steps in wrong

order or skip a step
forgetting to make a turn at a waypoint
Examples:
Perceptual
Errors
UNSAFE
ACTS
misperceive running light

configuration
hear a verbal order wrong
misjudge distance or rate
Examples:
Routine
Violation
routine shortcuts in a
maintenance procedure
habitually not donning personal
protective gear
Violations
Examples:
Exceptional
Violation
excessively risky ship-handling

intentionally miswiring a circuit to
irk supervisor
Figure 10.
Classification of Unsafe Acts
(modified from Shappell & Wiegmann, 2000, and TSB, 1998)
4.1.2
Skill-based Errors
Skill-based errors can occur in the execution of skills or procedures that have become so welllearned that they are performed almost automatically. Routine maintenance tasks, taking
navigational bearings, monitoring equipment displays, and other repetitive operations would be
considered skill-based tasks. These types of tasks are sometimes performed improperly due to
a failure of attention or memory. Consider the hapless soul who locks himself out of the car or
misses his exit because he was either distracted, in a hurry, or daydreaming. These are both
examples of attention failures that commonly occur during highly automatized behavior. Types
of attention failures include omitting a step in a procedure, reversing the order of two steps, or
Page 29 of 141
Houston, Texas
doing the right thing at the wrong time. Attentional deficits can also result in failing to detect a
problem while monitoring equipment.
In contrast to attention failures, memory failures often appear as omitted items in a checklist,
place losing, or forgotten intentions. Failures in memory can result in forgetting to do a planned
activity or losing ones place in a series of tasks. For example, most of us have experienced
going to the refrigerator only to forget what we went for. Likewise, it is not difficult to imagine
that when under stress during an operational emergency, critical steps in the emergency
procedures can be missed. Even when not particularly stressed, individuals can forget to
complete certain steps within a procedure.
If one of these types of errors is found during an incident investigation, it is a signal to look
deeper. Merely telling an operator to pay better attention next time will not solve the problem.
These types of errors are symptoms of underlying system failures. Take, for example, the fatal
accident aboard the RIX HARRIER (MAIB, 1997). On a July afternoon, the vessel was being
moored to a jetty on the River Humber. A mooring rope had been led around a fairlead, which
was situated on top of the aft bulwark rail. As the rope tightened, it sprang over the top of the
fairlead, striking the officer on his right arm and throwing him against the accommodation
bulkhead. Neither the officer nor the crew member helping him noticed that the mooring rope
had been passed inadvertently around the fairlead. This was an error resulting from lack of
attention. The investigation determined that the design of the aft mooring arrangement
increased the likelihood that such an error would be made. The investigation also determined
that, due to the ships work schedules, it was likely that the officer and crew member had
endured days of fragmented sleep and were suffering from chronic fatigue, a state that
increases the probability of attentional deficits. So in this case, both a ship design flaw and a
problem with the ships work schedules appeared to contribute to the attentional errors that
caused the death of the officer.
4.1.3
Perceptual Errors
Not unexpectedly, when ones perception of the world differs from reality, errors can, and often
do, occur. Typically, perceptual errors occur when sensory input is degraded, such as
navigating at night. Visual illusions, for example, occur when the brain tries to fill in the gaps
and make sense out of sparse information. In the earlier example of the CGC CUYAHOGA, the
captain made a perceptual error in his interpretation of the configuration of the running lights on
the SANTA CRUZ II. Had he seen the vessel in daylight, there would have been many visual
cues available to determine the type and heading of the vessel; but at night, with little visual
information available, it is all too easy to misinterpret.
Another common type of perceptual error occurs when trying to communicate in a noisy
environment. Static over the radio or noise from engines and generators can muffle or degrade
spoken words and commands. Again, the brain will attempt to fill in what wasnt heard often
based on the listeners expectations, be they correct or incorrect. As an example, a ship was
Page 30 of 141
Houston, Texas
transiting restricted waters when the Third Engineer noticed that the lube oil pressure was low.
He shouted (across a noisy engine room) to a cadet to adjust the pressure. The cadet
misunderstood (perception error) and closed the valve, causing the engine to go to dead slow
and creating a dangerous situation by greatly reducing the ships maneuverability in the hightraffic waterway (McCallum, Raby, Rothblum, Forsythe, Slavich, & Smith, 2000, unpublished).
4.1.4
Routine Violations
As discussed above, errors occur when someone is trying to follow the rules and do the right
thing, but gets an unexpected result. By contrast, a violation is when someone intentionally
ignores or bends a rule. Routine violations tend to be habitual by nature and are often
tolerated by supervision (in the case of not following a standard operating procedure) or by the
governing authority (in the case of not following a regulation) (Reason, 1990). Consider, for
example, the individual who routinely drives 5-10 mph faster than the posted speed limit (a
routine violation). Since the police will rarely pull someone over for such a minor infraction, they
are tolerating the violation and implicitly reinforcing the unsafe behavior. If the police were to
crack down on minor speeding, people would be less likely to violate the speed limit. Therefore,
if a routine violation is identified during an incident investigation, the investigator must look
further up the supervisory chain to identify those individuals in authority who are not enforcing
the rules.
4.1.5
Exceptional Violations
Unlike routine violations, exceptional violations appear as isolated departures from authority,
not necessarily indicative of the individuals typical behavior pattern nor condoned by
management (Reason, 1990). For example, an isolated instance of driving 105 mph in a 55
mph zone is considered an exceptional violation. Note that the violation is not considered
exceptional because of its extreme nature. Rather, it is considered exceptional because it is
neither typical of the individual nor condoned by authority. The fact that such behavior is not
typical of the individual makes it difficult to predict and deal with exceptional violations.
4.2
HFACS: Preconditions for Unsafe Acts
Although unsafe acts can be linked to the vast majority of incidents, simply focusing on unsafe
acts is like focusing on a fever without understanding the underlying disease causing it. Thus,
investigators must dig deeper into why the unsafe acts took place. As a first step, it is useful to
consider any preconditions for unsafe acts. There are two major subdivisions of unsafe
conditions (preconditions): substandard conditions of the operators and the substandard
practices they commit (Fig. 11). Substandard conditions are broken down into Adverse Mental
States, Adverse Physiological States, and Physical/Mental Limitations. Types of Substandard
Practices include Crew Resource Mismanagement and Personal Readiness. Each of these
subcategories is discussed below.
Page 31 of 141
Houston, Texas
Examples:
Adverse
Mental
States
loss of situational awareness

overconfidence
complacency
Examples:
Substandard
Condition of
Operators
Adverse
Physiological
States
Physical or
Mental
Limitations
PRECONDITIONS
FOR
UNSAFE ACTS
illness
circadian maladjustment
drug-induced impairment
Examples:
physical impairments (e.g.,
obesity, arthritis, poor vision)
insufficient aptitude for job
slow reaction times
Examples:
Crew
lack of pre-op team briefing
Resource
inadequate crew coordination during
Management an operation
Substandard
Practices of
Operators
Examples:
Personal
Readiness
violating work-rest rules

poor judgement such as doing
exhausting work just before
reporting for duty
Figure 11.
Classification of Preconditions for Unsafe Acts
(modified from Shappell & Wiegmann, 2000, and TSB, 1998)
4.2.1
Adverse Mental States (Substandard Conditions of Operators)
Being prepared mentally is critical in nearly every endeavor. As such, the category of Adverse
Mental States was created to account for those mental conditions that affect performance. Key
examples in the maritime and off-shore industries are loss of situational awareness,
overconfidence, and complacency. Predictably, if an individual loses situational awareness, the
likelihood increases that an error will occur. In a similar fashion, pernicious attitudes such as
overconfidence and complacency increase the likelihood that a violation will be committed.
Clearly then, any framework of human error must account for pre-existing adverse mental states
in the causal chain of events.
Page 32 of 141
Houston, Texas
4.2.2
Adverse Physiological States (Substandard Conditions of Operators)
This category refers to those medical or physiological conditions that preclude safe operations.
For example, illness can have a negative impact on our performance. Nearly all of us have
gone to work ill, dosed with over-the-counter medications, and have generally performed
sufficiently well. However, the side-effects of antihistamines, and the fatigue and sleep loss that
often accompany an illness can be detrimental to decision-making. For example, over-thecounter antihistamines decrease vigilance, performance on divided attention tasks, and shortterm memory, resulting in a 14% loss of productivity and an increase in errors (Kay, 2000).
Sleep loss, even in healthy individuals, increases the risk of accidents. Every April when the
U.S. springs ahead to daylight savings time, there is a significant increase in automobile
accidents: and this is from a mere one hour decrease in sleep time (Coren, 1998; Monk, 1980).
Therefore, it is incumbent upon any safety professional to account for these sometimes subtle
medical and physiological conditions within the causal chain of events.
4.2.3
Physical/Mental Limitations (Substandard Conditions of Operators)
The final substandard condition involves individual physical and mental limitations. Specifically,
this category refers to those instances when task or situational requirements exceed the
capabilities of the operator. For example, the human visual system is severely limited at night.
Yet, most people do not take this into account when driving a car at night, and do not slow down
or take other precautions. Similarly, there are occasions when the time required to complete a
task exceeds an individuals capacity. Individuals vary widely in their abilities to process and
respond to information. It is well documented that if individuals are required to respond quickly
(i.e., less time is available to consider all the options thoroughly), the probability of making an
error goes up markedly. Consequently, it should be no surprise that when faced with the need
for rapid processing and reaction times, as is the case in emergencies, all forms of errors would
be exacerbated.
In addition to the basic sensory and information processing limitations described above, there
are at least two additional instances of physical and mental limitations that need to be
addressed, albeit they are often overlooked by most safety professionals. These limitations
involve individuals who simply are not compatible with a given job, because they are either
unsuited physically or they do not possess the aptitude to do it. For example, some individuals
simply do not have the physical strength required to operate manual valves or haul heavy
equipment. Likewise, not everyone has the mental ability or aptitude for every job. The difficult
task for the safety professional is identifying whether physical or mental aptitude might have
contributed to the incident causal sequence.
Clearly then, numerous substandard conditions of operators can, and do, lead to the
commission of unsafe acts. Nevertheless, there are a number of things that we do to ourselves
that set up these substandard conditions. Generally speaking, the substandard practices of
Page 33 of 141
Houston, Texas
operators can be summed up in two categories: crew resource mismanagement and personal
readiness.
4.2.4
Crew Resource Mismanagement (Substandard Practices of Operators)
Operations in the off-shore and maritime industries depend on good communications and
teamwork. Communication and coordination is essential, not just between workers on a given
task, but between teams working on complementary or coordinated tasks. On a ship,
communications may be important between members of the same department (e.g., two
engineers repairing a piece of equipment, or passing information during a watch relief), between
departments (the deck officer may need to notify engineering of an upcoming maneuver),
between ships (for meeting and passing arrangements), and between the ship and other groups
or authorities such as Vessel Traffic Service, bridge tenders, dock workers, and the vessel
agent. The need for communication and coordination is often overlooked, leading to incidents.
One study of maritime casualties found that a lack of communication contributed to 18% of
vessel casualties and 28% of personnel injuries (McCallum, Raby, Rothblum, Forsythe, Slavich,
& Smith, 2000).
Heres an example of how such crew resource mismanagement can result in a serious incident.
A barge was moored to a quarry loading facility by a pull cable that was controlled from the
facility. The deckhand on the barge noticed that the pull cable was caught under a deck fitting,
and walked over to free it. Before he reached it, a dock worker started the winch to take the
slack out of the mooring line. As the cable tightened, it snapped off the fitting and struck the
deckhand with such force that he required surgery. In this case, both the deckhand and the
dock worker should have but didnt alert the other to their plans: an obvious failure of crew
coordination. A serious injury was the unhappy consequence of this lack of crew resource
management (Rothblum, 2000).
4.2.5
Personal Readiness (Substandard Practices of Operators)
In every occupation, people are expected to show up for work ready to perform at optimal levels.
Nevertheless, personal readiness failures occur when individuals fail to prepare physically,
mentally, or physiologically for duty. For instance, violations of work-rest rules, use of
intoxicants and certain medications, and participating in exhausting domestic or recreational
activities prior to reporting for duty can impair performance on the job and can be preconditions
for unsafe acts. While some of these maladaptive behaviors may be addressed by rules and
regulations, most are left up to the judgement of the individual. It is necessary for the individual
to understand that some off-time activities can be detrimental to subsequent job performance.
The incident investigator needs to probe for personal readiness and activities that may have
degraded it.
Page 34 of 141
Houston, Texas
4.3
HFACS: Unsafe Supervision
In addition to investigating those causal factors associated directly with the operator, it is
necessary to trace the possible causal chain of events up the supervisory chain of command
(Reason, 1990). It has been estimated that 80% of offshore platform accidents have their
predominant roots in supervisory and organizational factors (Bea, Holdsworth, & Smith, 1997).
There are four categories of unsafe supervision: inadequate supervision, planned inappropriate
operations, failure to correct a known problem, and supervisory violations (Fig. 12). Each is
described briefly below.
4.3.1
Inadequate Supervision
The role of any supervisor is to provide the opportunity to succeed. To do this, the supervisor,
no matter at what level of operations, must provide guidance, training opportunities, leadership,
and motivation, as well as the proper role model to be emulated. Unfortunately, this is not
always the case. For example, it is not difficult to conceive of a situation where adequate crew
resource management training was either not provided, or the opportunity to attend such
training was not afforded to a particular crew member. Conceivably, coordinated teamwork
would be compromised, and if an emergency situation arose, the risk of an error being
committed would be exacerbated and the potential for an incident would increase markedly.
In a similar vein, sound professional guidance and oversight is an essential ingredient of any
successful organization. While empowering individuals to make decisions and function
independently is certainly essential, this does not divorce the supervisor from accountability.
The lack of guidance and oversight has proven to be the breeding ground for many of the
violations that have crept into the cockpit. As such, any thorough investigation of incident
causal factors must consider the role supervision plays (i.e., whether the supervision was
inappropriate or did not occur at all) in the genesis of human error.
4.3.2
Planned Inappropriate Operations
Occasionally, the operational tempo and/or the scheduling of personnel is such that individuals
are put at unacceptable risk, crew rest is jeopardized, and ultimately performance is adversely
affected. Such operations, though arguably unavoidable during emergencies, are unacceptable
during normal operations. Therefore, the second category of unsafe supervision, planned
inappropriate operations, was created to account for these failures.
Page 35 of 141
Houston, Texas
Examples:
Inadequate
Supervision
failed to provide guidance

failed to track qualifications
failed to provide operational
doctrine
Examples:
Planned
failed to provide adequate briefing
Inappropriate improper manning/teaming
Operations
inadequate opportunity for crew
rest
UNSAFE
SUPERVISION
Failed to
Correct
Problem
Examples:
failed to identify an at-risk worker
failed to report unsafe procedure
failed to initiate corrective action
Examples:
Supervisory
Violations
authorized unnecessary hazard

failed to enforce rules and regs
authorized unqualified person for
task
Figure 12.
Categories of Unsafe Supervision.
(modified from Shappell & Wiegmann, 2000)
Take, for example, the issue of improper team complements. In aviation it is well known that
when very senior, dictatorial captains are paired with very junior, weak co-pilots, communication
and coordination problems are likely to occur. This type of personality mismatch is apt to
happen in any team environment, and can (and do) contribute to tragic accidents (such as the
crash of a commercial airliner into the Potomac River shortly after takeoff in 1982). When team
member selection is not taken into account, gross perceived differences in authority and
experience can cause more junior team members to be ignored, effectively eliminating an
important input to the team as a whole.
Page 36 of 141
Houston, Texas
4.3.3
Failure to Correct a Known Problem
The third category of known unsafe supervision, Failed to Correct a Known Problem, refers to
those instances when deficiencies among individuals, equipment, training or other related
safety areas are known to the supervisor, yet are allowed to continue unabated. For example,
a given worker might have a reputation for risky behavior or cutting safety margins too closely.
If the supervisor knows this and allows the behavior to continue, an incident may be the
unsurprising consequence. The failure to correct the behavior, either through remedial training
or, if necessary, removal from the job, can put the entire operation at risk. Likewise, the failure
to consistently correct or discipline inappropriate behavior fosters an unsafe atmosphere and
promotes the violation of rules.
4.3.4
Supervisory Violations
Supervisory violations, on the other hand, are reserved for those instances when existing rules
and regulations are willfully disregarded by supervisors. Although relatively rare, supervisors
have been known occasionally to violate the rules and doctrine when managing their assets.
For instance, sometimes individuals are assigned to do a task for which they are unqualified,
either through the lack of sufficient training, or even lacking the appropriate license. The failure
to enforce existing rules and regulations or flaunting authority are also violations at the
supervisory level. While rare and possibly difficult to identify, such practices are a flagrant
violation of the rules and invariably set the stage for the tragic sequence of events that
predictably follow.
4.4
HFACS: Organizational Influences
As noted previously, fallible decisions of upper-level management directly affect supervisory

practices, as well as the conditions and actions of operators.
Unfortunately, these
organizational errors often go unnoticed by safety professionals, due in large part to the lack of
a clear framework from which to investigate them. Generally speaking, the most elusive of
latent failures revolve around issues related to resource management, organizational climate,
and operational processes, as detailed in Figure 13.
Page 37 of 141
Houston, Texas
Examples:
Resource
Management
personnel selection, training,

manning
excessive cost cutting
purchase of unsuitable equipment
Examples:
Organizational chain-of-command, delegation of

authority, accountability
Climate
ORGANIZATIONAL
INFLUENCES
promotion and firing policies

culture -- values and norms
Examples:
Organizational op tempo, time pressure, quotas

procedures -- standards,
Process
documentation
risk mgmt and safety programs
Figure 13.
4.4.1
Organizational Factors Which Contribute to Incidents

(modified from Shappell & Wiegmann, 2000)
Resource Management
This category encompasses the realm of corporate-level decision making regarding the
allocation and maintenance of organizational assets such as human resources (personnel),
monetary assets, and equipment and facilities. Generally, corporate decisions about how such
resources should be managed center around two distinct objectives the goal of safety and the
goal of on-time, cost-effective operation. In times of prosperity, both objectives can be easily
balanced and satisfied in full. However, there may also be times of fiscal austerity that demand
some give and take between the two. Unfortunately, accident reports show us time and again
that safety is often the loser in such battles and, as some can attest to very well, safety and
training are often the first to be cut in organizations having financial difficulties. If cutbacks in
such areas are too severe, worker proficiency may suffer, leading to errors and incidents.
Page 38 of 141
Houston, Texas
Excessive cost-cutting could also result in reduced funding for new equipment or may lead to
the purchase of equipment that is sub-optimal and inadequately designed for the task. Other
trickle-down effects include poorly maintained equipment and workspaces, and the failure to
correct known design flaws in existing equipment. The result is a scenario involving
unseasoned, less-skilled workers using poorly maintained equipment under less than desirable
conditions and schedules. The ramifications for safety are not hard to imagine.
4.4.2
Organizational Climate
Climate refers to a broad class of organizational variables that influence worker performance.
In general, organizational climate can be viewed as the working atmosphere within the
organization. One telltale sign of an organizations climate is its structure, as reflected in the
chain-of-command, delegation of authority and responsibility, communication channels, and
formal accountability for actions. Just like in the operational arena, communication and
coordination are vital within an organization. If management and staff within an organization are
not communicating, or if no one knows who is in charge, organizational safety clearly suffers
and incidents do happen (Muchinsky, 1997).
An organizations policies and culture are also good indicators of its climate. Policies are
official guidelines that direct managements decisions about such things as hiring and firing,
promotion, retention, raises, sick leave, drugs and alcohol, overtime, incident investigations,
and the use of safety equipment. Culture, on the other hand, refers to the unofficial or
unspoken rules, values, attitudes, beliefs, and customs of an organization. Culture is the way
things really get done around here.
When policies are ill-defined, adversarial, or conflicting, or when they are supplanted by
unofficial rules and values, confusion abounds within the organization. Indeed, there are some
corporate managers who are quick to give lip service to official safety policies while in a public
forum, but then overlook such policies when operating behind the scenes. Safety is bound to
suffer under such conditions.
4.4.3
Organizational Process
This category refers to corporate decisions and rules that govern the everyday activities within
an organization, including the establishment and use of standardized operating procedures and
formal methods for maintaining checks and balances (oversight) between the workforce and
management. For example, such factors as operational tempo, time pressures, incentive
systems, and work schedules are all factors that can adversely affect safety (Fig. 13). There
may be instances when those within the upper echelon of an organization determine that it is
necessary to increase the operational tempo to a point that overextends a supervisors staffing
capabilities. Therefore, a supervisor may resort to the use of inadequate scheduling
procedures that jeopardize crew rest and produce sub-optimal crew complements, putting the
operation and its workers at an increased risk of a mishap. Organizations should have official
procedures in place to address such contingencies as well as oversight programs to monitor
such risks.
Page 39 of 141
Houston, Texas
Regrettably, not all organizations have these procedures nor do they engage in an active
process of monitoring operator errors and human factors problems via anonymous reporting
systems and safety audits. As such, supervisors and managers are often unaware of the
problems before an incident occurs. It is incumbent upon any organization to fervently seek out
the holes in the cheese and plug them up, before they create a window of opportunity for
catastrophe to strike.
4.5
The Benefits of Using HFACS
The Human Factors Analysis and Classification System (HFACS) framework bridges the gap
between theory and practice by providing investigators with a comprehensive, user-friendly tool
for identifying and classifying the human causes of incidents. The system, which is based upon
Reasons (1990) model of latent and active failures (Shappell & Wiegmann, 1997a),
encompasses all aspects of human error, including the conditions of operators and
organizational failure. Still, HFACS and any other framework only contribute to an already
burgeoning list of human error taxonomies (see, for example, Appendix C) if it does not prove
useful in the operational setting. In this regard, HACS has recently been employed by the U.S.
Navy, Marine Corps, Army, Air Force, and Coast Guard for use in aviation accident investigation
and analysis. To date, HFACS has been applied to the analysis of human factors data from
approximately 1,000 military aviation accidents. Throughout this process, the reliability and
content validity of HFACS has been repeatedly tested and demonstrated (Shappell &
Wiegmann, 1997b). HFACS has also been implemented by other types of organizations; an
example of its use by the Marine Facilities Division of the California State Lands Commission in
the investigation of incidents at marine terminals is provided in Section 8 and Appendix E.
Given that accident/incident databases can be reliably analyzed using HFACS, the next logical
question is whether anything unique will be identified. Early indications within the military
suggest that the HFACS framework has been instrumental in the identification and analysis of
global human factors safety issues, such as trends in operator proficiency, causes of specific
accident types, and problems such as failures of crew resource management (Shappell &
Wiegmann, 2000). Consequently, the systematic application of HFACS to the analysis of
human factors accident data has afforded the U.S. Navy and Marine Corps (for which the
original classification system was developed) the ability to develop objective, data-driven
intervention strategies.
Additionally, the HFACS framework and the insights gleaned from database analyses have
been used to develop innovative incident investigation methods that have enhanced both the
quantity and quality of the human factors information gathered during incident investigations.
However, not only are safety professionals better suited to examine human error in the field, but
using HFACS, they can now track those areas (the holes in the cheese) responsible for the
incidents as well. Only now is it possible to track the success or failure of specific intervention
programs designed to reduce specific types of human error and subsequent incidents. In so
doing, research investments and safety programs can be either readjusted or reinforced to meet
the changing needs of safety.
Page 40 of 141
Houston, Texas
5.0
PUTTING IT ALL TOGETHER: INVESTIGATING AN INCIDENT FOR HUMAN FACTORS

CAUSES
5.1
Introduction
As stated earlier, in an incident investigation, the investigator starts with the immediate actions
and events surrounding the incident and then works backwards to uncover contributing causes.
Who, where, when, what, and how are all useful questions to get information relevant to the
incident; but asking why is what will help the investigator drill down into the contributing, latent
conditions that need to be identified and resolved in order to avoid similar incidents in the
future. Remember from our introductory discussion that its not just the people you want to
concentrate on, but also the ways in which technology, environment, and organizational factors
influenced human performance.
5.2
Incident Investigation: Going Beyond the Obvious
If we are to learn from an incident, it is very important to go beyond the obvious cause and
ferret out the underlying, contributing causes. Heres an example. During the early hours of a
November morning, the DOLE AMERICA, a Liberian-registered refrigerated cargo vessel,
collided with the Nab Tower, a conspicuously-lit, man-made construction in the eastern
approaches to The Solent off the Isle of Wight (MAIB, 1999). The ship had left her berth in
Portsmouth and was proceeding seaward with the Norwegian captain, a Filipino officer, and a
helmsman on the bridge. The captain was in charge, and he set a course to pass to the east of
the tower. Suddenly, he saw on the starboard bow what he thought was the red portside light of
a vessel at close range, crossing from starboard to port and presenting an imminent risk of
collision. The captain ordered starboard helm before going to the front of the bridge to confirm
what he thought he had seen. He then called the officer to join him, and the officer confirmed
the presence of a red light and reported a second red light to starboard of the first. The captain
then ordered hard to starboard helm. When no further lights were seen ahead, the captain
ordered hard to port helm, still with the intention of passing to the east of the Nab Tower. The
ship struck the tower shortly afterwards.
The immediate cause of the collision was the masters inappropriate and unquestioned helm
order to port (unsafe act decision error). However, the following contributing factors were
important to this casualty:
From his position at the front of the bridge, the captain was unaware of the ships
heading and her exact position in relation to the tower (precondition for unsafe act
substandard practice crew resource mismanagement).
No discussions took place between the captain and the officer concerning the ships
progress (precondition for unsafe act substandard practice crew resource
mismanagement).
Page 41 of 141
Houston, Texas
The captain and the officer failed to work as an effective team, probably due, in part,
to their differing nationality and social backgrounds, and to an autocratic
management style (precondition for unsafe act substandard practice crew
resource mismanagement and organizational factor organizational climate).
With no dedicated lookout to refer to, the captain called the officer to join him at the
front of the bridge, thereby removing his only source of navigational information
(unsafe act decision error).
The ships manager provided no specific instructions to its officers regarding

voluntarily offering relevant information to the captain (organizational factor
organizational climate).
While the immediate cause was the captains poor decision making due to inadequate
information, future avoidance of this type of incident depends on correcting the underlying
unsafe conditions and organizational factors. Changing the autocratic management style that
was in place in this company to one of crew resource management, and training the bridge
team to operate more effectively by empowering the officer to actively contribute to navigational
decisions (particularly relevant to a multi-national crew) are keys to preventing such a casualty.
Had the investigation stopped with the obvious cause, the true precursors to this incident
would have remained hidden, and remedial actions based only on the immediate cause would
have been ineffective.
5.3
A Tool for Investigation: Events and Causal Factors Charting
Before one can begin identifying the human error causes of an incident, one needs a way to
represent how an incident happened. There are a number of tools that can be used to get
varying levels of detail surrounding the events of an incident and what might have contributed to
it. Some of these include timeline analysis, link analysis, barrier analysis, work safety analysis,
human error HAZOP, and human error analysis. Most of these can be used either during a
safety audit (to understand the work conditions and identify risks before an incident occurs) or
during an incident investigation. A good introduction to these methods may be found in Kirwan
(1997).
Another method that is more directly related to understanding the progression and causes of an
incident is Events and Causal Factors Charting (Hill & Byers, 1992a). This method was
originally developed by the National Transportation Safety Board for the analysis of accident
investigations. It highlights the major events in the progression of an incident and also
associates contributing causes to each event. Contributing causes include not only active and
latent human errors, but also equipment problems, weather, and anything else which may have
influenced the events surrounding the incident. Events and causal factor charting can be
helpful in organizing and understanding the sequence of events and also in identifying holes or
inconsistencies in the incident information collected.
Page 42 of 141
Houston, Texas
To illustrate how Events and Causal Factors Charting can be used, lets take a closer look at
the TORREY CANYON incident that was introduced in Section 2 (Hill & Byers, 1992b).
5.3.1
TORREY CANYON Synopsis
The Captain of the TORREY CANYON was experienced, careful, and a stickler for details. The
TORREY CANYON was traveling from Kuwait to Wales with a cargo of 100,000 tons of oil.
They were heading for Angle Bay, British Petroleums (BP) deep-water terminal on the western
tip of Wales. The day before the TORREY CANYON was due to arrive at Angle Bay, the
captain was contacted by BPs agent, who told him of impending decreases in the tide at Milford
Haven, at the entrance to Angle Bay. He was told that if the TORREY CANYON did not catch
high tide on the next evening, it would have to wait outside the harbor for most of a week for the
next tide high enough to get the ship in. Now, to have a ship of that size sitting idle for five days
is very expensive, and the captain was determined to reach Milford Haven on time. This didnt
seem to present any problem at the time; to be ready to catch the high tide the next evening the
TORREY CANYON had to get to Milford Haven and it had to transfer cargo from the midship
tanks to the fore and aft tanks to even out the ships draft. At sea, the tanker drew 52 feet 4
inches amidships, but that was too deep to make it into Angle Bay, so they had to shift cargo.
The captain estimated the transfer would take about four hours and planned to make the
transfer after they reached Milford Haven. Still, there seemed to be plenty of time.
The next morning, the captain asked to be called when the Scilly Islands were sighted. The
Scillies are made up of 48 tiny islands and contains a number of submerged large rocks and
sandbars. There were 257 shipwrecks there between 1679 and 1933. The captain was
intending to sight the Scillies to starboard, pass them to the west and then go into Milford
Haven. However, when the Scillies were sighted, and he was called, they were off the port bow.
Rather than turn and go west around the islands, the captain decided he needed to save time
and would pass between the Scilly Islands and Lands End, the southwesternmost tip of
England. The passage between Lands End and the Scillies is divided into two parts by an
island and each of those parts have further obstructions within them. The captain decided to
take the western channel. He did not have a copy of the Channel Pilot for the region and he
was not particularly familiar with the area. The TORREY CANYON was making full speed when
it met some fishing boats in the channel, which delayed it making a turn. After taking a bearing
from the unfamiliar landmarks, the captain realized that he had overshot his turn and the
channel. When he ordered hard to port, and the helmsman turned the wheel, nothing
happened. The captain realized that the steering selector switch was set incorrectly on
autopilot, reset it to manual, and the turn to port was begun. The TORREY CANYON then ran
into a granite reef so hard that it could not be pulled off. The Royal Air Force eventually
bombed the wreck in an effort to burn some of the oil before it washed up on the beaches.
Page 43 of 141
Houston, Texas
During the official inquiry which followed, it was pointed out that the captain had plenty of time to
get to Milford Haven if he had transferred his cargo while underway8. The chairman of the
board of inquiry reportedly stated after the hearing, He [the captain] didnt want to dirty his
deck, to come into port looking sloppy (Hill & Byers, 1992b). Perrow (1984, p. 184) points out,
that as most accidents do, this accident involves many if only statements:
If only the captain had not forgotten to put the helm on manual, they might have turned
in time;
If only the fishing boats had not been out that day, he could have made his turn earlier;
If only he had prudently slowed down once he saw the fishing boats, he could have
turned more sharply;
once deciding to risk going through the Scilly Islands he used a peculiar passage
through them if only he had used another passage, it might have been safer (even
faster).
Well never know precisely why the captain made the decisions he made.
5.3.2
An Event and Causal Factors Chart of the TORREY CANYON Incident
To do an Event and Causal Factors Chart, we begin by determining the major events that
occurred. Working backwards from the accident, there are four major events:
The TORREY CANYON fails to make its turn in time and runs aground;
The captain takes the western channel between the Scilly Islands and Lands End;
The TORREY CANYON goes east of the Scilly Islands;
The Scilly Islands are sighted to the NW (port) rather than to the NE (starboard).
These four events would be placed in boxes (to denote that they are events) across the top of
the page, as shown below in Figure 14.
This accident occurred years before the Exxon Valdez and the environmental protection legislation that
followed. At the time of the TORREY CANYON incident, transferring oil while underway was standard
operating procedure for many companies.
Page 44 of 141
Houston, Texas
Sights Scilly
Islands to NW
instead of to NE
Goes East of
Scilly Islands
Takes
western
channel
Runs
aground
Unfamiliar
route
Tide schedule
means he has to
hurry
Unaware of
advice in
channel pilot
Fishing boats
in channel
Neglected
currents &
winds
Economic
pressure to
hurry
No copy of
channel pilot
on board
Unfamiliar with
landmarks and
waters
Doesnt want to
dirty decks by cargo
transfer en route
Unfamiliar
route
Failed to
slow down
Steering selector
switch on
autopilot
Figure 14.
Events and Causal Factors Chart for TORREY CANYON

Grounding
(after Hill & Byers, 1992)
Below each event box, we list the perceived causes of the event, depicted as ovals (solid ovals
for established causes, and dotted ovals for assumed causes). We drill down into the causes
by asking why. For example, when we ask why the TORREY CANYON failed to make the turn
in time (thereby running aground), we find four different issues:
The TORREY CANYON failed to begin its turn in time because there were fishing boats
in the way;
The captain was unfamiliar with the waterway and landmarks, making him late to
recognize the turning point and initiate the turn;
He failed to slow down and was going too fast to make the turn;
Page 45 of 141
Houston, Texas
The steering selector switch was in the wrong position (autopilot), and by the time it was
corrected, it was too late to make the turn.
While the Events and Causal Factors Charting often ends at this point, we want to take it a step
further for the purpose of determining active and latent human factors causes. To do this, it is
helpful to remember the human-system approach discussed in Section 2 (refer to Fig. 2).
Human errors result not only from errors made by a given person, but, more importantly, human
errors usually result from suboptimal interactions between people and organizational,
technological, or environmental factors. As you drill down seeking contributing causes, keep
asking yourself the following questions to help you identify human errors:
Did some aspect of the involved person contribute to the incident? (consider
knowledge, skills, experience, motivation, alertness, physical and mental states, use of
medicines or drugs, personal problems, etc.)
Did some aspect of interpersonal interactions contribute to the incident? (consider

communication, definition of roles and responsibilities, experience working as a team,
autocratic vs. empowered style, etc.)
Did some aspect of the interactions between the people and the organization contribute
to the incident?
(consider training and qualification requirements, crew/team
complement, work schedules, safety culture, supervision, policies regarding economic
pressure, etc.)
Did some aspect of the interactions between the people and the technology contribute to
the incident? (consider equipment layout, whether equipment is designed to do the job,
how information is provided from the equipment to the user, whether controls can be
easily operated, whether displays are legible, whether the design obstructs proper
maintenance, etc.)
Did some aspect of the interactions between the people and the environment contribute
to the incident? (consider the workplace environment in terms of lighting, noise,
temperature, vibration, ship motion, fog, snow, etc.; also consider the regulatory and
economic environment and their impact on job behavior)
The human-system approach and HFACS are complementary ways of looking at human errors.
In the human-system approach, we identify the locus of the error. That is, we determine
whether the error resulted because of a deficiency in a given persons actions or decisions, or
whether there was a poor interaction between multiple people or between people and
technology, organization, or environment. HFACS identifies the parts of the companys
organization that had the responsibility for preventing the error. That is, with HFACS we
determine whether the individual (Preconditions for an Unsafe Act), the line supervisor (Unsafe
Supervision), and-or management (Organizational Influences) had the responsibility for
preventing the error. By identifying the level(s) of the organization that had the responsibility for
Page 46 of 141
Houston, Texas
preventing an incident, we identify the part(s) of the organization where changes must be made
to solve the human error problems.
Now let us look again at the Events and Causal Factors Chart of the TORREY CANYON
grounding and use the human-system approach and HFACS to identify the underlying human
factors causes of the last event, Runs aground. To do this, we would continue to ask why. To
ask why there were fishermen in the way doesnt get us anywhere with this particular case9.
Note that the presence of the fishermen is, in fact, an environmental factor that interacts with
the captain; but since neither the captain nor the company have any control over the presence
of fishermen, that aspect of the incident is ignored. The captain did have control over his
response to the presence of the fishing vessels, and that will be discussed in the third causal
factor.
Drilling down on the second causal factor, the fact that the captain was unfamiliar with the
waterway, is more enlightening. The captain was unfamiliar with the route, because it was not
the route he had planned to take. Furthermore, he did not have a copy of the Channel Pilot on
board, and so could not avail himself of helpful information. Why did he not have a copy of the
Channel Pilot, when he had planned to take a nearby route? While we dont have the answer to
this question, it might have involved an error in HFACS Organizational Influences
Organizational Process if the company had not established that ships carry information about all
routes it would transit. Or, perhaps the company had the policy, but the captain had failed to
ensure that the document was onboard and available for use; this would be an example of
Unsafe Supervision Supervisory Violation. As you can readily see, depending on who was
responsible (the company or the captain), the type of corrective action needed would be vastly
different.
Asking why the captain failed to slow down when he saw the fishing vessels (which was
determined by the board of inquiry to be the prudent action to have taken), the probable answer
is that the captain felt pressured to make good time, and that pressure negatively influenced his
judgement. While theres no question that the captains unsafe act (going too fast) was based
on a decision error, we need to continue to ask why. Why did the captain feel such a
compulsion to make good time? As pointed out in the incident synopsis, had the captain
missed the evening high tide at Milford Haven, the TORREY CANYON may have had to sit idly
waiting for several days before the next tide of sufficient depth to allow her to pass. This
spawns several issues for further investigation. Given that the tidal depths at Milford Haven
were known, why did the company elect to send a tanker that could only get through Milford
Haven on certain days (i.e., did the company consider lightening the TORREY CANYONs load
or sending a different vessel(s) that had a draft more compatible with the tides at Milford
9
Although this question is a good one in certain sections of the U.S. where fishing regulations limit
fishing seasons to only a day or two, causing greatly congested waterways on those days. Asking why in
these cases may point out a flaw in the regulations, showing that these regulations need to consider not
only conservation but waterway mobility, as well.
Page 47 of 141
Houston, Texas
Haven? potential HFACS Organizational Influences Resource Management issue)? If the

TORREY CANYON had gotten into Milford Haven too late to make the high tide, how would that
have affected the captain? Did the company have a policy which would have penalized the
captain for missing the tide, thus encouraging him to take risks (HFACS: Organizational
Influences Organizational Climate)? Or, perhaps, was this a self-imposed pressure due to the
captains pride at making schedule and resulting in his taking unnecessary risks (Unsafe
Supervision Supervisory Violation)? There is an obvious tension between the pressure to
make good time and safe navigation. Depending on how it evolved (organizational policy or
supervisory deficiency), the appropriate correction will vary.
Finally, we have the causal factor of the steering selector switch being in the wrong position.
When we ask why, we will find a procedural issue: why was the ship on autopilot in a
hazardous navigating environment? The answer to this question could be a combination of
factors such as: it was put there inadvertently (HFACS: Unsafe Act Skill-based Error); it was
put there intentionally to allow the helmsman to do another task (Unsafe Act Routine
Violation); or perhaps it was left on autopilot too long, due to inadequate bridge team
coordination (Precondition Substandard Practices Crew Resource Mismanagement). Using
the human-system approach, we might ask whether there was a suboptimal interaction between
the people and the technology. Specifically, why didnt the helmsman know that the ship was
on autopilot? The answer is that there was no indication of the steering selector setting at the
helm! The steering selector control was located where it could not be seen when standing at
the helm. Whereas the locus of this error is the interaction between the user and the
technology, note that the responsibility for the error is at an entirely different level. On the one
hand, this is a design error on the part of the equipment manufacturer; however, since the
company most likely had no control over the manufacturers design, this is not a useful avenue
to pursue. But on the other hand, this might also be an error on the part of company
management for selecting and purchasing unsuitable equipment and-or installing it in an
unsuitable manner (HFACS: Organizational Influences Resource Management).
By charting first the primary events that led to the incident, and then by asking why until the
(many) contributing causes are found, one can establish the reasons why an incident happened
and trace the layers of responsibility from the individual(s) to the preconditions, to supervisory
errors, and to unsafe organizational influences. The combination of Events and Causal Factors
Charting, the human-system approach, and the Human Factors Analysis and Classification
System (HFACS) can be a powerful tool for ferreting out the true underlying causes of incidents
and identifying parts of the system which need to be corrected in order to prevent recurrences of
similar problems.
Page 48 of 141
Houston, Texas
5.4
The Art of Incident Investigation
Incident investigation is somewhat of an art. It takes great skill to build a rapport with and
interview people associated with the incident10. It also takes a great deal of knowledge about
the technical and human factors aspects of the incident to ask the right questions and identify
the important issues (this is the reason for having investigation teams no one person is likely
to have all the requisite skills and knowledge). Likewise, there is no one method or best set of
questions to ask that will work for all investigations (Appendix H gives a sample of human
factors questions that could be asked). It takes experience to spot potential issues and to know
what avenues of questioning will be most fruitful in a particular incident. It takes experience and
skill to hunt for underlying causes and to fit the pieces of the puzzle together.
Incident investigation is also not a serial endeavor: it is highly iterative. As Figure 15 depicts,
there is a cyclical process of gathering information, organizing it into the sequence of events
and causes that led up to the incident, and looking for the underlying human factors causes.
Oftentimes after the initial round of interviews, it will become apparent that the sequence of
events has missing pieces, or that some of the information appears contradictory. This
necessitates additional interviews or re-interviews. As you begin to feel you understand the
sequence of events, you may find that you still lack the information to identify the types of
human errors that were made and the underlying causes of these errors. More questioning is
needed. Sometimes an exploration of the human errors will bring to light that the sequence of
events is still incomplete. And so it goes, back and forth, asking questions, organizing data,
finding holes, and asking more questions, until you can finally produce a set of events, causes,
and underlying human errors that hang together and make a sensible explanation for how the
incident evolved. Patience and persistence are two traits of successful investigators!
In this section we have repeatedly stressed that the good investigator keeps asking why? It is
useful to consider when to stop asking why. Asking why is a great tool for identifying underlying
contributors to the errors that caused the incident. But if taken to extremes, it can become
almost absurd. In general, we want to keep asking why as long as the answer is still something
that has practical significance to the incident and is under the companys control to make
changes11. In the TORREY CANYON incident, for example, we elected not to consider why
10
For a discussion of interviewing techniques, as well as other good material on the collection of
evidence, see Center for Chemical Process Safety (in preparation).
11
There are times when we do want to look for causes that are beyond the companys control. Some
incident causes have ramifications for offshore or maritime safety in general and may necessitate
changes to equipment, legislation, or codes of practice. For example, if a regulation appeared to be a
cause of the incident, then that information needs to be brought to the attention of the appropriate
authorities so that the regulation can be revised. In the TORREY CANYON example, the poor design of
the steering selector mechanism, because it could easily cause similar problems for others, should have
been discussed with the manufacturer, to prompt a redesign, and reported in industry publications to
warn others of the hazard. While the main thrust of this paper is to help companies use human factors
incident investigation to improve their own safety, we all have the responsibility to share this knowledge
in order to improve the safety of the industry as a whole.
Page 49 of 141
Houston, Texas
there were so many fishing vessels in the channel, because that was totally out of the control of
the company. But we did elect to consider the captains response to those fishing vessels (i.e.,
the fact that he did not slow down), because his response is within his (and the companys)
control. There will always be extenuating circumstances. We need to focus on how the
company responds to those circumstances in a way that gets the job done safely.
Page 50 of 141
Houston, Texas
Begin Investigation
Events & Causal
Factors Charting
how?
wh
o?
1. Ask
Questions
en
wh
at?
wh
wh
er e
2. How Did
Incident Evolve?
why?
why?
why?
3. Identify Human Errors

Locus:
Responsibility:
Human System Approach
HFACS
No
Figure 15.
Is
incident
scenario complete?
Does it make
sense?
Yes
Investigation
Complete
The Incident Investigation Process
Page 51 of 141
Houston, Texas
6.0
ANALYZING INCIDENT DATA
Now that youve been running your incident investigation program for awhile, what do you do
with all those data? There are many ways to make good use of the data collected from incident
investigations. Here are a few ideas.
6.1
Learning from Individual Incidents
This is the simplest way to learn from your incident investigation program. Individual incidents
can be discussed at safety meetings, allowing personnel to gain an understanding of how
different incidents evolved and how an accident was avoided. This type of information sharing
can stimulate discussion of similar occurrences and potential changes to procedures, policies,
training, equipment usage, etc. that might help prevent future incidents.
While case-by-case studies can be beneficial, they have a drawback. By focusing on a single
incident, there is no way to know what facets of that incident may represent general problems
as opposed to things that were unique to that particular incident. Most companies would rather
spend their money fixing frequently-occurring problems than smaller, once-in-a-lifetime
problems. The way to get a feeling for the importance and the frequency of a problem is
through data analysis, and several approaches to analysis are discussed below.
6.2
Identifying High-Risk Activities or Facilities Using Simple Frequency Analysis
Frequency analysis can be an effective way to identify problem areas on which you need to
focus. For example, say an offshore drilling company has three rigs and wants to know whether
all three have about the same number of incidents. A simple frequency analysis entails adding
up the number of incidents reported over a given period of time (e.g., one year) by the crews of
each rig (Fig. 16).
Number of Incidents Reported by Rig
Number of Incidents
Reported
30
25
20
15
10
5
0
A
Drilling Rigs
Figure 16.
Example of a Frequency Analysis
Page 52 of 141
Houston, Texas
In the example shown in Figure 16, it is obvious that Rig C has reported over twice as many
incidents as has either Rig A or Rig B. Looks like Rig C has a problem, right? Well, not
necessarily. You need to go deeper and find out why Rig C has reported more incidents.
Maybe the crew members on Rig C are very excited about the new incident investigation
program and are being vigilant for and reporting every incident that arises, while the crews on
Rigs A and B are only reporting the incidents they consider to be important in some respect.
Or maybe Rig C is larger, does more operations, and has more people working on it than either
A or B if Rig C is doing four times the work that is done by either Rig A or B, then Cs true
incident rate would actually be lower than that on A or B! This shows the necessity for
considering how to make an apples to apples comparison. However, if Rigs A, B, and C are
roughly equivalent in all respects, then C might truly have a safety problem that needs to be
identified and solved. One way to look into this in more depth would be to do more frequency
analyses by type of operation, or by type of equipment used, or by some other relevant factor.
In this way you can isolate which operations or equipment appear to be related to the higher
incident rates (for example, perhaps when looking just at Operation X, Rig Cs incident rate is
the same as that for A and B; but when looking just at Operation Y, Rig Cs incident rate is
much higher than that for A or B).
6.3
Looking for Trends
A simple extension of the frequency analysis discussed above is to compare frequencies over
time to look for trends in the data. Perhaps you have made some changes to a standard
operating procedure to reduce injuries. Is the new SOP helping? To find out, you could plot the
number of injuries in the years prior to the new SOP and compare that to the number of injuries
since the new SOP was put into place. Or maybe the SOP has been helpful in reducing certain
types of injuries but not others. A plot, like that presented in Figure 17, could help the company
spot areas of concern. In this example, most of the injury rates are fairly consistent over the
four years shown. However, three injury categories show some interesting changes. Struck
By/Against and Slips/Trips/Falls both show marked decreases. In this particular case, the
company had built and put into service 18 new ships in 1996-1999. The learning curve for
operating the new ships may have contributed to exaggerated Struck By/Against rates in 1998
and 1999, with the decrease in 2000 showing that the crews had become familiar with the new
ships. The decreased rate of slips, trips, and falls is attributed to the companys purchase of
new safety shoes (designed for the restaurant industry to keep traction on wet floors) a
successful safety intervention! The third item of interest is the relatively greater rate of injuries
in the Chemical Spray category over the last two years. This appears to be due to the fact
Page 53 of 141
Houston, Texas
Types of Injuries
1Q98 - 3Q01
40
34
35
30
28
24
25
21
19
20
19
15
15
12
11
10
10
7
55
44
5
1
22
11
2
0
Sp
ra
y
ns
m
ic
al
he
In
ha
le
t
C
c
Ex
pl
os
io
ur
re
n
r
th
e
O
00
El
ec
tri
98
hi
n
Pu
s
g/
P
ul
lin
ns
io
n
O
ve
re
xt
e
Li
fti
ng
w
Sw
al
lo
Te
m
/A
bs
or
b/
Ex
tr e
m
ee
n
w
au
gh
t
Be
t
s/
Tr
ip
s/
Sl
ip
St
ru
ck
By
/A
ga
in
s
Fa
lls
99
01
Figure 17.
Example of a Trend Analysis
(Data courtesy of Stolt-Nielsen Transportation Group)
that new crews were hired to man some of the newer ships, and these crew members did not
have previous experience on a chemical tanker. Stolt is addressing this problem with an
intensive training course taught by experienced Captains and Chief Mates, a course which has
proved very successful in the past. This type of finding alerts a company to the need for getting
additional data to understand the contributing causes to such incidents so that productive safety
interventions can be designed and implemented.
Frequency and trend analysis are often supplemented with cost data to help a company decide
where to focus its next safety intervention. While frequency data tell you how often a given type
of incident happens, it does not tell you the severity or consequence of the incident. Therefore,
it is helpful to combine frequency data with cost (or some other measure of severity) to
determine which types of incidents are most in need of controls. For example, Figure 17 shows
a much higher frequency of slips, trips, and falls than injuries associated with temperature
extremes. Going solely by the frequency data, one would assume the slips, trips, and falls are
more important to control. However, if most of the slips, trips, and falls have a relatively minor
consequence (that is, the injury caused is mild), then it may not warrant much attention. Lets
say that on the average, the injury caused to the workers in these slips, trips, and falls is so
minor that nothing more than first aid is needed and the employee can return immediately to the
Page 54 of 141
Houston, Texas
job. However, lets say that the result of the average injury due to extreme temperatures (e.g.,
heat stroke or frostbite) requires hospitalization and one week of lost time. So even though the
frequency of the temperature-related incidents is only about a quarter of that due to slips, trips,
and falls, their severity is much worse. The total cost (frequency x average cost) of the
temperature-related incidents is much greater than that for the slips, trips, and falls, making the
reduction of temperature-related incidents a higher priority for the company.
A word of warning about trend analysis: just because the data appear to show a trend does not
necessarily mean there really is a significant trend present. Data can be highly variable (that is,
the number of incidents can fluctuate greatly from one time period to the next). It is not
uncommon to see rather large changes in the numbers of incidents from year to year. When
you are dealing with a small number of data points (e.g., comparing yearly incident frequencies
from 1999 to 2000), you cannot see the underlying variability. One way to get a better
appreciation for the variability is to look at the data by month or by quarter instead of by year.
The most accurate way to identify true differences in frequencies and establish real trends is by
using statistical analysis (for example, fitting the data to a linear function (linear regression) and
determining whether the slope is significantly different from zero).
Both frequency analysis and looking for trends are ways to identify operational variables that
may require closer examination. Notice that while the frequency or trend analysis will show you
areas of concern, it does not answer the question as to why these differences are occurring.
Once youve isolated the types of activities, operations, or situations of interest, you can use the
next analysis technique to explore your incidents further.
6.4
Looking for Similar Incidents
If each incident report is coded with the type of incident (e.g., oil spill), and the activity during
which it happened (e.g., filling a tank), one can search the database to identify all the incidents
which had these features in common. One immediate advantage to this is it helps to identify
your high-frequency events. Another big advantage is that you can now re-read the incident
narratives and look for other similarities that might lead to the identification of a safety hazard
that needs to be fixed. For example, in the case of spills caused by overfilling a tank, it may be
that an overfill alarm needs to be added, or that the standard operating procedure needs to be
changed so that the tank is continuously monitored by a crew member, and-or the fill rate
decreased as the tank gets close to being filled. This type of analysis is an excellent way to
identify equipment design flaws and poor operating procedures. The FAA-sponsored, NASArun Aviation Safety Reporting System, has used this method successfully to identify equipment
defects, runway design problems, and to make improvements to air traffic control protocols.
6.5
Determining Under What Conditions a Given Error Happens
In the construction of the incident database, it can be useful to do a risk assessment of various
operations to identify things that might go wrong. The database can then be used to see
under what conditions things actually do go wrong. For example, in a study of communications
Page 55 of 141
Houston, Texas
errors, the types of operations that depend on good communications were identified (e.g.,
vessel navigation with a marine pilot on board requires good communications between the pilot
and the ships bridge team; safe meeting and passing agreements depend on good ship-to-ship
communications). The potential need for communication was then tagged in the database by
answering five simple questions, such as Was there a pilot navigating? or Were there two or
more vessels involved in this casualty?. In analyzing the accident data, it was found that of the
accidents in which one of these five questions was answered yes, 76% of them had a
communications error as a contributing factor to the accident. This is a powerful way to identify
high-risk activities and situations.
6.6
Looking for Underlying Causes Using Meta-Analysis
If your incident investigators are adept at asking why enough to get to underlying human error
causes, then you can use your incident database to determine which types of causes may
precipitate many of your incidents. To extend the example above of communications-related
casualties, the incident investigators used a form to identify specific communications problems,
such as did not communicate, did not send information in a timely manner, message was
interrupted, did not interpret the information correctly, and others. For each problem
identified, the investigator went on to consider a list of contributing factors, such as inadequate
knowledge of company policies for communications, limited English skills or knowledge, did
not operate communications equipment correctly, distracted or interrupted by other tasks,
assumed there was no need to communicate, and others (see App. G for the complete
communications investigation protocol).
A frequency analysis of the communication problems showed that the single biggest problem
was a failure to communicate. That is, in 68% of the accidents, someone had information that
could have prevented the accident, but chose not to tell anyone. These failure to
communicate casualties were isolated and a frequency analysis was done to identify the most
frequent contributing factors. While this identified factors such as incorrect interpretation of the
situation, assumed incorrectly that other party already knows, and others, it didnt give a good
sense of what might be at the crux of these accidents.
To get a clearer picture of what was going on, a meta-analysis was done in which the
narratives of the different casualties were reviewed and additional characteristics of the
situations were identified. The result was the finding that the most common apparent underlying
cause in 92% of these failure to communicate accidents was that the person did not perceive
a safety threat, either because he had misinterpreted the situation or because he failed to think
about the ramifications of the situation beyond his own specific job responsibilities (that is, he
did not consider how his actions might affect other people). These types of behaviors show a
deficit in situation awareness. The meta-analysis also showed that in almost half of these
accidents, there was a second person who did not speak up. This person perceived the safety
threat, but assumed it was not his job to say anything (he assumed someone else was aware of
Page 56 of 141
Houston, Texas
the problem and would take care of it). This failing shows a lack of ship resource
management. The meta-analysis allows one to go beyond the specific data categories in the
incident database and to find underlying causes that may tie other, seemingly-disparate causes
together. In a sense, a meta-analysis is like putting a puzzle together. Each database element
is a piece of the puzzle, but the meta-analysis helps us see how to put the pieces together and
get greater meaning from them. In this case, it was the meta-analysis which most effectively
pointed us at the true underlying problems, suggesting the types of interventions (improvements
in situation awareness and ship resource management) that would be productive.
6.7
Identifying Relationships Among Incident Attributes by Statistical Analysis
Statistics can be used to draw out meaningful relationships among elements of incidents, and
sometimes they can be used to infer probable cause. Statistical analysis was used in a couple
ways in a recent study of fatigue-related accidents (McCallum, Raby, & Rothblum, 1996). The
purpose of the research was to understand not only how many marine accidents were related to
fatigue, but also to look for underlying contributing factors (in other words, what was causing the
fatigue). A database was established using scientific literature to identify the questions that
should be asked (the resulting fatigue investigation questions can be found in App. G). The
fatigue investigation was administered during routine casualty investigations, and the database
was used to find out what attributes were significantly related to fatigue (that is, what data items
had statistically different values in the fatigue-related accidents compared to accidents that did
not result from fatigue). One set of tests looked at the number of hours worked by mariners who
caused injuries (either to themselves or to another crew member).
The statistical tests (t-tests) determined that the number of hours on duty at the time of the
accident, and the number of hours worked in the last 24/48/72 hours were all significantly
different for the fatigue-related and non-fatigue injury cases. The averages for each of these
comparisons are plotted below (Fig. 18). One must be careful about jumping to causal
conclusions. In some cases, the difference seen between two groups may be due to something
very different than what is being tested (remember our example of the number of incidents
reported by the different oil rig crews). However, in this case, the scientific literature supports
the relationship between long work hours and increasing fatigue. Therefore, these data were
taken as strong evidence that an underlying cause of these fatigue-related injuries was long
work hours.
Statistical analysis was used a second way in this study. It was used to consider all the
different factors that were correlated with fatigue-related accidents and to come up with a quick
screening test for fatigue. Because the full fatigue investigation took about 40 minutes, it was
desirable to find a few questions that would indicate whether fatigue appeared to play a role,
and whether, therefore, the investigator should collect all the fatigue data. A multiple regression
analysis was performed to determine which factors were most predictive of fatigue-related
casualties. The result was a simple Fatigue Index equation consisting of just three questions
(the number of hours slept in the last 24 hours, the number of hours worked in the last 24 hr.,
40
34.9
urs
35
Fatigue
30
Page 57 of 141
Houston, Texas
Figure 18.
Result of Statistical Analysis of Incident Database Items

(from McCallum, Raby, & Rothblum, 1996)
and the number of fatigue symptoms experienced by the mariner). When the data from both
the fatigue and non-fatigue casualties were put into this equation, it was found to be 80%
correct in its ability to identify whether a given casualty was fatigue-related or not (the fatigue
index worksheet is in App. G).
The purpose of statistical analysis is to determine whether apparent differences in data points
are just due to the variability of the data or whether they are due to a true underlying difference
or trend. For example, in Figure 17, the incidence of caught between injuries varies greatly
over the four years shown. Is the downward trend from 1999 through 2001 real, or is it just a
matter of the normal variability of the data? Statistical tests can be used to determine this. It
should be noted that just because something turns out to be statistically significant, it does not
necessarily mean that it is important or significant in practical terms. The size of the
significant difference could be very small. For example, Figure 16 compared the number of
incidents on three drilling rigs. It might be that the incident rate on Rig B is significantly lower
than the rate on Rig A. But since both have a relatively low rate of incidents, from a practical
standpoint, the difference just isnt interesting (not worthy of taking action).
6.8
After the Analysis
Websters Dictionary (G & C Merriam Company, 1973) defines data as, factual information (as
measurements or statistics) used as a basis for reasoning, discussion, or calculation. The
operative word is basis. The analysis of incident data serves as nothing more than a basis or a
point of departure for discussion, reasoning, and perhaps additional study. It is not until the
data have been pondered, organized, understood, and put into proper context that these bits of
Page 58 of 141
Houston, Texas
disparate facts turn into valuable and useful information. This subsection discusses two areas
in which one needs to go beyond the data analysis in order to fully understand the data and take
appropriate action.
6.8.1
Data-Driven Research: the Link Between Data Analysis and Solving Safety Problems
You cannot solve safety problems simply by analyzing incident data. In general, the analysis of
incident data will identify a potential problem that then needs to be studied before it can be
solved. For example, if you plot your companys injury data and find that a large percentage of
the injuries are from slips, trips, and falls, you have identified a problem, but not a solution. The
next step would be to investigate further and understand what seems to be causing the slips,
trips, and falls. Where do these injuries occur and under what conditions? Are deck surfaces
slippery or are stairway treads worn? Have personnel been provided with proper footwear, and
if so, are they wearing it? (Youd be surprised at the number of times companies provide
personnel protection gear of one sort or another, only to find that employees refuse to use it
because its either uncomfortable or interferes in some way with other aspects of their jobs.
Protective gear must be designed to be compatible with the workers needs and workplace
tasks.)
As was discussed earlier (see Fig. 8 in Sec. 3), the analysis of incident data is the precursor to
data-driven research used to understand the problems identified by the analysis. Research
may be as simple and low-tech as a discussion with employees and line supervisors to get their
perceptions of the problems and potential solutions, or it can be as detailed and intensive as a
full-blown scientific study. The point is that the analysis of the incident data is a starting point,
and that it takes follow-up study to understand the genesis of a problem and to devise
successful safety interventions.
6.8.2
It May Be Data, But Its Not Necessarily Telling You Anything
One final caution: a database is only as good as the data that are put into it. If the investigator
doesnt ask all the relevant questions, the database cannot, by definition, have the relevant
data. This harks back to the recommendations given in Section 3 for building a successful
incident investigation program. If the company does not promote an open, fair, and
improvement-oriented culture, or if there isnt a common understanding about the scope and
purpose of the incident investigation, or if the investigators are not appropriately trained, or if the
incident database is hard to use, the data that populate the incident database may be less than
accurate and complete. Obviously, analyses based on such data will be of questionable value
(garbage in, garbage out).
Even with the best of intentions, things may happen which affect the database. For example, a
simple change in policy affecting which incidents will be investigated may result in the
appearance of a greatly increased (or reduced) incident rate when comparing data from periods
before and after the policy went into effect. Lets say a company decides to forego incident
investigation on any incident which costs the company less than $10,000. If there are types of
incidents which are predominantly low-cost, the frequency of those incidents will appear to be
dramatically reduced after the policy takes effect (even though the true frequency of the
Page 59 of 141
Houston, Texas
incidents has not changed, or even increased they just arent being investigated anymore).
Training for investigators may result in a better understanding of the classification scheme (such
as HFACS) they are using. While this should result in improved data reliability in the future, it
may also give the appearance of changes in certain types of incident rates (because some
incidents may have been misclassified prior to the training).
Never put blind faith in your incident data analysis: always be on the lookout for procedural or
other reasons (unrelated to actual incident rates) that might be affecting the analytic process.
Keep records of changes made to the database, investigation policy, and investigator training
these could be great time-savers in understanding mysterious trends. Keeping records of
important company policy or procedural changes can also be helpful in understanding changes
in incident frequencies. The more you know about changes in the way you do business both
in the company at large and in the incident investigation program specifically the better you
will be able to differentiate between spurious trends and true safety issues.
6.9
Summary of Analysis Techniques
Data analysis can be used to identify areas in need of safety interventions. Oftentimes, data
analysis shows an interesting trend, but does not give you sufficient information to take action.
This will require follow-up studies to better define the problems and suggest workable solutions.
Data analysis can also point out where the database and-or investigation procedures are
lacking. For example, you may find out the company has a high rate of slips, trips, and falls
but that doesnt tell you enough about the problem. The database might need to be modified to
add information on the types of slips, trips, and falls (e.g., where they occur, what operations
were in progress, how much lost time resulted), and the incident investigators may need to ask
additional questions to illuminate the causes of these accidents. Recall that Figure 8 shows a
feedback loop from Database Analysis back to Incident Investigation. Data analysis is a great
way to learn about the strengths and weaknesses of your investigation methods and database.
This section has provided examples of ways you can learn from your incident data. Frequency
analysis and looking for trends are simple procedures that anyone can do quickly with the aid of
a spreadsheet application. Looking for similar incidents and determining conditions which tend
to be associated with them are fairly simple procedures, although a database with relevant
index variables is helpful for doing such analyses efficiently. Statistical analysis, while requiring
more expertise on the part of the safety analyst, can provide great benefits by finding underlying
correlations and relationships. The important thing is not to let your incident data just sit there:
analyze it and make it work for you.
Page 60 of 141
Houston, Texas
7.0
FINDING SAFETY SOLUTIONS
Lets say youve collected incident information and run some analyses. Now you want to
develop measures to prevent these incidents from recurring. What do you do? How do you find
solutions that will be effective? Oftentimes, finding effective solutions is elusive. A study of
offshore operations in an international oil and gas company demonstrates what can happen
(Bryden, OConnor, & Flin, 1998). This company had an incident investigation program. The
database contained information on technical and human factors causes of the incidents and
suggested remedial actions. Analysis of the recommended remedial actions showed that only
10% of them addressed the underlying causes of the incidents, while another 31% addressed
only direct causes and no underlying causes. The shocker was that 59% of the recommended
remedial actions were quick fixes which did not address the causes of the incidents at all! The
ineffective quick fixes tended to be things such as telling the worker not to do it again, or
mentioning the danger at the next safety meeting. A safety program based on trying to motivate
the worker not to repeat a dangerous action, without taking steps to solve the underlying
causes, is doomed to failure.
On the other hand, it is not so surprising that companies might fall into such a quick fix trap.
One might say that the biggest problem with having a successful incident investigation program
is that now there are data about which management must not only think, but also do something
constructive! Perrow put it this way (Perrow, 1986, as quoted by Hollnagel, 2000, p.1):
Formal accident investigations usually start with an assumption that the operator
must have failed, and if this attribution can be made, that is the end of serious
inquiry. Finding that faulty designs were responsible would entail enormous
shutdown and retrofitting costs; finding that management was responsible would
threaten those in charge; but finding that operators were responsible preserves
the system, with some soporific injunctions about better training.
Remember Reasons (1990) Swiss cheese model each slice of cheese (excluding unsafe
acts) represents a layer of system defenses. The fact that incidents are occurring means that
one or more of these layers of system defenses requires repair: they are not effective barriers
to prevent unsafe outcomes. Hollnagel (2000), like Reason, has suggested that in order to
prevent incidents, we must go beyond finding a single root cause (or making the operator the
scapegoat, per Perrow) and understand how to improve the barriers (system defenses).
Barriers can either avert an incident from taking place, or reduce the magnitude of the negative
consequences (prevention and mitigation, respectively, as depicted in Fig. 8).
7.1
The Triangle of Effectiveness: A Guide to Safety Interventions
Gerry Miller (2000; Miller et al., 1997) adds his voice to those of Reason and Hollnagel in
decrying the past tendency to place the blame for industrial incidents solely on operator error.
Instead, it is his contention that even the most safety-conscious employee will occasionally
Page 61 of 141
Houston, Texas
initiate unsafe acts at the job site, and that sometimes these acts are encouraged, led, or even
coerced upon the employee by a variety of factors beyond the employees control. However,
Miller states that these acts can be prevented, or at least the consequences of the acts
mitigated, through the application of barriers or safety interventions. He illustrates this concept
through his triangle of effectiveness (see Fig. 19), which presents eight levels of barriers that
can be used to prevent or mitigate incidents12. Starting at the base of the triangle, these eight
elements are:
Policies and culture13 management policies and corporate culture which promote a
safe, human-centered work environment;
Workplace design ergonomically-designed and arranged equipment;
Environmental control keeping lighting, temperature, noise, etc. within humancompatible ranges;
Personnel selection selecting the right people for the job;
Training and standard operating procedures (SOPs) ensuring workers have the
necessary knowledge and skills to do the job, and that SOPs are correct and
consistent with best practices;
Interpersonal relationships (communication)

information between team members;
Job aids understandable, easy-to-use task instructions and warning placards;
Fitness for duty ensuring that workers are alert, focused, and capable of safe job
performance.
the
exchange
of
necessary
All eight barriers are important, Miller concludes, and must be included in a total behaviorallybased safety program. It should be emphasized, however, that the elements at the base of the
triangle (i.e., policies & culture, workplace design, and environmental control) have the most
significant impact on safety and should form the backbone of a companys safety program.
(Each of these eight barriers will be discussed in more detail in the next section)
In Reasons model, these elements (at the base of the triangle) are controlled by the
Organizational Factors layer of defenses. When the organization (company management)
makes poor decisions, such as the selection of equipment which is not designed to support the
12
Miller uses this triangle both as a model for accident causation and as a guide to selecting safety
interventions. Like Reasons framework, Millers emphasizes the multiplicity of causes of a given
incident and attributes the causes to the lack or failure of barriers (system defenses in Reasons jargon).
In this paper we have chosen to focus on Millers triangle as a means for selecting interventions, since it
bridges the gap between Reasons organizational model of system defenses (management line
supervisor worker) and the concrete needs of a shipping or offshore company to select specific means
to solve identified safety problems.
13
Miller calls this factor Management Participation. Workshop participants felt Policies and Culture
was a more intuitive label.
Page 62 of 141
Houston, Texas
human operator, that single poor decision has an enormous trickle down effect because so
many operators and operations are affected. Such poor decisions at the Organizational Factors
layer very often become latent contributors to incidents. In a similar fashion, good decisions
made at the Organizational Factors layer, such as the selection of well-designed equipment (or
other human-centered decisions contained at the base of the triangle), contribute very positively
to the safety program, again because of the numbers of people and operations they touch.
Interventions based solely on elements at the top of the triangle (such as fitness for duty and job
aids) will have the least impact on workplace safety, and therefore should have a lesser
emphasis within the companys safety program. The factors at the top of the triangle depend
primarily on the actions of individual workers. Interventions at this level are on a one-by-one
basis a less efficient and less effective way of dealing with safety issues.
Fitness
for Duty
Job Aids
Interpersonal
Interaction
Training/SOPs
Personnel Selection
Environmental Control
Workplace Design
Policies & Culture
Figure 19.
The Triangle of Effectiveness for
Safety Interventions to Reduce Human Error
(after Miller, 2000)
Page 63 of 141
Houston, Texas
An example will help to clarify this. Lets say Joe needed emergency medical attention because
he accidentally sheared off one of his fingertips while cutting metal sheeting to make a repair.
Telling Joe to be more careful will not likely have a big impact on safety. Training all the repair
crews on the correct procedure for cutting sheet metal will have more of an impact, since more
of the workforce is made aware of the problem and a way to protect themselves (assuming, of
course, that supervision and peer pressure encourage and reinforce their behavioral changes).
But the best way to prevent this type of incident is by having equipment that has been designed
with a guard to prevent ones fingers from contacting the cutting mechanism (workplace
design level).
Hollnagel (2000) points out that the purpose of an incident investigation program is to identify
barriers (system defenses) that have failed or barriers that were missing which allowed an
incident to happen. A good safety intervention program repairs and-or develops as many of
these barriers as possible. Millers (2000) addition to this line of thinking is that when it is not
possible to implement all the relevant barriers, selecting those towards the base of the triangle
will reap better protection than selecting only those towards the tip. Just as we need to probe
deeper to find the underlying latent factors which cause incidents, we also want to make safety
fixes and focus our safety program at the deepest levels possible (at the base of the triangle).
7.2
Relationship Between Reasons Slices of Cheese and the Triangle of

Effectiveness
As shown in Figure 20, different layers of system defenses are related to different elements
within the triangle of effectiveness. The Organizational Factors layer has the greatest span of
control, and therefore, the greatest capacity for effective intervention. Remember that
organizational factors refers to the policies, procedures, and decisions put into place by upper
management. Management is usually responsible for designing the procedures and developing
the work policies implemented by the line supervisors and workers. As such, this layer of
system defenses can influence seven of the eight elements within the triangle of effectiveness,
and is the only layer of defense which can effectively impact the most important lower three
elements (see top of Fig. 20).
The Supervision layer of defenses represents the interventions that can be controlled by line
management. Note that whereas the organizational factors layer is generally in charge of
designing and developing policies and procedures, supervisors are responsible for carrying out
those policies and procedures. This automatically limits the effectiveness that supervisors can
exert, since they often cannot change existing policies and procedures, only report back on
those which may appear to be latent factors in incidents. The middle section of Figure 20
displays the types of interventions to which supervisors can contribute. While they are not
always directly involved in hiring and firing, line supervisors generally are involved in personnel
selection from the standpoint of assigning people to tasks. In a similar vein, while they might
not be involved in training or writing standard operating procedures (SOPs), they are
responsible for seeing that these are properly carried out. They may also be the ones who
recommend workers for remedial or advanced training.
Page 64 of 141
Houston, Texas
The worker has the smallest span of control over safety interventions. The workers level is
basically contained within the layer of system defense called Preconditions for Unsafe Acts. It
becomes the workers responsibility to adhere to standard operating procedures, learn to use
equipment properly, communicate clearly, use job aids when needed, and to stay fit for duty. If,
for example, a standard operating procedure is deficient or a management work-rest schedule
causes excessive fatigue on the job, the workers span of control is too limited to allow for
meaningful intervention at that level. This is why it is so important to look for interventions at
the base of the triangle, at the organizational factors layer of system defense.
Now lets discuss each the intervention elements within the triangle of effectiveness and see
how they relate to the layers of system defenses.
7.2.1
Policies & Culture
Management policies and corporate culture depend on the active participation of upper
management in promoting a human-centered work environment and worksite. As such, policies
and culture are key to an effective error reduction program: they are the base on which
everything else rests. Management participation should be demonstrated in a variety of ways.
It should be visible in its support and active encouragement of an open, safety first corporate
culture, where safety first is not just a motto but a corporate mission. An atmosphere that
provides incentives for personnel to question and improve work environments and standard
operating procedures shows a caring management philosophy. Management actions and
decisions should be human-centered, enabling the best personnel performance. Examples of
Page 65 of 141
Houston, Texas
Fitness
for Duty
Job Aids
Organizational
Factors
Interpersonal
Interaction
Training/SOPs
Personnel Selection
Workplace Design
Policies & Culture
Fitness
for Duty
Supervision
Job Aids
Interpersonal
Interaction
Training/SOPs
Personnel Selection
Workplace Design
Policies & Culture
Fitness
for Duty
Preconditions
for Unsafe Acts
Job Aids
Interpersonal
Interaction
Training/SOPs
Personnel Selection
Workplace Design
Policies & Culture
Figure 20.
Types of Interventions Needed to Plug the Holes in the Layers

of Defenses.
Different layers of system defenses (slices of cheese at left) can be
bolstered by and have the power to implement different sets of
interventions (unshaded elements in the triangles).
Page 66 of 141
Houston, Texas
how management can demonstrate its commitment to a good employee safety program include
establishing safe work loads and schedules based on known human physiological limitations
and requirements (e.g., work-rest cycles); the creation of easily-understood and achievable
company policies and practices; the establishment and consistent application of rewards (or
punishments) for compliance (or lack thereof) with company policies; creation of reasonable
product delivery schedules; providing physical facilities and equipments designed to match
human capabilities and limitations; insistence on good facility maintenance; and a commitment
to uncovering the underlying causes of incidents. By definition, management policies and
corporate culture are a part of the Organizational Factors level of defenses. Whenever such
organizational factors (like the examples just given) are discovered during an incident
investigation, management policies and culture should be considered when designing
interventions to stop future such incidents.
7.2.2
Workplace Design
Good, ergonomic workplace design can be extremely effective in reducing incidents. As

underscored earlier in this paper, poor design of vessels, platforms, equipment, and work
environments is an underlying precondition that can set up the human operator to make
errors. In fact, the American Bureau of Shipping states that 88% of shipboard injuries and 50%
of fatalities are the result of poor design (McCafferty, 2000). Unfortunately, examples of poor
design abound in the maritime industry. As Miller says (2000, p. 7), You cannot overcome
human errors induced by poor design of the work place with more training, more manuals or
written procedures, exhortations to work safer, or threats of punitive actions for job accidents.
Good design is not a mystery: there have been decades of human factors research into a
plethora of design facets. Further, there has been considerable experience acquired over the
past decade in the offshore and shipping industries with applying HFE design criteria to new
and remodeled facilities. In addition, well-established and accepted human factors design
standards and guidance now exist for the maritime industry (ASTM, 1995; American Bureau of
Shipping [ABS], 1998; ABS, 2001). In fact, when HFE design principles are introduced at the
beginning of the design of a facility or system, it can drastically reduce life-cycle costs and
prevent the need for costly modifications down the road (because it was built right the first time).
Good workplace design requires proactive management involvement; it is almost always
beyond the span of control of the workers or line supervision. Thus, incidents resulting from
workplace design flaws, even though some may be considered equipment preconditions, must
look to the Organizational Factors layer of the organization for effective safety interventions.
7.2.3
Lack of environmental control is another widespread, latent cause of maritime and offshore
incidents. This element of the triangle refers to the work environment: temperature, humidity,
Page 67 of 141
Houston, Texas
lighting, noise, and vibration. Human beings have safe operating ranges just like equipment.
Put a human in an environment that is outside his safe operating range, and he becomes an
accident waiting to happen. Miller (2000, p.8) relates an all-too-frequent occurrence:
As just one example, studies have shown that crane accidents are the second
most frequent cause of injuries and fatalities on offshore platforms in the Gulf of
Mexico (GOM). Yet, operator cabs on these cranes are traditionally not
environmentally controlled. Some years ago during a visit to an offshore platform
it was noted ... [that the temperature inside a crane cab was] 122 degrees F.
With this combination of heat and July GOM humidity is it any wonder that an
operator error of omission (i.e. the operator did not complete an act that he was
suppose to have done) occurred that day resulting in damage to the crane and
platform.
Emphasizing the importance of proper workplace environments, the American Bureau of
Shipping (ABS) has prepared human performance-based environmental standards for both
ships and offshore structures. These standards will be issued in 2002.
7.2.4
Personnel Selection
It has long been recognized that certain jobs require special physical, mental, or social skills not
possessed by everybody who would like to work in those jobs. Finding the right people for the
job is what personnel selection is all about. As an example, special physical and psychological
screening tests have been used for at least the last thirty years in the public safety sector to
screen out those who would not be suitable for the law enforcement or fire safety professions. If
a person is selected to work in a job for which he/she is not suited, that can result in an
increased probability of that person contributing to a workplace accident.
Personnel selection should consider the personality traits and special abilities needed for a
given job. Just because someone has the desire to work in a particular job, or even has spent
twenty years working elsewhere in the company, that doesnt necessarily qualify that person for
the open position. A given job may have certain physical requirements (e.g., good color vision
for an electrician, or good visual acuity and depth perception for a crane operator), as well as
intellectual aptitudes (e.g., good communication skills for a supervisor) and psychological
requirements (e.g., good judgement and coolness under stress for a ship master or offshore
installation manager).
Page 68 of 141
Houston, Texas
As a barrier to incidents, personnel selection is active at two different levels. The

Organizational Factors level (upper management) usually presides over personnel selection in
terms of hiring and firing by developing job descriptions and qualifications requirements. A
company which takes the quality of its personnel seriously can do much to ensure the right
people are placed in the right jobs. The Supervision layer of system defenses also contributes
significantly in terms of the way personnel are assigned to tasks. Again, if the supervisor takes
this job seriously, the right numbers of people, fit for duty, and with the right qualifications can
be assigned to operational and maintenance duties, helping to ensure safety.
7.2.5
Training and Standard Operating Procedures (SOPs)
Training and SOPs make up the next element in the triangle14. More training has too often
been managements sole, knee-jerk remedy to incidents. As already stated, more training can
not make up for inattention given to other barriers such as workplace design, environmental
control, and management policies. Another frequent limitation to training is that it is assumed
that someone who is experienced at a given task is therefore qualified to train others. Such
training is haphazard, often poorly performed, and usually incomplete. There are rigorous
methods (such as Instructional Systems Development) for analyzing tasks, determining
performance objectives, and training and testing to these objectives (McCallum, Forsythe,
Smith, Nunnenkamp & Sandberg, 2001). As automated systems become more prevalent in the
maritime and offshore industries, thorough training, not just in the task at hand, but also in the
operational parameters of the equipment, becomes increasingly necessary (Sanquist, et al.,
1996). In summary, before resorting to training, make sure it really is the answer to the
problem; and if training is whats needed, then its worth doing right.
The second part of this element is standard operating procedures. Many times incidents occur
not because the worker lacked skills or knowledge, but because the SOP was not designed
appropriately for the given conditions. The Training/SOP element acts as a safety intervention
in all three layers of system defenses (Fig. 20). At the Organizational Factors layer, the
responsibility is to institute sufficient training and effective, safe SOPs. At the Supervision layer,
the training/SOP defense is to ensure trained personnel are assigned to tasks and that they use
the SOPs. At the Preconditions layer, the worker must ensure that he/she has the required
training to do assigned work and that he/she understands and consistently uses SOPs. While
all three layers of system defense are necessary to ensure safe operations, it is clear that a
hole in the Organizational Factors layer (e.g., a decision to provide only the most basic
training/SOPs, or to shirk the responsibility and place it wholly on supervisions shoulders) will
do the most harm, since that missing barrier will affect the entire workforce.
14
Miller refers to this element just as training. However, a recent study (McCallum, Forsythe, et al.,
2000) notes that many incidents attributed to inadequate knowledge and skills are actually promoted
through incorrect SOPs, as opposed to insufficient training, per se.
Page 69 of 141
Houston, Texas
A final note on standard operating procedures. Many of them are written so poorly that
personnel are unable to use SOPs effectively. SOPs are important for safe operation, and they
should be well-written in order to convey the needed information to the users. A good tutorial on
writing SOPs is given by Information Mapping15.
7.2.6
Interpersonal Relationships (Communication)
Interpersonal relationships (communication) is what makes teams work. Offshore, shipboard,

and dockside activities all depend on teamwork. In earlier times, the Offshore Installation
Manager or the ships Master was the unquestioned authority in a one-way, top-down, chain of
command (Miller, 2000). However, most shipping and offshore companies (and many other
industries, such as aviation and nuclear power) have recognized that crew resource
management is an essential component of safety. Personnel need to feel empowered to speak
up, question, and double-check decisions and actions of other team members (including
supervisors). Research has shown that when crew members do not communicate effectively,
accidents result: 28% of personnel injuries and 18% of vessel casualties were related to
inadequate communications (McCallum, Raby, Rothblum, Forsythe, Slavich & Smith, 2000,
unpublished).
In the maritime and offshore environments, poor communications can be caused by factors in
addition to a lack of crew resource management. Physical constraints (separating persons who
need to be together via poor layout of rooms or facilities), inappropriate organizational structure
which puts too many chains of command between individuals who need to communicate, and
overloading workers so that they do not have the time to communicate can all contribute to poor
interpersonal communications. This is a factor that is often, and mistakenly, overlooked in
incident investigations.
Good communication is a safety intervention that must be established at all three layers of
system defense. As was the case with Training/SOPs, the Organizational Factors layer must
lay the foundation by making two-way communication a part of company culture and facilitate it
through good workplace design and policies. Supervisors must actively support and encourage
effective communications within and among teams. And to keep communications from
becoming a Precondition for Unsafe Acts, workers need to be involved, responsive members of
their work teams.
15
For more information about seminars by Information Mapping, please see their web site at
http://www.infomap.com or call (800) INFO MAP (463-6627).
Page 70 of 141
Houston, Texas
7.2.7
Job Aids
Job aids come in several forms such as hazard identification (warning) signs, operator and
maintainer manuals, and specific operating procedures. These can be of help in reducing
human-induced incidents, especially when learning a new task, performing a task that is done
infrequently, or completing a job that must be performed in an exact sequence. However, a
poorly prepared job aid can lead to incidents rather than prevent them. Examples of critically
important job aids found on ships and platforms are the lifeboat launching instructions and the
operating instructions for manually releasing the fire fighting suppressant system. Unfortunately,
these safety-critical instructions are typically confusing and difficult to understand (Miller, 2000).
There is a lot of research available on how to prepare good job aids, instructional placards, and
warning signs (Curole, McCafferty & McKinney, 1999; Laughery, Wogalter & Young, 1994;
Wogalter, Young & Laughery, 2001; and seminars by Information Mapping11). One concept,
called information mapping (Curole, et al., 1999), utilizes research on the human learning
process to provide very specific guidelines on how to prepare manuals, procedures, checklists
and other printed and-or pictorial job aids. Properly prepared job aids can be a useful barrier to
the prevention of maritime incidents. By the same token, poorly-written or missing job aids can
contribute to incidents and are an important aspect to be considered both during the incident
investigation and in the preparation of preventive recommendations.
7.2.8
Fitness for Duty
Fitness for duty is another term for adverse mental or physiological states that are severe
enough to reduce the individuals capacity to perform. These states can be due to physiological
conditions of illness or fatigue, or to the use of alcohol or drugs (including over-the-counter
medications). These states can also be psychological in nature such as emotional trauma due
to family or financial problems, or from a neurotic or even psychotic disorder. If any of these
things is sufficient to distract or otherwise impact the persons performance of safety-related
duties, it can be a definite contributor to an incident. This is another factor that is often
overlooked during the incident investigation process, but should receive attention.
Essentially, fitness for duty is the responsibility of the worker to keep it from becoming a
precondition. Line supervisors have a responsibility to ensure that workers are, in fact, fit for
duty. This is the one element of the triangle that the Organizational Factors layer doesnt
explicitly address (for example, work-rest policies and worksite environment, both of which can
affect an employees physiological and mental fitness for duty, would fall under other triangle
elements).
Fitness-for-duty testing is a controversial area. There are some tests available for determining
whether a person is under the influence of drugs or alcohol or severely fatigued; however, the
reliability of most of these tests is a hotly-debated issue. Some trucking companies have
successfully used a simulator-type test to ensure a truckers driving performance is up to par
before getting on the road. Such scientifically-validated and operationally-relevant screening
techniques have yet to be developed for the maritime and offshore industries.
Page 71 of 141
Houston, Texas
An alternative approach to testing is to implement a crew endurance program (Comperatore,

Rothblum, Kingsley, Rivera, & Carvalhais, 2001). This type of program educates personnel as
to how fitness for duty can affect not only job performance but long-term health and assists
personnel in controlling the hazards that affect fitness for duty (note that such a program would
be part of the Policies & Culture and Training/SOP elements of the triangle, not the Fitness for
Duty element). Since many of the variables that influence fitness for duty are job-related (e.g.,
fatigue from poorly designed work schedules, or lack of coordination and manual dexterity from
working in too-cold temperatures), a crew endurance program helps to prevent latent factors at
all three layers: organizational factors, supervision, and preconditions for unsafe acts.
To summarize, the precursor for effective safety solutions is an in-depth analysis of incidents.
Only by understanding all the latent factors which contributed to an incident can one determine
what barriers would be effective in either averting or reducing the effects of similar incidents in
the future. It is helpful to establish barriers or interventions within all layers of system
defenses in order to reduce the likelihood of future incidents. However, when this is not
feasible, try to implement interventions in the areas of policies & culture, workplace design, and
environmental control, as these are often latent factors for many incidents and are usually more
effective at preventing incidents than interventions towards the tip of the triangle of
effectiveness. One last key to successful interventions: make your safety interventions
SMART; that is, they should be specific, measurable, attainable, reasonable, and timely.
8.0
USER EXPERIENCES WITH STARTING AN INCIDENT INVESTIGATION PROGRAM
This section provides two examples of agencies which have begun to incorporate human
factors incident investigations. Their experiences demonstrate how human factors incident
investigation is used and provide some lessons learned.
8.1
U.K. Marine Accident Investigation Branch
8.1.1
Background
The Marine Accident Investigation Branch (MAIB) is an independent division of the United
Kingdoms Department for Transport, Local Government and the Regions (DTLR). The chief
inspector reports directly to the Secretary of State, and is empowered to investigate marine
accidents and hazardous incidents occurring onboard or to UK registered ships worldwide, and
to all other vessels within UK territorial waters. Each year MAIB receives over 2,000 incidents.
Presently, about 60 field investigations are undertaken annually by 13 inspectors working
individually or as a team. Such field investigations, including the formal report produced, take
about 10-12 months. MAIB inspectors also investigate about 550 additional incidents by paper
and telephone.
The fundamental purpose of an MAIB investigation is to determine the circumstances and
causes of an accident or incident with the aim of improving the safety of life at sea and the
Page 72 of 141
Houston, Texas
avoidance of accidents in the future. It is not the purpose to apportion liability nor to apportion
blame: MAIB is not an enforcing authority; that role is taken by the Maritime and Coastguard
Agency, a totally separate organization within the DTLR. An MAIB investigation is conducted in
accordance with the provisions of The UK Merchant Shipping (Accident Reporting and
Investigation) Regulations 1999, and aims to determine: what happened; how it happened; why
it happened; and what can be done to prevent it from happening again.
8.1.2
Events Leading to Human Factors Investigations
In the 1980s the Surveyor General Organization (predecessor of the Maritime and Coastguard
Agency) commissioned the Tavistock Institute of Human Relations, London, to carry out a
series of studies on the human element in shipping casualties. This was done in order to
understand, and hence reduce, the dangers associated with human frailty in the United
Kingdom merchant fleet. This was the first systematic attempt to consider human error in
shipping casualties in the UK. Some of the recommendations from the study were that:
More attention should be paid to the structure, order, and timing of questions in
accident investigation.
A check-list of human element questions should be developed to help the

investigator in relating his thoughts to the general body of human factors knowledge.
A search for alternative explanations to accidents should be consciously developed.

This would discourage a pre-occupation with finding a single best explanation, which
may be counterproductive to revealing the true facts about specific casualties.
Regular seminars should be conducted to exchange experiences about casualty

investigation.
A computerized and flexible accident data system should be developed to aid human
factors research.
MAIB was set up as a separate organization in July 1989. It was chartered to investigate
accidents, keeping this function separate from the Maritime and Coastguard Agencys
responsibility for the regulation of ship safety.
8.1.3
MAIBs Classification of Human Factors Causes
MAIB began with a relatively simple taxonomy of human factors, looking mainly at operator error
and organizational factors. In 1994, MAIB developed a more comprehensive classification of
human factors contributions to accidents and incidents. Based on Reasons model of accident
causation, the classification developed aimed to show how active human errors or violations are
shaped by latent failures. MAIB currently has six top-level human factors classifications:
External bodies liaison (e.g., regulations)
Company & organization
Page 73 of 141
Houston, Texas
Crew factors
Equipment
Working environment
Individual
There are sub-classifications under each of these headings (see Appendix D for the full
taxonomy). The database was designed to encourage the examination of accident context
either from the individual outwards to the regulatory and policy context, or from the context and
company inwards to the vessel and the individuals who operate it. In this way, MAIB increases
the likelihood of identifying contributing factors at all levels. Some of these levels may be
related: for example, company policy on training may influence skills and knowledge at an
individual level. In many cases, there will be unrelated human factors areas which contribute to
an accident. All the human factors causes which can be identified from the evidence available
should be classified.
8.1.4
Human Factors Training for MAIB Inspectors
The development of inspector skills and understanding of human factors investigation has been
an evolutionary process. The thirteen MAIB inspectors work under one roof in Southampton
which provides an ideal opportunity to share with each other, on a daily basis, experiences with
accident investigation. Initially, MAIB inspectors attended human factors training courses
provided by the Transportation Safety Board of Canada and by Det Norske Veritas (DNV) in
Atlanta. Inspectors were encouraged to attend seminars and lectures on human factors.
This was an ad-hoc, but to some extent effective, approach to introducing a more formal
method of human factors investigation. However, a significant drawback was the difference in
terminology used by the varying sources of training and guidance. This hindered investigation
team effectiveness and made quality assurance of the investigation process and reporting
difficult and inconsistent.
To enable a common understanding of human factors investigation, an MAIB training course
was developed to achieve a more consistent and reliable approach by inspectors to evaluate
the human factors causes of accidents. A certain amount of consistency of reporting and data
input has been achieved, but inconsistency and occasional confusion does sometimes arise.
MAIB tries to overcome these problems by internal seminars on human factors, regular reviews
and audits of the investigation process and outcomes, and attendance of the MAIB course. The
course content adapts to the changing needs and experience of the inspectors.
The most recent MAIB course lasted two days and included topics such as: a general
introduction to human factors (human performance, teamwork, basic methods, terms, and
tools); human factors and accident investigation, including models of causation; human error
and error analysis (such as, what is human error, why does it occur, how is it assessed,
Page 74 of 141
Houston, Texas
performance influencing factors, the SHEL model, preventing errors and violations and
minimizing the impact of errors, safety management, and safety culture).
A formal approach to human factors investigation has also highlighted the importance of
effective interviewing techniques. Consequently, attendance by all inspectors on an in-house
interviewing techniques course run by an experienced trainer is mandatory. Continuous
development of inspector competence in human factors investigation is also promoted through
mini-coaching sessions, self-study, and the day-to-day application of MAIBs formal
investigation process. As a result, MAIB now has a team of inspectors competent in the
investigation and identification of human factors contributors to marine casualties.
8.1.5
Benefits of MAIBs Database
More than 1,000 accidents investigated by MAIB inspectors are recorded in the database. A
measure of success of the database is the increasing demand from diverse interests in the
marine industry. Excluding MAIB, the main users of the database are the Maritime and
Coastguard Agency (MCA), university researchers, and consultants, all seeking patterns and
trends in accident types and causes. Information in the database has been particularly helpful
to the MCA who, along with other flag states, is introducing codes of operational inspections
and risk assessments and certificates of competency based on performance standards.
The historical information from similar accidents has also proven to be a powerful tool to
promote MAIB arguments for safety changes. For example, MAIB analysis has uncovered
trends in accidents during lifeboat launching and recovery caused by a multiplicity of human
factors. The study was able to identify common factors leading to these accidents, and the
risks associated with lifeboat launching systems, by examining the common problems
encountered. Operators make mistakes in maintenance and operation of launching equipment
because of overly complex design and inadequate operator manuals. Over the years, the size
and weight of lifeboats and equipment have increased, diminishing the ability of seamen to
handle launching and recovery operations safely. The database analyses have allowed MAIB
to understand the problems and to make safety recommendations.
8.2
California State Lands Commission, Marine Facilities Division
8.2.1
Background
The Marine Facilities Division (MFD) of the California State Lands Commission is
headquartered in Long Beach, California. Created in 1990, the MFD is tasked with pollution
prevention at marine oil terminals. Towards this end, MFD inspectors monitor activities and
enforce regulations at 85 marine facilities along the California coast. Inspectors oversee and
evaluate the safety of such operations as oil transfers to and from oil tankers and barges and
make comprehensive inspections of marine oil terminals and pipelines. At a Facilitys request,
Division Specialists also conduct safety management assessments aimed at identifying
potential trouble areas in an organizations defenses against adverse incidents.
Page 75 of 141
Houston, Texas
8.2.2
Initiation of Human Factors Investigations
MFD specialists and inspectors also investigate oil spills as a means of informing prevention
strategies. Up until recently, those inquiries identified personnel, organizational, and equipment
factors as primary or secondary causes of spills, but without clearly distinguishing active failures
from latent system conditions. Additionally, it was difficult to capture within the investigation
framework the multiple factors that often conspired to bring about a single adverse event. In
May of 2001, MFD introduced its inspectors to the Human Factors Analysis and Classification
System (HFACS; Wiegmann & Shappell, 1999) and has begun using that model to support
inquiries and to analyze system failures that contribute to spills. The HFACS taxonomy was
selected in part because it is particularly well-suited for prevention, in that it encourages a focus
on failed system defenses, rather than on individual failures. This allows users of the resulting
data to address the appropriate system components in devising prevention strategies.
8.2.3
Adaptations to HFACS
It was initially clear that MFD needed to extend the tool to cover certain structural and
mechanical faults as well as environmental conditions in order to cover the spectrum of
contributors to oil spills. Note that this is not a deficiency on the part of HFACS: HFACS was
intended to guide the human factors portion of an investigation. Naturally, there will be
equipment factors, weather factors, and other non-human factors that contribute to many
casualties and incidents. It will be necessary for any company to develop its own set of
classifications to capture these types of problems. However, MFD observed that most of the
contributing factors to oil spills were, in fact, human factors, so they found it handy to use
HFACS as their main classification tool and add the equipment and environmental factors to it.
Their complete investigation taxonomy is shown in Appendix E (see HFACS Layer Guides).
Equipment factors were appended to the Layer 1 Guide (Unsafe Acts from HFACS) as
Structural/Mechanical Damage/Failure. This is the equipment analog of an unsafe act, in that
the damage or failure appears as the immediate cause of an incident (e.g., the oil spill appeared
to be caused by a damaged valve). Similarly, equipment and environmental factors were added
to the Preconditions layer (Layer 2). Just as complacency is an adverse mental state
(precondition) that can lead to a routine violation (unsafe act) like taking a shortcut that causes
an incident, substandard equipment design, such as an ambiguous display, can be the
precondition for an unsafe act (misreading the display and causing an incident). In this way,
MFD combined the major oil spill causal factors both human and non-human into a single,
HFACS-like taxonomy.
Adaptations were made to the human factors taxonomy as well. HFACS was originally
developed for aircraft accidents, and thus incorporates certain terminology and causal factors
related to aviation (such as hypoxia and spatial disorientation). These terms were dropped.
Maritime industry specific terms are captured using event data forms, which require an incident
specific statement of the actor (an individual or group, or a structure/part) and a situation-
Page 76 of 141
Houston, Texas
specific description of the action/inaction or system failure that contributed to the incident. This
remains an ongoing process.
8.2.4
Training and Job Aids
HFACS was introduced to MFD when their human factors analyst took a full-day workshop on
the topic led by Drs. Wiegmann and Shappell. He then developed and provided training on
HFACS to MFDs inspectors and specialists. Discussions among MFD staff occur during
monthly meetings, and along with input from other maritime industry representatives, these
meetings led to modifications in the taxonomy (discussed above) and to the development of job
aids. The job aids used by MFD inspectors are provided in Appendix E.
New job aids and revisions of existing ones are considered regularly in response to issues that
arise during monthly meetings. Situation-specific guidance sheets are presently under
development to guide team members in collecting essential information in response to
particular circumstances. Guidance is in a bulleted If then form. For example,
If the incident involves turning an incorrect valve,
Then
photograph the valve and its immediate area;
ask about a history of such incidents;
ask whether the involved personnel were experienced with the equipment;
detail communications leading up to the action.
The intention of these sheets is to assure that relevant data are collected before they are lost.
8.2.5
Benefits of Human Factors Investigations
MFD is less than a year into using the revised HFACS investigation tool, and is still learning
about the process and making changes to its procedures. However, an early analysis of six
incidents shows that inspectors are learning to use the tool. Of the 21 causal factors identified,
20 could be completely categorized by the inquiry team which included marine safety
inspectors, specialists and a human factors analyst using HFACS.
Not surprisingly, inquiry teams were more successful at identifying unsafe acts and
preconditions than they were at finding problems in other latent factors like unsafe supervision
and organizational influences. Team members felt that they had sufficiently considered and
identified all relevant unsafe acts in five (of the six) incidents, and had identified all the
preconditions in four of the incidents. However, teams judged that they had identified all the
potential types of unsafe supervision in only three of the incidents, and had identified all the
organizational factors in only one incident.
There are five issues that may contribute to the difficulty of identifying latent factors. The first is
that it can take substantial time and resources (on the parts of both the investigator and those
being investigated) to dig beneath the surface and unearth latent factors. Sufficient time is not
Page 77 of 141
Houston, Texas
always available. Second, when a regulatory body (such as MFD) or an employer is also the
investigator, the fear of punishment can be a disincentive for those under investigation to fully
cooperate and help identify latent factors. Third, necessary information may be unavailable,
either because it is confidential (e.g., personnel records) or unrecoverable (e.g., a momentary
state of mind, an absent maintenance record). Fourth, organizational influences in particular
sometimes only become apparent over the course of several incidents rather than in a single
one. Team members have addressed this through incorporating a structured note rather than
assigning a contributing factor when they have reason to believe but not definitive evidence
that an organizational factor is among weakened defenses in a particular case. Finally, those
conducting the inquiry may lack the know-how or experience to ask the appropriate questions.
This can occur for those with considerable maritime knowledge as well as those with human
factors knowledge, since for any one investigator that knowledge is likely to be centered around
particular areas (ship, terminal, company operations, or management), and their ability to
establish a comfortable rapport with key individuals related to the incident will vary accordingly.
While there are too few cases yet to allow for meaningful analysis, there have already been
lessons learned that can help commercial companies improve safety. MFD is starting a
newsletter as a way of sharing this information with the marine terminal companies. Another
benefit from these initial uses of HFACS has been the discovery of an area in which MFD can
improve its reviews of preventative maintenance programs. Additionally, notes have been
expanded to capture instances when outside influences factors other than terminal and
vessel organizations contribute to incidents. In summary, MFDs human factors incident
investigation program has gotten off to a good start and shows promise in discovering how
marine terminals can change their policies and operations to improve safety and reduce oil
spills.
8.3
Stolt-Nielsen Transportation Group
8.3.1
Background
Stolt-Nielsen S.A. is one of the worlds leading providers of transportation services for bulk
liquid chemicals, edible oils, acids, and other specialty liquids. The company, through its parcel
tanker, tank container, terminal, rail and barge services, provides integrated transportation for
its customers. Stolt-Nielsen Transportation Group owns 72 ships involved in the chemical
parcel trade: 51 ships in world-wide trade and 21 ships in coastal trade in Europe and the Far
East.
Stolt-Nielsen has developed a full International Ship Management (ISM) Quality and Safety
Program for its ships. This program is to the highest standards of the industry and is audited by
three classification societies. The safety program tracks and investigates all incidents and uses
statistical process control methods to identify trends. In turn, this information is used to develop
training and educational programs designed to reduce risks and losses. As an industry leader,
Page 78 of 141
Houston, Texas
Stolt-Nielsen also shares its data and benchmarks with other companies through groups such
as the National Safety Council16.
8.3.2
Human Factors Incident Investigation
Stolt-Nielsen has been doing incident investigation for a number of years. Since it already had
an ongoing training and Quality Assurance program, the incident investigation program was
added under the same umbrella at minimal cost to the company. In 1992, a human factors
investigation taxonomy was added. Stolts investigation form (see App. F), while simpler than
those used by MAIB and MFD, has nevertheless proven to help the company detect and correct
a variety of safety hazards.
Incident investigation is a combined responsibility of ships officers and of Stolts Division of
Marine and Safety Services. Officers are trained on incident investigation, and other Quality
Assurance topics, every three years. Because Stolts officers and crew are from around the
world, training is an expensive undertaking. All incident forms are sent to the Assistant
Manager of Marine and Safety Services, who completes the investigation. By having a single
person ultimately responsible for the incident data helps to keep the data reliable (consistent
use of terms and coding). The major problem with investigating incidents is that the fleet is
distributed worldwide, making timely and accurate reporting a challenge. However, persistence
on the part of the investigator and a shared understanding of the importance of the incident
program has led to a successful program.
The Assistant Manager is also responsible for the data analysis, which is a plus, since the
analyst knows the terminology and understands the constraints under which the data were
collected, which in turn reduces the likelihood that unwarranted data comparisons will be made.
Stolt analyzes their incident data quarterly, looking both for trends within the company and
benchmarking their incident rates against those of other shippers that are members of the
National Safety Council. Lessons learned from incident analyses are disseminated widely
through the company via Loss Control Bulletins and training programs.
8.3.3
Benefits of the Incident Database
Stolt enjoys a lower incident rate than the industry average (Fig. 21), due in large part to its
attentive tracking of incidents and responsive safety interventions. Stolt uses frequency
analysis and analyzes trends over years to identify safety problems and track the success of its
interventions. As described in the Analysis section, Stolt had used incidents to determine that
crew members suffered a high frequency of slips, trips, and falls. In response, the company
acquired new safety shoes designed for better traction on wet surfaces. Follow-up statistics
16
The Waterborne Transport Division of the National Safety Council currently keeps safety data that
members can use for benchmarking. There are plans to produce guidance and training to help
members improve their safety analysis and benchmarking capabilities. For more information, please see
their website at http://www.waterbornetransport.com or contact William Boehm by phone at (281) 8605043 or by email at wboehm@stolt.com .
Page 79 of 141
Houston, Texas
Lost Time Injury Rates (Industry vs. Stolt)

(Incidents per 200,000 person-hours worked)
1.4
LTIR INDUSTRY
LTIR STOLT
1.2
0.8
0.6
0.4
0.2
0
1992
1993
1994
1995
1996
1997
1998
1999
2000
Figure 21.
Comparison of Lost Time Injury Rates between Stolt

and an Industry Average from 1992 through 2000
This is an example of how Stolt uses incident data to benchmark
its safety relative to other companies. (Data courtesy of StoltNielsen Transportation Group.)
revealed a marked decrease in slips, trips, and falls, which appear to show the success of the
intervention.
Another use of incident data is shown in Figure 22, which depicts a comparison of the
frequencies of various types of injuries (data for the first three quarters of 2001). As can be
seen in the graph, injuries to the head and eyes happened most frequently, and these findings
were supported by data from prior years. Thus, Stolt has reviewed the types of safety glasses
and hardhats used by crew members and have identified problems with the current safety
equipment. An alternate type of safety glasses is now being tried which hopefully will provide
better protection. Regarding hardhats, one of the problems discovered was that current
hardhats were uncomfortable, and crew members did not wear them consistently. A new type
of head protection a ball cap with a butchers hard cap inserted inside is being trialed as a
result.
Page 80 of 141
Houston, Texas
Analysis of Body Parts Injured

(1Q01 - 3Q01)
8
120%
7
100%
Number of Incidents
80%
5
60%
3
40%
Sum Total Percent of Incidents
2
20%
1
Figure 22.
w
N
ec
k
N
os
e
O
th
er
Sk
in
To
e
r
Ea
bo
El
ee
ld
er
W
r
A b is t
do
m
en
C
he
st
C
hi
n
ou
Kn
Sh
ot
Fo
An
kl
e
0%
Ey
e
H
ea
d
Fi
ng
er
H
an
d
Ba
ck
Le
g
M
ou
th
Ar
m
Number of Incidents as a Function of Body Part Injured.

(Data from first three quarters of 2001; courtesy of StoltNielsen Transportation Group.) This shows how Stolt uses
incident data to identify safety hazards.
Stolt regards its incident investigation program to be a resounding success. By using incident
investigation as a part of its overall Quality Assurance program, Stolt has been able to identify
and correct safety problems, many of which have human factors causes. Through a consistent
focus on incident causes and efforts to remediate those causes, Stolt has achieved lower injury,
accident, and pollution rates than the industry average. The company is justifiably proud of its
safety record.
Page 81 of 141
Houston, Texas
9.0
SUMMARY
As we have seen, human error (and usually multiple errors made by multiple people and at
multiple levels of the organization) contributes to the vast majority (over 80%) of marine
casualties and offshore incidents, making the prevention of human error of paramount
importance if we wish to reduce the number and severity of maritime and offshore incidents.
Many types of human errors were described, the majority of which were shown not to be the
fault of the human operator. Rather, most of these errors tend to occur as a result of
technologies, work environments, and organizational factors which do not sufficiently consider
the abilities and limitations of the people who must interact with them, thus setting up the
human operator for failure.
Effective
Safety Programs
Data-Driven
Research
Triangle of
Effectiveness
Incident
Investigation
Events & Causal
Factors Charting
Human
Error
Human-System
Approach
HFACS
HF questions
(App. H)
Incident
Database
HFACS
Database
Analysis
Section 6.0
Generic
incident database (App. B)
Specialized HF
items (App. G)
Feedback
Figure 23.
Tools for an Effective Incident Investigation Program
Incident investigation that includes an analysis of human error is needed if we are to prevent
these incidents in the future. This paper has presented several different tools that can be used
to perform a human factors incident investigation and to use the resulting data to improve the
companys safety program (Fig. 23). Appendix H presents a selection of human-related
questions that can be asked to identify potential human error issues. The human-system
Page 82 of 141
Houston, Texas
approach helps the investigator consider suboptimal interfaces and interactions between
people, technology, organization, and environment that may have contributed to the incident.
Reason (1990) created a useful framework for categorizing the types of human error. His
Swiss cheese model considers not just the unsafe acts of the operator, but also considers
several layers of system defenses that may need mending if the safety program is to be
effective: preconditions for unsafe acts, unsafe supervision, and organizational factors.
Reasons model has been captured in the Human Factors Analysis and Classification System
(HFACS; Shappell & Wiegmann, 1997a, 2000; Wiegmann & Shappell, 1999), an incident
investigation system which has been used widely and successfully in military and industrial
incident investigation. HFACS was presented in this document because it is relatively easy to
learn and use, and because it has a history of enabling successful safety programs. HFACS
can be used by the offshore and maritime industries to supplement existing incident data
systems with human factors information. For companies that have not yet begun an incident
investigation program, additional classification schemes, both for specific human-related errors
(fatigue, communications, and skills and knowledge) and for non-human incident data (e.g.,
vessel or platform type, activities/operations during which the incident occurred, environmental
and weather conditions that may have played a role) are provided in the Appendices. Event
and Causal Factors Charting was introduced as an additional tool to aid in understanding the
events that led to an incident and the causal factors that underlie those events. Used together,
Event and Causal Factors Charting followed by an HFACS analysis of the causes can provide a
powerful way to represent the development of an incident and to identify the system failures that
generated and perpetuated the incident.
It cannot be overemphasized that a good incident database is only the starting point for a
successful incident prevention program. An open, fair, improvement-seeking culture, a
common understanding of the purpose and scope of the incident investigation program,
appropriate training for incident investigators, a simple, user-friendly database, and feedback on
the results of the incident investigation program are all essential elements to the collection of
valid and complete incident data. In addition, regular analysis of the incident data is required to
identify potential problems and to evaluate the results of new safety programs. Several data
analysis techniques were summarized that can help companies make the most of their incident
data. As was pointed out, one cannot do data analysis blindly one must consider changes in
policies and procedures that may have had an effect on the way data were collected and
classified. Thoughtful analysis will help to distinguish spurious results from real trends that may
require intervention.
Follow-up studies (data-driven research) are usually needed to
thoroughly understand a given safety issue and determine what types of interventions may be
needed.
Finally, we considered how to select interventions based on the types of system defenses that
have failed. By linking Reasons Swiss cheese with Millers triangle of effectiveness, we
have a tool for finding the most effective ways to solve safety problems. While traditional safety
Page 83 of 141
Houston, Texas
management seems to focus on reprimanding, cajoling, and more training, the triangle of
effectiveness shows that these are the least effective ways for reducing incidents. A safety
culture must start at the top, and so, too, must the most effective interventions. Management
participation, human-centered workplace design, and human-compatible environmental control
may require more up-front effort than yet another training course, but because these elements
are integral to the safe design and operation of the workplace, they will reap much larger safety
benefits. The safety-conscious organization starts at the top when developing safety
interventions to protect its employees, products, and the environment.
Human errors can be reduced significantly. Other industries have made tremendous progress
in controlling human error through careful documentation of incidents, analysis of incident data,
follow-up studies, and top-down, human-centered interventions. Indeed, maritime/offshore
industries can do the same: the U.K.s Marine Accident Investigation Branch, Californias
Marine Facilities Division, and Stolt-Nielsen Transportation Group have all shown that the
maritime/offshore sector can put human factors incident investigation to effective and profitable
use. By using human factors incident investigation to identify weaknesses in our system
defenses, and by crafting safety interventions through the human-centered design of
technologies, work environments, and organizations, we can support the human operator and
foster improved performance and fewer incidents.
10.0
REFERENCES
1.
American Bureau of Shipping [ABS] (1998). Guidance Notes on the Application of

Ergonomics to Marine Systems. New York: Author.
2.
American Bureau of Shipping [ABS] (2002). Guide for Crew Habitability on Ships.
December, 2001. Houston, TX: Author.
3.
American Society of Testing and Materials [ASTM] (1995) Standard practice for
human engineering design for marine systems, equipment, and facilities. Standard
F1166-95a. West Conshohocken, PA: Author.
4.
Bea R. (1997) Accident and Near-Miss Assessments and Reporting. In R.G. Bea,
R.D. Holdsworth, and C. Smith (Eds.), 1996 International Workshop on Human
Factors in Offshore Operations. New York: American Bureau of Shipping.
5.
Bea R., Holdsworth R., & Smith C. (1997) Human and Organizational Factors in the
Safety of Offshore Platforms. In R.G. Bea, R.D. Holdsworth, and C. Smith (Eds.),
1996 International Workshop on Human Factors in Offshore Operations. New York:
American Bureau of Shipping.
6.
Bryant D.T. (1991) The Human Element in Shipping Casualties. Report prepared
for the Dept. of Transport, Marine Directorate, United Kingdom.
Page 84 of 141
Houston, Texas
7.
Bryden R., OConnor P., & Flin R. (1998) Developing CRM for the Oil and Gas
Industry: A human factors approach. Poster presentation at the Human Factors and
Ergonomics Society 42nd Annual Meeting, October 5-8, 1998.
8.
Card J.C. (1997) The Coast Guard, Offshore Industry, and the Human Element. In
R.G. Bea, R.D. Holdsworth, and C. Smith (Eds.), 1996 International Workshop on
Human Factors in Offshore Operations. New York: American Bureau of Shipping.
9.
Center for Chemical Process Safety. (1992) "Guidelines for Investigating Chemical
Process Incidents". American Institute of Chemical Engineers: New York.
10.
Center for Chemical Process Safety. (in preparation) Guidelines for Investigating
Chemical Process Incidents (update to the 1992 edition). American Institute of
Chemical Engineers.
11.
Comperatore CA, Rothblum AM, Rivera PK, Kingsley LC, Beene D, and Carvalhais
AB. (2001) U.S. Coast Guard Guide for the Management of Crew Endurance Risk
Factors - Version 1.1. U.S. Coast Guard Research and Development Center Report
No. CG-D-13-01. Washington, D.C.: U.S. Dept. of Transportation, U.S. Coast Guard.
Available on the web at: http://www.rdc.uscg.gov/Reports/CGD1301Report.pdf
12.
Coren S. (1998) Sleep deficit, fatal accidents, and the spring shift to daylight
savings time. Presented at INABIS 98 5th Internet World Congress on Biomedical
Sciences at McMaster University, Canada, Dec. 7-16, 1998. Available on the web at
http://www.mcmaster.ca/inabis98/occupational/ coren0164/index.html
13.
Cormier P.J. (1994) Towing Vessel Safety: Analysis of Congressional and Coast
Guard Investigative Response to Operator Involvement in Casualties Where a
Presumption of Negligence Exists. Masters Thesis, University of Rhode Island.
14.
Curole M., McCafferty D., and McKinney A. (1999) Human and Organizational
Factors in Deepwater Applications. Paper presented at the Offshore Technology
Conference (OTC# 10878), Houston, TX, May 3-6, 1999.
15.
Det Norske Veritas (1995). Presentation on safety management systems to the U.S.
Coast Guard, February 9, 1995.
16.
Edwards E. (1972) Man and machine: Systems for safety. In Proceedings of the
BALPA Technical Symposium, London.
17.
Ferguson S.J. & Landsburg A.C. (1998) BIMCO/USCG partner for safety:
aboard for NMSIRS! BIMCO Bulletin, 93 (6), 42-48.
18.
Franklyn C. (2000). Incident Investigation Training Course. Houston, TX: Risk,

Reliability and Safety Engineering (RRS), Inc.
19.
G & C Merriam Company (1973) Websters New Collegiate Dictionary. Springfield,

MA: Author.
Page 85 of 141
all
Houston, Texas
20.
Gordon R., Flin R. & Mearns K. (2001) Designing a human factors investigation tool
to improve the quality of safety reporting. Proceedings of the Human Factors and
Ergonomics Society 45th Annual Meeting, 1519-1523.
21.
Gudmestad O.T., Gordon R., Bellamy L., Brabazon P., Craig M., Lee G., Moore J.F.,
Rettedal W.K., Snell R., & Vegge A. (1997) The Role of Human and Organisational
Factors (HOF) in the Fabrication, Installation and Modification (FIM) Phases of
Offshore Facilities.
22.
Hawkins F.H. (1987) Human Factors in Flight. Aldershot, U.K.: Gower Technical
Press.
23.
Helmreich R., Klinect J., et al. (2000) Models of threat, error and CRM in flight
operations. In Tenth International Symposium on Aviation Psychology. Columbus,
OH: The Ohio State University.
24.
Hill S.G. & Byers J.C. (1992a) How to Apply Human Factors in Investigation.
Lesson 5 from A Human Factors Module for the U.S. Coast Guard Investigation
Department Course.
25.
Hill S.G. & Byers J.C. (1992b) Introduction to the Human Element in Marine
Casualties. Lesson 1 from A Human Factors Module for the U.S. Coast Guard
Investigation Department Course.
26.
Hill, S.G., Byers, J.C., and Rothblum, A.M. (1994) Gathering and recording humanrelated causal data in marine and other accident investigations. In Proceedings of
the Human Factors and Ergonomics Society 38th Annual Meeting, 1994, pp. 863867.
27.
Hollnagel E. (2000) Barrier Analysis and Accident Prevention. Paper presented at

the 5th Australian Aviation Psychology Symposium, November, 2000.
28.
Howard M., Fleming M.T., Holt J., McSween T., Moreno M., Perron P., & Reece
M.D. (1997) Improving Offshore Drilling, Workovers, Production Operations and
Maintenance Through Practical Application of Human and Organizational Factors.
In R.G. Bea, R.D. Holdsworth, and C. Smith (Eds.), 1996 International Workshop on
Human Factors in Offshore Operations. New York: American Bureau of Shipping.
29.
Kay G. (2000) Medical Workshop (talk on neuropsychology, antihistamines, fatigue,

and performance). Paper presented at Alertness 2000: Tools for Managing Fatigue
in Transit. Oct. 26-27, 2000, Arlington, VA.
30.
Kirwan B. (1997) Evolving Human Factors in Offshore Operations. In R.G. Bea,

R.D. Holdsworth, and C. Smith (Eds.), 1996 International Workshop on Human
31.
Kontogiannis T. (1999) User strategies in recovering from errors in man-machine

systems. Safety Science, 32, 49-68.
Page 86 of 141
Houston, Texas
32.
Laughery K. R. Sr., Wogalter M.S., and Young S.L.. (Eds.) (1994) Human Factors
Perspectives on Warnings. Santa Monica, CA: Human Factors and Ergonomics
Society.
33.
Marine Accident Investigation Branch [MAIB]. (1977) Inspectors Report of an

Investigation into a Fatal Accident on board RIX HARRIER at Immingham Oil Jetty
on 22 July 1997.
34.
Marine Accident Investigation Branch [MAIB]. (1999) Report on the Investigation of

the Collision of DOLE AMERICA with the Nab Tower in the Eastern Approaches to
The Solent on 7 November 1999. MAIB Marine Accident Report.
35.
Marine Transportation Research Board [MTRB]. (1976) Human Error in Merchant

Marine Safety. Washington, DC: National Academy of Science. AD/A-028 371.
36.
Maritime Safety Committee [MSC]. (1998) Role of the Human Element in Maritime
Casualties. Report of the Joint ILO/IMO adhoc Working Group on Investigation of
Human Factors in Maritime Casualties. MSC 69/13/1. London: International
Maritime Organization.
37.
McCafferty D.B. (2000) HF Design Standards. Paper presented at Maritime Human

Factors Conference 2000, Linthicum, MD, March 13-14, 2000.
38.
McCallum M.C., Forsythe A.M., Raby M., Barnes A.E., Rothblum A.M., and Smith
M.W. (2000, unpublished) Skill and Knowledge Limitations in Marine Casualties: A
Research Study to Develop and Evaluate Investigation, Reporting, and Analysis
Procedures. Study performed for the U.S. Coast Guard Research and Development
Center.
39.
McCallum M.C., Forsythe A.M., Smith M. W., Nunnenkamp J.M., and Sandberg G.R.
(2001) Developing Performance-Based Assessments of Mariner Proficiency. U.S.
Coast Guard Research & Development Center Report # CG-D-06-01 (ADA390929).
Washington, D.C.: Department of Transportation, U.S. Coast Guard.
40.
McCallum M.C., Raby M., and Rothblum A.M. (1996) Procedures for Investigating
and Reporting Human Factors and Fatigue Contributions to Marine Casualties.
Washington, D.C.: U.S. Dept. of Transportation, U.S. Coast Guard Report No. CGD-09-97. AD-A323392.
41.
McCallum M.C., Raby M., Rothblum A.M., Forsythe A.M., Slavich A., and Smith M.
W. (2000, unpublished) Communications Problems in Marine Casualties: A
Research Study to Develop and Evaluate Investigation, Reporting, and Analysis
Procedures. Study performed for the U.S. Coast Guard Research and Development
Center.
42.
Miller G.E. (1997) U.S. Coast Guard Human Factors Engineering (HFE) Program
Training Course. Course prepared and delivered for U.S. Coast Guard Vessel
Compliance Division (G-MOC).
Page 87 of 141
Houston, Texas
43.
Miller G.E. (1999) Human Factors Engineering (HFE): What It Is And How It Can
Be Used To Reduce Human Errors in The Offshore Industry. Paper given at the OTC
Conference (OTC# 10876), Houston, TX, May 3-6, 1999.
44.
Miller G.E. (2000) The Current and Suggested Role of Human and Organizational
Factors (HOF) for the Reduction of Human Induced Accidents and Incidences [sic] in
the Maritime Industry. Paper presented at Maritime Human Factors 2000, Linthicum,
MD, March 13-14, 2000.
45.
Miller G.E. (2001) The Human Element and the Maritime, Changing the Mindset.
Paper given at the Marine Log Sixth Annual Conference on the Role of the Human
Element in Maritime Operations, Washington, D.C. April 24-25, 2001.
46.
Miller G.E., Bockosh G., Edmondson N., Holdsworth R., Khandpur R., Mello J., &
Mirabella J. (1997) Further Developments of Standards, Specifications, and
Guidelines Related to Human and Organizational Factors (HOF) to Reduce Human
Error in Offshore Facilities Operations. In R.G. Bea, R.D. Holdsworth, and C. Smith
(Eds.), 1996 International Workshop on Human Factors in Offshore Operations.
New York: American Bureau of Shipping.
47.
Monk, T.H.
(1980)
Traffic accident increases as a possible indicant of
desynchronosis. Chronobiologia, 7, 527-529.
48.
Muchinsky P.M. (1997) Psychology Applied to Work (5th edition). Pacific Grove,
CA: Brooks/Cole Publishing Company.
49.
National Research Council [NRC]. (1990)

Washington, DC: National Academy Press.
50.
National Transportation Safety Board [NTSB]. (1981) Major Marine Collisions and
Effects of Preventive Recommendations. Report No. NTSB-MSS-81-1.
51.
Perrow C. (1984) Normal Accidents: Living with High-Risk Technologies. Basic

Books.
52.
Rasmussen J. (1987) The definition of human error and a taxonomy for technical
system design. In J. Rasmussen, K. Duncan, and J. Leplat (Eds.), New Technology
and Human Error. Toronto: Wiley.
53.
Reason J. (1990) Human Error. New York: Cambridge University Press.
54.
Rothblum A. (2000) Human Error and Marine Safety. Presented at the Maritime
Human Factors Conference 2000, Linthicum, MD, March 13-14, 2000.
55.
Rothblum A., Chaderjian M., and Mercier K. (2000). Development and Evaluation of
the International Maritime Information Safety System (IMISS). Presented at the
Society of Naval Architects and Marine Engineers, IMISS Public Meeting, May 2000.
56.
Rothblum A.M. (2000). Dont Just Stand There, Say Something: Communications
Errors in Maritime Casualties. Paper presented at the Maritime Human Factors
Conference 2000, Linthicum, MD, March 13-14, 2000.
Page 88 of 141
Crew Size and Maritime Safety.
Houston, Texas
57.
Sanquist T.F., Lee J.D., McCallum M.C., and Rothblum A.M. (1996). Evaluating
Shipboard Automation: Application to Mariner Training, Certification, and
Equipment Design. Proceedings of the Public Forum on Integrated Bridge
Systems, held March 6-7, 1996 in Vienna, VA. Washington, D.C.: National
Transportation Safety Board.
58.
Sasou K. & Reason J. (1999) Team errors: Definition and taxonomy. Reliability
Engineering and System Safety, 65, 1-9.
59.
Shappell S.A. & Wiegmann D.A. (1997a) A human error approach to accident
investigation: The taxonomy of unsafe operations. International J. Aviation
Psychology, 7(4), 269-291.
60.
Shappell S.A. & Wiegmann D.A. (1997b) A reliability analysis of the Taxonomy of
Unsafe Operations. Aviation, Space, and Environmental Medicine, 68, 620.
61.
Shappell S.A. & Wiegmann D.A. (2000) The Human Factors Analysis and
Classification System-HFACS.
FAA Office of Aviation Medicine Report #
DOT/FAA/AM-00/7. Washington, D.C.: Department of Transportation, Federal
Aviation Administration.
62.
Transportation Safety Board of Canada. (1994) Working Paper on Tankers Involved

in Shipping Accidents 1975-1992.
63.
Transportation Safety Board of Canada [TSB]. (1998) An Integrated Process for

Investigating Human Factors. Montreal: Author.
64.
U.K. P&I Club (1992). The United Kingdom Mutual Steam Ship Assurance
Association (Bermuda) Limited. Analysis of Major Claims.
65.
U.S. Coast Guard (1995) Prevention Through People: Quality Action Team Report.
Washington, DC: U.S. Coast Guard.
66.
Wagenaar W.A. and Groeneweg J. (1987) Accidents at sea: Multiple causes and
impossible consequences. Int. J. Man-Machine Studies, 27, 587-598.
67.
Wenk E. Jr. (1997) Safety, Corporate Culture and Corporate Character. In R.G.
Bea, R.D. Holdsworth, and C. Smith (Eds.), 1996 International Workshop on Human
68.
Wiegmann, D. A., and Shappell, S. A. (1999) A Human Factors Approach to

Accident Investigation. Workshop presented at the Human Factors and Ergonomics
Society 43rd Annual Meeting.
69.
Wogalter M.S., Young S.L., and Laughery K. R. Sr. (Eds.) (2001) Human Factors
Perspectives on Warnings, Vol. 2: Selections from Human Factors & Ergonomics
Society Proceedings, 1994-2000. Santa Monica, CA: Human Factors and
Ergonomics Society.
Page 89 of 141
Houston, Texas
APPENDIX A. SAMPLE INCIDENT REPORTING FORM17
The first step in learning from incidents is to find out that they happened! This appendix
presents a form that can be used by workers to report incidents and near-misses. If your
companys incident database is to contain information from both investigated (i.e., someone
acts on the information in the reporting form and opens an investigation of the incident) and
non-investigated (i.e., the reporting form is the only source of information) incidents, it becomes
important to be able to judge the goodness or validity and completeness of the incident data.
You may want to add a field to your database to show whether the incident was investigated or
not (or the degree to which it was investigated). Another example of rating the completeness of
the incident data is shown in App. E at the bottom of the HFACS Event Data Form. All these
incidents are investigated, and the investigators mark whether they believe they were able to do
a full investigation.
The reporting form in this appendix was developed for the anonymous reporting of near-misses.
As such, it requests information about both the incident and the person reporting the incident.
Information about the reporter is requested in order to infer the likelihood that the reporter has
experience in the incident context (that is, is the reporter likely to know what correct procedures
should have been, and would he/she understand the ramifications of the steps that led to the
incident). Question 6 seeks to aid the assessment of the reliability of the reported information
by finding out whether the person who reported it was actually involved in the incident first-hand.
In this way, one can attempt to weigh the potential value of the information in the incident report.
17
based on a prototype of the International Maritime Information Safety System (IMISS; see Rothblum,
Chaderjian, & Mercier, 2000).
Page 90 of 141
Houston, Texas
Incident Reporting Form
Tell us about yourself
q
q
q
q
q
q
1. In what maritime sector are you involved?

deep draft
inland
off-shore
terminal
recreational
other _____________________________
2. What is your specific job? (e.g., Ordinary

nd
Asst.
Engineer,
drilling
Seaman,
2
supervisor, dock worker, USCG XO)
3. Type of vessel / facility / structure where

you typically work? (e.g., crude oil tanker,
tension leg platform, liquid bulk terminal)
Incident / Potential Hazard

4. Type of incident or hazard (e.g., near-collision,
near-injury; buoy off-station, dangerous work
practice)
5. When did the incident / hazard occur?

Date: ____/____/____
mm / dd / yy
Time: _________ am/pm
6. How were you involved in the incident (or

discovering the hazard)?
directly involved I was an active participant.
indirectly involved I was at the scene, and
saw/heard everything that happened.
observed from a distance I did not see/hear
everything.
not involved I heard about this from
someone else.
other (describe):
q
q
q
q
q
q
q
q
q
dawn
daylight
dusk
night
7. Type of vessels, platforms, structures, or

facilities involved (e.g., lineboat, trawler,
shipyard, jack-up rig):
8. Describe the visibility, weather, and water

conditions.
(e.g., overcast, light rain, no wind, river stage 8
ft. and rising)
9. Where did it happen? (specify the waterway, port, location on ship, etc.):
10. What specific operation(s) was occurring at the time of the incident or hazard? (e.g., normal
bridge/pilothouse watch, normal engineroom watch, cargo transfer operation, ballasting, making
tow, fishing)
Page 91 of 141
Houston, Texas
11. Describe what happened. What were the events which led up to the problem? How was the
problem discovered? What happened next? (be as specific as possible, and put events in the
order in which they happened) For a potential hazard, describe the situation and what could have
happened.
12. What do you think caused the incident or contributed to the events surrounding the incident?
(Consider: decisions; actions; inactions; information overload; communication; fatigue; drugs or
alcohol; physical or mental condition; procedures; policies; design of equipment / ship / facility /
waterway; crew / workers (experience, manning); weather; visibility; equipment failure (why did it
fail?); maintenance.)
13. What went right? How was an accident avoided? (Consider: corrective actions; contingency
plans; emergency procedures; luck.)
14. How can we prevent similar incidents (correct the hazard)? What changes need to be made?
By whom? This block is also for describing Lessons Learned, Safety Tips, and Suggestions.
Page 92 of 141
Houston, Texas
APPENDIX B. DATABASE ITEMS ON GENERAL INCIDENT INFORMATION18

(to be collected in addition to the human error data)
Time, Location, and General Conditions

Date:
/
/
.
mm / dd / yy
Month and Year:
/
.
mm / yy
Day of Week: Su Mo Tu We Th Fr Sa
Local Time:
(24-hr clock)
Location of vessel/platform/facility
Port/Harbor
Terminal
Pier
At Anchor
Restricted waters (marked channel, bay, etc.
Ocean ( 12 nm)
Coastal (< 12nm)
Inland waters
River
Great Lakes
Lake
Bay / Sound / Strait
Offshore Platform in State Waters ( <3nm)
Offshore Platform in Federal Waters ( >3nm)
Other ______________________________
Specific Location
Lat: _____
Long: ______
Type of Aid to Navigation: _____
Waterbody / Waterway name: __________________________
Port / Harbor name: ____________________________
Water depth _______ ft.
Mile Marker_________
18
based on a prototype of the International Maritime Information Safety System (IMISS; see Rothblum,
Chaderjian, & Mercier, 2000).
Page 93 of 141
Houston, Texas
Light
Dawn (morning twilight)
Daylight
Dusk (evening twilight)
Night
Visibility
Visibility: ______ nm
Clear
Cloud cover
Fog
Other
Weather
Calm
Thunderstorm
Rain
Sleet
Snow
Hurricane
Hail
Tornado
Waterspout
Other ________________________
Water Conditions
Salinity (fresh/salt) _________
river stage ____ ft. Rising OR Falling
Flood stage ____ Above gauge level ___ At gauge level _____
Below gauge level ____ Low gauge level ____
Swells:_______ ft. OR sea state _____
Wave height _______ ft; Wave period_________ sec; Wave Direction ______degrees
Current velocity: _________
Current direction__________
Ebb/Flood
Tide: Rising/ Falling Hours since high/low water_______
Obstructed/Floating debris: Describe___________________________
Other ________________________
Wind:
direction: _____
speed: ____ kts.
Number of vessels involved _____

Number of facilities/platforms involved ______
Number of people involved _____
Page 94 of 141
Houston, Texas
Vessel activity, Vessel #1 (check all that apply)

Vessel not underway
Docked
At anchor
Moored
Vessel in port
Docking/undocking
Mooring/Releasing Lines
Maneuvering
Shifting dock-to-dock
Embark/disembark pilot
Embark/disem.
Passengers
Transfer oil/chemicals
Transfer containers
Transfer bulk cargo
Transfer break bulk
Bunkering
Tug escort
Break Bulk Cargo
Other cargo activity _____
Other port activity ______
Fishing
Fishing
Trolling
Trawling
Longlining
Dive
Siening
Dragging
Hauling Gear
Setting Gear
Setting pots/traps
Fish processing
Other fishing __________
USCG / Military
CG patrolling
CG boarding
CG interdiction
CG transit
CG towing
CG - setting buoys
CG - ice breaking
CG assist
CG - helo ops
UNREP Ops
Other CG ops _________
Other military ops _______
Towing
Towing/pushing/hip
Locking
River upbound
River downbound
Make/break tow
Other towing _________
Test & Repair

Testing main engines
Shut-down Engines/Cargo
Equipment testing
Deck Machinery
Electronics
Maintenance
Repairs
Safety Equipment Test
Engine Dept Equipment
Test
Deck Equipment Test
Cargo Monitor Equipment
Test
Transiting
Other Commercial Activity

Production (platform)
Channel inbound
Channel outbound
Logging
Drilling Crew shift
Port inbound
Port outbound
Open waters transit
Great Lakes transit
Meeting
Tank cleaning
Ballasting/deballasting
Lightering
Bunkering/Fueling
Passing/Overtaking
Crossing
Overtaking
Tug escort
Other transit _________
Transfer-related
Cargo Transfer at Anchor
Diving
Touring
Launch Service
Other commercial _______
Other Vessel Activities

Recreational boating
Evacuation
Prep for extreme event
(e.g., hurricane)
Other _______________
Vessel Activity, Vessels #2 and #3 (repeat above)
Page 95 of 141
Houston, Texas
Facility Activity (check all that apply)

Maintenance
Equipment testing
Dry docking
Dock loading/unloading
Dock crane operations
Loading Containers
Staging Containers
Rail Operations
Truck Operations in Yard
Training Workers
Oil Transfer
Dry Bulk Transfer
Break Bulk Transfer
Chemical Transfer
VTS operations
Bridge opening/closing
Lock operations
Other activity ________
Incident/Hazard Event (check all that apply)

Vessel/platform event
Abandonment
Allision
Blowout
Capsize
Collision
Damage to cargo
Damage to environment
Emergency response
Personnel event
Amputation
Asphyxiation
Broken bone
Burn
Crush
Cut
Evasive maneuver
Explosion
Fire
Flooding
Fouling
Grounding
Implosion
Loss of electrical power

Loss of maneuverability
Loss of stability
Material failure
Material failure, diving
Set adrift
Sinking
Other ________________
Drowning
Electrocution
Fall into water
Paralysis
Severe bleeding
Slip/trip/fall
Sprain/strain
Struck by object
Other ________________
Page 96 of 141
Houston, Texas
Type of Vessel: Vessel # 1 (check one)

Tank Vessel
Petroleum
LNG tanker
Chemical
Other tanker ___________
Freight Vessel
Breakbulk
Bulk, Ore
Bulk, Grain
Bulk, Other ___________
Container ship
Tow and Barge
Towing vessel
Tow and barge combo
Barge w/o towboat
Tank barge _________
Freight barge ________
Other barge _________
Tug escort
Composite Unit
Passenger
Passenger, small
Passenger, large
Ferry
Casino boat
Cruise ship
Other Vessels/Platforms
Fishing vessel
Fish processor
Offshore supply vessel
Platform/rig
Workboat (platform)
MODU
OBO
Heavy lift
Dredge
Pilot boat
Other commercial _______
HSC [High speed craft]
Hydrographic survey
Emergency Response
USCG / Military
CG cutter
CG small boat
CG ice breaker
CG buoy tender
CG other _____________
Other military ___________
Recreational
Jet Ski
Sail boat
Powerboat
Other recreational boat ___
(repeat above for vessels #2 and #3)

Type of Facility: Facility # 1 (check one)
Cargo Facilities
Port / pier / dock
Container terminal
Liquid Bulk terminal
Solid Bulk with self loading systems
Bulk cargo facility
Break bulk cargo facility
Oil/chemical facility
Container facility
Other cargo facility _______
Moorings
Single point moorings
Multi buoy moorings
Offshore moorings
Platforms / Drilling
Offshore, Fixed platform
Jack-up rig
Mobile Oil Production Units
Compliant Tower
Tension Leg Platform
Other
deep
water
production/drilling facilities
Lifeboat
Other Facilities
Designated waterfront facility
Shipyard / Dry dock
Marina
VTS
Waterfront Facility
(repeat above for Facility # 2)
Page 97 of 141
Houston, Texas
Person # 1 (check one)

Crew on Deep Draft Ship
Master
Chief mate
1 Mate
2 Mate
3 Mate
4 Mate
AB
OS
Boatswain
Other Deck ____________
Chief Engineer
1 Asst. Eng.
2 Asst. Eng.
3 Asst. Eng.
QMED/Oiler/Wiper
Maintenance Dept.
Other Eng. ____________
Chief Steward
Steward
Chief Cook
Cook
Purser
Waiter
Other Deep Draft _______
Cadet
Trainee
Pumpman
Entertainer
Crew on Towing Vessel

Captain/Pilot (towing)
1st Class operator
Mate (towing)
2nd Class Operator
Deckhand
Engineer
Cook
Other towing __________
Port / Dock Workers
Dock/Harbor Master
Port Engineer
Port Captain
Port Pilot
Port Safety Rep.
Mooring Master
Line Handlers
Hose Crew (Hook up)
Hose Watch
Terminal Person in Charge
Longshoreman
Dock Worker
Gaugers
Pilot
Docking Pilot
Other Port _____________
Shipyard / Inspections
Shipyard worker
Ship chandler
Government Inspector
Class society inspector
Insurer
Surveyor
Other ____________
Page 98 of 141
Crew on USCG/Military
CG Command (CO,XO,EO)
CG bridge crew
CG engineering
CG other ____________
Other military _________
Waterway Personnel
Lock Master
Bridge Tender
VTS
Federal Pilot
State Pilot
Other Waterway ________
Platform / Rig Crew
Operations supervisor
Logging
personnel/contractor
Drilling supervisor/engineer
Platform foreman
Platform crew
Drilling supervisor/engineer
Drill rig personnel
Other _____________
Other Personnel
Vendor
Agent
Visitor
Passenger
Recreational boater
Ship Chandler
Other _______________
Houston, Texas
Person #1s Activity at Time of Incident (check as many as apply)

On-board Activity
Watch / work / OT
Off duty
Navigation watch
Engineering watch
Eng. Maintenance/repair
Deck maintenance
Handling lines
Make/break tow
Load/discharge cargo
Load/discharge
passengers
Meal prep./serve
Other on-board ________
Facility Crew Activity
Describe __________
Platform Crew Activity

Drilling operations
Normal
production
operations
Crane operations
Maintenance
Testing
Evacuation
Off duty
Other platform _________
Shipyard Crew Activity
Dry dock
Ship repair
Logging
Other shipyard _________
Military Crew Activity

Watch / work
Off duty
Collateral duty
Navigation watch
Engineering watch
Maintenance
CG boarding team
CG helo ops
CG setting ATONs
CG launching small boat
CG other ____________
Other military _________
Other _____________
Person #1s Impairments (check as many as apply to incident)

Fatigue
Suspected alcohol intoxication
Suspected drug intoxication
Rx/OTC medications
Illness
High stress
Chronic health condition: __________________
Psychological condition: __________________
Other ____________________
Hearing impairment
Vision impairment
Impaired mobility
Speech impediment
Poor or no English
(repeat Person, Activity, and Impairments for Persons # 2, 3, and 4)
Page 99 of 141
Houston, Texas
Equipment Factors (check as many as apply to the incident)

Vessel Equipment which
failed:
Steering gear
Engine
Generator
Reefers
Radar/ARPA
Electronic chart
Communications system
Other bridge equipment
Winch
Lines
Crane
Platform Equipment which Problem:

failed:
Drill rig
Inoperable
Drill equipment (tanks)
Operated incorrectly
Pipelines & valves
Crane
Cause:
Compressors
Lack of maintenance
Safe shut-down equipment
Improper maintenance
Gas handling units
Design/manufacture error
Hoses
Improper use
Separators
Operator error
Fire detection/control
Inadequate training
Command
&
Control Other _________________
equipment
Electrical
O-ring/gasket
Piping
Other platform ________
Location of equipment
Door/hatch
Bridge
Sanitation
Engine room
Anchor/chain
Deck
Safety & Personal
Protection Gear
Vent system
Firefighting equipment
Tanks
Pump
Fire suppression equipment Other ______________
Valves
Life jackets/preservers
Remote control/monitoring Life boats
systems
Cold water immersion suits
Other vsl equip__________ Hearing protection
Eye protection
Head protection
Chemical protection gear
Respirators
Other safety ___________
Waterway Factors (check all that apply)

high/low river stage
strong current
buoy off-station
nav aid not working
poorly marked waterway
bridge not marked
Last Chart Survey of Area
Where charts corrected for latest
survey (ship on dock)
Additional permitted activities
Debris in water
dangerous crossing
dangerous traffic scheme
channel width/depth not as charted
Narrow channel
dangerous sandbars/shoals
dangerous port design/layout
Nav aid not available such as DGPS off air
Severe weather/waves
Other:__________
Page 100 of 141
Houston, Texas
APPENDIX C. EXAMPLES OF HUMAN FACTORS TAXONOMIES

(in alphabetical order by system name)
Incident/Accident System
Agency
Reference
Aviation Safety Reporting

System (ASRS)
run by NASA for the FAA; used by

commercial aviation (flight crews,
ground crews, and air traffic
control)
Reynard W.D., Billings C.E.,

Cheaney E.S., & Hardy R.
(1986). The Development of the
NASA Aviation Safety Reporting
System. NASA Reference
Publication 1-1-14. National
Aeronautics and Space
Administration.
Confidential Hazardous
Incident Reporting Programme
(CHIRP)
Concerted Action Committee on

Casualty Analysis (European
Commission)
Concerted Action Committee on

Casualty Analysis (1998).
Confidential Hazardous Incident
Reporting Programme. FP4
Waterborne Transport Tasks 21
and 36. (draft)
Human Error Root Cause

Analysis (HERCA)
U.S. Dept. of Energy
Collopy M.T. & Waters R.M.
Human Factors Analysis and

Classification System
(HFACS)
US Navy and Marine aviation
Shappell S.A. & Wiegmann D.A.

(1997) A human error approach
to accident investigation: The
taxonomy of unsafe operations.
International J. Aviation
Psychology, 7(4), 269-291.
Human Factors Checklist, and

others based on the SHELGEMS models
International Civil Aviation

Organization (ICAO);
Transportation Safety Board
(TSB) of Canada;
ICAO (1993). Investigation of

Human Factors in Accidents and
Incidents. Human Factors
Digest No. 7, ICAO Circular 240AN/144. Montreal, Canada:
Author.
US Coast Guard; International

Maritime Organization (IMO)
TSB of Canada (1988). An

Integrated Process for
Investigating Human Factors.
Montreal: Author.
International Maritime
Organization (1998). Role of the
Human Element in Maritime
Casualties. Maritime Safety
Committee, 69th Session,
Agenda Item 13 (MSC/69/13-1).
London: Author.
Page 101 of 141
Houston, Texas
Human Factors Investigation

Tool (HFIT)
prototype in testing
Gordon R., Flin R. & Mearns K.

(2001). Designing a Human
Factors Investigation Tool to
Improve the Quality of Safety
Reporting. Proceedings of the
Human Factors and Ergonomics
Society 45th Annual Meeting,
1519-1523.
Human Performance
Investigation Process (HPIP);
Nuclear Regulatory Commission

and various industries
Paradies M., Unger L., Haas P.,

& Terranova M. (1993)
Development of the NRCs
Human Performance
Investigation Process, Vol. 1-3.
Report # NUREG/CR-5455.
Washington, DC: Nuclear
Regulatory Commission.
TapRooT (a commercial
revision of HPIP)
Paradies M., Unger L., & Busch

D. (199x). Root Cause Tree
Users Manual, Revision 3.
Influence Networks
UK Maritime and Coastguard

Agency (MCA); UK Health and
Safety Executive (HSE);
Organization (IMO)
Embrey, D (1992) Incorporating

Management and Organisational
Factors into Probalistic Safety
Assessment. Reliability
Engineering, 38, 199-208.
BOMEL Limited (2001) FSA of
Bulk Carriers Develop-ment
and Quantification
of Influence Networks
(undertaken for MCA, UK).
Information Safety System
(IMISS)
prototype system developed by a

SNAME working group, sponsored
by US Coast Guard and US
Maritime Administration
Rothblum A., Chaderjian M., and

Mercier K. (2000). Development
and Evaluation of the
Information Safety System
(IMISS). Presented at the
Society of Naval Architects and
Marine Engineers, IMISS Public
Meeting, May 2000.
Marine Safety Reporting

System (MSRS)
Human Factors Group
Safe Marine Transportation

(SMART) Forum of Puget
Sound, 1997. Marine Reporting:
The Development of a
Northwest Marine Safety
Reporting System. Seattle, WA:
Author. (see App. D)
Page 102 of 141
Houston, Texas
APPENDIX D. HUMAN FACTORS TAXONOMY USED BY THE

U.K. MARINE ACCIDENT INVESTIGATION BRANCH
External Bodies Liaison
Non compliance
Communication
Equipment design - manufacturer
Training, skills, knowledge
Working environment/workplace
Incorrect installation/defective equipment
Company & Organisation
Company standing orders inadequate, insufficient, conflicting

Manufacturers instructions
Communication
Pressures - organisational
Inadequate resources
Training, skills, knowledge
Crew Factors
Communication
Management and supervision inadequate
Allocation of responsibility inappropriate
Procedures inadequate
Manning (rotation/watches)
Training
Discipline - crew/passengers
Unsafe working practices
Equipment
Equipment misuse
Equipment not available as needed
Equipment poorly designed for operational use
Equipment badly maintained
Personnel unfamiliar with equipment/not trained in use
Automation means crew not trained in use of manual alternatives
Working Environment
Performance affected by noise

Performance affected by vibration
Performance affected by temperature
Performance affected by humidity
Performance affected by visual environment/visibility
Performance affected by ship movement/weather effects
Poor housekeeping
Layout unsuitable for task
Accommodation
Individual
Communication
Page 103 of 141
Houston, Texas
Competence and skill

Training, inexperience, knowledge
Violation of procedures
Health: drugs/alcohol
Health: medical condition
Domestic issues
Fatigue and vigilance
Perceptual abilities
Poor decision making/information use
Perception of risk
Workload
Page 104 of 141
Houston, Texas
APPENDIX E. ADAPTATION OF HFACS BY THE CALIFORNIA STATE LANDS COMMISSION,

MARINE FACILITIES DIVISION
This appendix shows how the Marine Facilities Division (MFD) of the California State Lands
Commission has adapted the Human Factors Analysis and Classification System (HFACS) to
their specific needs. Below are the guides (job aids) provided to MFD inspectors for their use in
investigating the human factors and other causes of oil spills. The job aids contained herein
include:
HFACS Terminology
HFACS Layer Guides
HFACS Event Data Form (filled out during a spill investigation)
Event Data Form Instructions
Note that the term Layer refers to the slice of cheese being considered: Layer 1 equates to
Unsafe Acts; Layer 2 to Preconditions; Layer 3 to Unsafe Supervision; and Layer 4 to
Organizational Influences. The term Tiers is a short-hand notation for the levels of the
taxonomy being considered. For example, Tier 1 is the Layer or top level classification (e.g.,
unsafe act). Tier 2 would be the next level of classification, such as an Error (unintentional) as
the type of unsafe act. Tier 3 would be the bottom or most specific level of classification, such
as Skill-based errors. Pick-List refers to the items or types of the Tier 3 factor, such as
omitted step in procedure under Skill-Based Errors.
As discussed in the report, MFD adapted HFACS to its needs by incorporating non-HF items
into the investigation paradigm. For example, in addition to Unsafe Acts under Layer 1, MFD
included the category Structural/Mechanical Damage/Failure to help inspectors consider the
equipment factors that contributed to the incident and how human factors caused or
complemented the equipment factors in the evolving oil spill.
Page 105 of 141
Houston, Texas
HFACS Terminology
Accident: An unintended event which results in personal injury, illness, property damage, or
environmental impairment.
Near Miss: An unintended event which has the potential for causing personal injury, illness,
property damage or environmental impairment.
Incident: An unintended event which results in or has the potential for causing personal injury,
property damage, or environmental impairment.
Routine Violation (from Reason, 1990): Two factors, in particular appear to be important in
shaping habitual violations: (a) the natural human tendency to take the path of least effort; and
(b) a relatively indifferent environment (i.e., one that rarely punishes violations or rewards
observance). Everyday observation shows that if the quickest and most convenient path
between two task-related points involves transgressing an apparently trivial and rarely
sanctioned safety procedure, then it will be violated routinely by the operators of the system.
Such a principle suggests that routine violations could be minimized by designing systems with
human beings in mind at the outset.
Error: a generic term to encompass all those occasions in which a planned sequence of
activities fails to achieve its intended outcome, and when these failures and when these failures
cannot be attributed to some chance agency.
Slip and lapse: errors which result from some failure in the execution and/or storage of an
action sequence, regardless of whether or not the plan which guided them was adequate to
achieve its objective. Slips are observable; lapses not.
Mistake: deficiencies or failures in the judgemental and/or inferential processes involved in the
selection of an objective or in the specification of the means to achieve it, irrespective of
whether or not the actions directed by this decision-scheme run according to plan.
Page 106 of 141
Houston, Texas
LAYER 1 GUIDE: UNSAFE ACTS

Unsafe Acts
Errors
Skill-based
Perceptual
Decision
failed to prioritize attention
misjudged distance/rate/time
improper procedure or
inadvertent use of controls
misread dial or indicator
misdiagnosed emergency
wrong response to emergency

omitted step in procedure, or failed to see/hear/otherwise
sense
executed step out-ofsequence.
poor decision
omitted checklist item or
completed check list item outof sequence.
Violations
Routine
Exceptional
(Fill in) Definition: Isolated departure from
(Fill in) Definition: Common or habitual
authority, rules and regulations (taking a
instance of breaking the rules and
shortcut) that is typically not condoned by
regulations (taking a shortcut) that is part of
management
a persons behavior pattern and is often
tolerated by the organization
Structural/Mechanical Damage/Failure
Vessel Control
System
Vessel Structural
Terminal Control
System
Terminal Structural
containment
containment
flange/gasket
flange/gasket
hose
hose
inert gas system
loading arm
loading arm
pipeline
pipeline
pump
piping
shell plating
pump
valve
shell plating
stern tube
valve
Page 107 of 141
Houston, Texas
LAYER 2 GUIDE: PRECONDITIONS FOR ADVERSE EVENTS

Preconditions For Adverse Events
Substandard Conditions of Operators
Adverse Mental States
channelized attention
Adverse Physiological States

impaired physiological state
Physical/Mental Limitations
insufficient reaction time
complacency
medical illness
poor vision/hearing
distraction
physiological incapacitation
lack of knowledge
mental fatigue
physical fatigue
incompatible physical
capability
haste
loss of situational awareness
misplaced motivation
task saturation
Substandard Practices of Operators
Crew Resource Mgt.
Personal Readiness
impaired communications due to language difference
self-medicated
interpersonal conflict among crew
inadequate rest
failed to use all available resources

failure of leadership
misinterpretation of traffic calls
failed to conduct adequate brief
impaired communication/conflict due to cultural difference
Substandard Work Interface
Substandard Design
design
inadequate for
use
design
inadequate for
maintenance
ambiguous instrumentation
inadequate layout or space
substandard illumination
Adverse Environmental
Conditions
Substandard
Maintenance
poorly maintained
equipment
poorly maintained
workspace
poorly maintained
communications
equipment
low visibility
storm
extreme temperature
extreme sea state
substandard
communications equipment
equipment substandard for
job
Page 108 of 141
Houston, Texas
LAYER 3 GUIDE: UNSAFE SUPERVISION

Unsafe Supervision
Inadequate Supervision
Planned Inappropriate
Operations
Failed to Correct a
Known Problem
failed to provide
guidance
failed to provide correct

data
failed to correct
document in error
failed to provide
operational doctrine
failed to provide adequate

brief time
failed to identify an atrisk behavior
failed to provide
oversight
improper manning
failed to initiate
corrective action
failed to provide training

failed to track
qualifications
failed to track
performance
adequacy of operational
procedure or plan
provided inadequate
opportunity for crew rest
Supervisory
Violations
authorized an
unnecessary
hazard
failed to
enforce rules
and regulations
authorized
unqualified
crew
failed to report unsafe

tendencies
LAYER 4 GUIDE: ORGANIZATIONAL INFLUENCES

Organizational Influences
Resource Management
inadequate management of
human resources
inadequate management of
monetary resources
inadequate design and
maintenance of facilities
Organizational Climate
adequacy of organizational
structure
adequacy of organizational
policies
adequacy of safety culture
Page 109 of 141
Organizational Process
adequacy of established
conditions of work
adequacy of established
procedures
adequacy of oversight
Houston, Texas
HFACS Event Data Form

Control #:
OES #:
Facility WO #:
Facility Name:
Vessel Name:
Event Date:
Reviewed by:
Event Time:
____
____
Substance:
Event Type
____
____
Quantity :
Who/What
Incident Causal Factor
Tier 1
Tier 2
gallons
Tier 3
Pick-List
___ Spill
___ Class 3 Violation

Evolution
___ Ballasting/Deballasting
___ Bunkering
___ Fueling
___ Idle
___ Intra-Terminal
___ Intra-Vessel
___ Load/Discharge
___ Maintenance
___ Maneuvering
___ Term/Term
___ Testing
___ Other
Event
___ Arrival
___ Depart
___ Disconnect
___ Hook-up
___ Start-Up
___ Steady
___ Stripping
___ Topping off
___ N/A
Transfer- Related: ___ Yes ___No
Responsible Party: ___ Vessel ___ Terminal ___ Other
Updated Substance:
Notes (Outside Influences; recommendations; contributors):
Quantity:
Page 110 of 141
Layer Completeness 1: Y N
gallons
2: Y N
3: Y N
4: Y N
Date:
Houston, Texas
Event Data Form Instructions:

Top Sections
1. Control #: The control # uniquely identifies the database record for this event. This number
gets assigned at the point of computer data entry. It can be left blank by the
inspector/specialist/analyst completing the form.
2. OES#: This number is assigned by The Office of Emergency Services and can be found on
the Hazardous Materials Spill Report associated with the spill event. For a class 3 violation,
this section should be left blank.
3. Event Date: Enter the month, day and year of the spill or violation event in mm-dd-yyyy
format.
4. Event Time: For spills, enter the time of day that the spill occurred. This time can usually be
found on the Hazardous Materials Spill Report. For class 3 violations, enter the time of day
the violation was noted.
5. Facility WO #: Enter the work order number associated with the facility where the event
occurred.
6. Facility Name: Enter the name of the facility where the event occurred.
7. Vessel Name: Enter the name of the vessel, if any, involved in the event.
8. Reviewed by: Each staff member that produces or reviews the completed form should initial
here. Each form should be produced by and reviewed by a specialist, inspector and human
factors analyst at a minimum before it is ready for data entry.
9. Substance: Enter the product(s) involved. For a violation, enter the product involved only if
the violation occurred during a transfer event.
10. Quantity: For a spill event, enter the amount of product spilled, in gallons. For a violation,
this section should be left blank .
Left Sections
1. Event Type: Check the appropriate blank indicating whether the event is a spill or a class 3
violation.
2. Evolution: Check appropriate blank(s) indicating the type of operation that was in progress
when the event occurred.
3. Event: If the event occurred during a transfer, note the phase of the transfer by checking the
appropriate phase.
Page 111 of 141
Houston, Texas
Center Grid Sections
1. Who/What: Indicate one of the following:
A person or group of persons associated with the incident causal factor identified identify
in column two. Use job titles rather than names; For example, tankerman, dockworker,
TPIC, VPIC, chief mate, operations manager, terminal manager, barge company,
shipping company, etc.
A damaged or malfunctioning facility structure or piece of equipment associated with the
incident causal factor identified in column two.
An environmental condition or event associated with the incident causal factor identified
in column two.
2. Incident Causal Factor: a prevailing condition, act, or omission that contributes to bringing
about an adverse event.
Note: For each row in the grid, The Who/What (column 1) entry and the Incident
Causal Factor (column 2) entry should combine to form a sentence.
Refer to the document Human Factors Analysis and Classification System Marine
Facilities: Definitions for a description of each tier & classification in HFACS. You can also
use the HFACS Tier and Layer Chart for a map of tiers and layers.
3. Tier 1: Enter the most general category for the who/what & incident causal factor listed.
4. Tier 2: Enter the subcategory for the who/what & incident causal factor listed.
5. Tier 3: Enter the subtype for the who/what & incident causal factor listed.
6. Pick-List: For subcategories and subtypes that have pick-list items listed in the HFACS
Layer Guides, select a specific factor from the list that best describes the incident causal
factor.
Bottom Sections
1. Transfer-Related: Check to indicate whether the spill or class 3 violation is transferrelated.
2. Responsible Party: For spills, check to indicate the party responsible for the release.
3. Layer Completeness: For each layer, circle yes if you believe all causal factors that
contributed to the event were identified as a result of the inquiry. Circle no if you believe
other causal factors could have been identified had you been able to get more information
during the inquiry.
4. Updated Substance: If the product determined to be involved in the event changes over
the course of the inquiry, note the change here.
5. Quantity: enter the final estimate of the amount of product spilled.
6. Date: Enter the date the HFACS EVENT DATA FORM is completed.
7. Notes (Outside Influences; recommendations): If groups outside the terminal-vessel
organizations contributed to the event, note it here. Examples of outside influences include
government agencies, local public political pressure and economic pressures. Also, specific
comments about the event, or recommendations that result from the inquiry should be
added here.
Page 112 of 141
Houston, Texas
APPENDIX F STOLT-NIELSEN ACCIDENT/INCIDENT INVESTIGATION

SHIP:
DATE, TIME & PLACE OF INCIDENT: CONT
LOCATION ONBOARD:
LOSS DESCRIPTION: (check all applicable)
___ DAMAGE
___ Property Damage
___ System/Equipment Damage
___ Cargo Loss/Damage/Contamination
___ Ship Damage
___ POLLUTION
___ Contained
___ Water Pollution
___ Air Pollution
___ Personnel Exposure
___ GROUNDING
___ INJURY
___ Death
___ Serious
___ Minor
___ ADMINISTRATIVE
___ Operational Delay
___ Regulatory Violation
___ Inspection Deficiency
___ FIRE/EXPLOSION
___ COLLISION
___ PIRACY/THEFT
___ OTHER (describe):________________________________________________________________

LOSS SEVERITY: (check)
__ Major
__ Serious
__ Minor
PROBABILITY OF OCCURRENCE: (check)

__ Frequent
__ Occasional
__ Seldom
DESCRIPTION OF THE EVENT:
IMMEDIATE CAUSES:
event?)
(What substandard actions and conditions caused or could cause the
UNDERLYING CAUSES: (What specific personal or job factors caused or could cause the event?)
REMEDIAL ACTIONS/RECOMMENDED CHANGES: (To control the causes listed)
Rev 9/9/99
(Contd...)
Page 113 of 141
Houston, Texas
Page 2:
RESPONSIBILITY (For remedial action/changes):
CODING OF IMMEDIATE CAUSES: (check all applicable)

SUBSTANDARD CONDITIONS:
__ 1. Inadequate guards or barriers
__ 2. Inadequate or improper protective equipment
__ 3. Defective tools, equipment or materials
__ 4. Congestion or restricted action
__ 5. Inadequate warning systems
__ 6. Fire or explosion hazards
__ 7. Poor housekeeping; disorder
__ 8. Hazardous environmental conditions
__ 9. Noise exposure
__10. Radiation exposure
__11. High/low temperature exposure
__12. Inadequate or excess illumination
__13. Inadequate ventilation
__14. Installation failure
__15. Equipment failure
__16. System failure
SUBSTANDARD ACTIONS:
__ 1. Operating equipment without authority
__ 2. Failure to warn
__ 3. Failure to secure
__ 4. Operating at improper speed
__ 5. Making safety devices inoperable
__ 6. Removing safety devices
__ 7. Using defective equipment
__ 8. Using equipment improperly
__ 9. Failing to use proper personal protective
equipment
__10. Improper loading
__11. Improper placement
__12. Improper lifting
__13. Improper position for task
__14. Servicing equipment in operation
__15. Horseplay
__16. Drugs/Alcohol
__17. Fatigue
__18. Procedural error
CODING OF UNDERLYING CAUSE: (check all applicable)

JOB FACTORS:
__ 1. Inadequate Leadership/Supervision
__ 2. Inadequate Personnel
__ 3. Inadequate Engineering
__ 4. Inadequate Maintenance
__ 5. Inadequate Tools/Equipment
__ 6. Inadequate Work Standards
__ 7. Wear and Tear
__ 8. Abuse or Misuse
PERSONAL FACTORS:
__ 1. Fatigue
__ 2. Inadequate capability
__ 3. Lack of knowledge
__ 4. Lack of skill
__ 5. Stress
__ 6. Improper motivation
SUBMITTED BY (print name): _____________
DATE OF REPORT: ____________
SAFETY COORDINATOR SIGNATURE: ______________________________________

SHIPS CAPTAIN CONCURRENCE: _________________________________
FLEET MANAGERS APPROVAL: _________________________________
(Please provide any additional comments on additional sheets and attach.)
Page 114 of 141
Houston, Texas
Appendix G. Sample Taxonomies for Investigating for Fatigue,

Communications Problems, and Inadequate Skills and Knowledge
The U.S. Coast Guard Research and Development Center has performed three studies of the
roles of specific human errors in marine casualties: fatigue, communications problems, and
inadequate skills and knowledge. These studies focused on critical casualties, that is,
casualties which resulted in significant damage to the vessel, cargo, or the environment, or
which resulted in significant injury to personnel.
The first two pages (following this introduction) represent data that were found useful for
determining whether or not fatigue played a role in the incident (McCallum, Raby, & Rothblum,
1996). Earlier studies (Marine Transportation Research Board, 1976; National Research
Council, 1990) had found fatigue to be the number one concern of mariners. This study
corroborated that concern, finding that fatigue played a role in 16% of critical vessel casualties
(groundings, collisions, allisions, etc.) and in 33% of critical personnel injuries (lacerations,
amputations, crushings, etc.). The types of data reflected in the Fatigue Information form focus
on scientifically validated causes, symptoms, and effects of fatigue. This form is recommended
for use by organizations interested in controlling crew fatigue and reducing the casualties
fatigue can cause.
As mentioned in Section 6 in the discussion of statistical analysis, the data from the fatigue
investigations were submitted to a multiple regression analysis. The analysis showed that the
Coast Guard could significantly streamline its initial investigation of fatigue to just three
questions: how many hours did the person sleep during the 24-hour period prior to the casualty;
how many hours did the person work during the 24-hour period prior to the casualty; and how
many fatigue symptoms did the person experience while on duty prior to the casualty? The
answers to these questions are put into an equation, as shown on the Fatigue Investigations
Worksheet (third page of this Appendix), and the result was found to correctly identify whether
fatigue was a causal factor in 80% of the cases. Please note that the Fatigue Index Score is
offered only as an example of the use of statistical analysis of incident data, and it is not
recommended for use by anyone other than the Coast Guard. The selection and weightings of
the variables in the Fatigue Index equation may well reflect the somewhat guarded atmosphere
of a USCG investigation (i.e., since ones license could be on the line, one is less likely to be
open and forthright about ones condition and everything that occurred). We would anticipate
that in the open, blame-free culture recommended for safety-conscious companies, a very
different set of weighting factors (and maybe a different set of variables altogether) would apply.
Page 115 of 141
Houston, Texas
Inadequate communications was found to play a role in 18% of critical vessel casualties and in
28% of critical personnel injuries (McCallum, Raby, Rothblum, Forsythe, Slavich, & Smith, 2000,
unpublished). The communications investigation procedures provided in this appendix are
based on a scientific model of communications (described in the Instructions for Investigating
Communications Problems in Marine Casualties). The gist of the procedure is to identify under
what types of situations communication is normally required a list that can easily be modified
to suit any companys operations then to determine whether necessary communications were
absent or ineffective and why. As described in Section 6, this investigation protocol was not
only sufficient to identify casualties in which inadequate communications were a contributing
cause, it also supported a meta-analysis to illuminate underlying causes of the communications
failures. In addition, it was found that the five screening questions (Step 1 on the
communications investigation form) captured 76% of the casualties in which communications
errors were involved, making it a simple and effective way to determine whether the complete
communications investigation needs to be performed.
The final set of investigation forms assess whether skill and knowledge limitations may have
contributed to the incident. A study of skill and knowledge errors showed that they played a role
in 22% of critical vessel casualties and in 32% of critical personnel injuries (McCallum,
Forsythe, Raby, Barnes, Rothblum, & Smith, 2000, unpublished). Step 4 of the Mariner Skill &
Knowledge Limitations Investigation Screening form essentially seeks to distinguish errors of
knowledge and skilled performance (Decision Errors in HFACS) from slips and lapses (Skillbased and Perceptual errors in HFACS) and from violations. If it appears that the individual did
not intentionally break a rule (violation) and failed to perform an action properly given an honest
effort, it is assumed that a lack of knowledge or skill is the cause. The remainder of the skill
and knowledge forms provide lists of common skill and knowledge areas relevant to different
types of activities (bridge, deck, engineering, and safety & emergency operations) obviously,
these forms can be modified to suit any type of operation. The crux of the forms is to identify
the type of error (skill or knowledge that was lacking) and to determine whether the error
resulted from a lack of training/experience on the part of the person or from a poorly designed
standard operating procedure (SOP) or policy. Note that while this analysis will determine
whether insufficient training, experience, or a poor SOP was a contributing cause, that the best
solution to such problems may turn out to be other types of interventions, such as a redesign of
equipment, tasks, or work environment (see Sec. 8).
Page 116 of 141
Houston, Texas
Fatigue Investigation Form

Form should be completed for each individual who was directly linked to the casualty.
Section 1: Casualty Day
1. Experience of involved individual
year(s)
month(s)
a) in the industry
b) with this company
c) in present job or position
d) on present vessel
2. Individuals activity at time of casualty
3. Specific location on vessel
4. Number of hours on duty at time of casualty
5.
6. Time and last meal (breakfast, lunch, dinner) taken prior

to casualty
7. Was the schedule on day of casualty different from usual?
No break
Yes, when
Any break (awake) prior to casualty?
No
a.
What was the difference?
b.
How long had you been on this schedule?
Yes
8. Did you experience any of these factors during the last 24 hours prior to the casualty? (Check all that apply.)
Disruptive ship vibrations

Disruptive ship motion
High noise level
Stormy weather
Cold temperature
Hot temperature
Boredom
None
High stress
Other
Demanding task
9. Did you experience any of the following while you were on duty prior to the casualty? (Check all that apply.)
Forgetful
Difficulty keeping eyes opened
Difficulty operating equipment
Sore muscles
Distracted
Less motivated
None
Other
Desire to sit or lay down
Section 2: Working Schedule

10. Number of days on tour at time of casualty (including shipyard)
11. Please shade the days on which you had 24 hours off in the previous 30 days
30 29
28
27 26
25
24
23 22
21
20
19
18
17
16
15 14
13
12
11 10
Day of
Casualty
(over)
Page 117 of 141
Houston, Texas
Fatigue Information - Side 2

12.
very
very
low
high
1 2 3 4 5
Work
Rec
Normal
Schedule
Sleep
0000
m id n ig h t
0200
0400
0600
08 00
1000
1400
2 pm
1200
noon
1600
4 pm
1800
6 pm
20 00
8 pm
2200
10 pm
2359
Workload level
Fatigue level
Sleep quality
From the time of casualty, trace back the work, recreation, and sleep periods for the last 72 hours (3 days) prior to the
casualty. Sleep also includes naps. Please mark time of casualty with vertical line.
Work
Rec
13.
very
very
low
high
1 2 3 4 5
Sleep
0000
m id n ig h t
Day
of
Casualty
14.
Work
Rec
Day
1
Prior to
Casualty
Sleep
15.
Work
Rec
Day
2
Prior to
Casualty
Sleep
0400
0600
08 00
1000
1400
2 pm
1200
noon
1600
4 pm
1800
6 pm
20 00
8 pm
2200
10 pm
2359
Workload level
Fatigue level
Sleep quality
0200
0400
0600
08 00
1000
1400
2 pm
1200
noon
1600
4 pm
1800
6 pm
20 00
8 pm
2200
10 pm
2359
Workload level
Fatigue level
Sleep quality
very
very
low
high
1 2 3 4 5
0000
m id n ig h t
0200
0400
0600
08 00
1000
1400
2 pm
1200
noon
1600
4 pm
1800
6 pm
20 00
8 pm
2200
10 pm
2359
Workload level
Fatigue level
Sleep quality
very
very
low
high
1 2 3 4 5
Work
Rec
Day
3
Prior to
Casualty
very
very
low
high
1 2 3 4 5
0000
m id n ig h t
16.
0200
Sleep
0000
m id n ig h t
0200
0400
0600
08 00
1000
17. Are there any company or union policies regarding work

hour limits?
No Yes
1400
2 pm
1200
noon
1600
4 pm
1800
6 pm
20 00
8 pm
2200
10 pm
2359
Workload level
Fatigue level
Sleep quality
18. In the involved individuals opinion, was fatigue a contributing factor to this
casualty?
Maximum hours:
No Yes
Why:
Section 3: For the Investigator Only

19. What was the decision/action that was considered improper given the existing circumstances?
Failure to secure equipment

Failure to
important
notice
something
Failure to take action at proper

time
20. a) Was alcohol/drug testing done?

b)Was the result:
Positive
Failure to recognize code/symbol

Failure to decide on an action
Forgetting to accomplish task
Prone to take risks
Yes No
Negative No result
22. On a scale of 1 to 5, do you feel that the mariner gave you true
and accurate information?
1
2
not at all
true & accurate
Erroneous judgment of situation

Erroneous calculations
Improper procedures
Slow reaction to circumstance
Other
21. In your opinion, was fatigue a contributing factor to this casualty?
No
Yes
Why
Additional Comments
5
extremely
true & accurate
Page 118 of 141
Houston, Texas
FATIGUE INVESTIGATION WORKSHEET

The formula
s(21.4) + wh(6.1) - sh(4.5) = Fatigue Index Score
has been established to
determine if fatigue may have affected an individual involved in a marine casualty. It should
only be used for examining cases in which direct human error contributions are present, i.e.,
those situations which involve an individuals decisions, actions or inactions as casual factors
occurring immediately before the casualty.
Using the Formula
s(21.4)
+
wh(6.1)
sh(4.5)
=
Fatigue
Index
Score
s = total number of fatigue symptoms (0-7) experienced while on duty before the casualty
wh = total number of hours worked in the last 24 hours (to the nearest tenth)
sh = total number of hours slept in the last 24 hours (to the nearest tenth)
1. Interview the individual whose errors directly contributed to the casualty (or another person
who can verify the information) to determine the total number of fatigue symptoms (s) listed
below, if any, that the individual experienced while on duty prior to the casualty.
Fatigue Symptoms (s)
- forgetful
- distracted
- sore muscles
- less motivated
- desire to sit or lay down

- difficulty keeping eyes opened
- difficulty operating equipment
2. Obtain the individuals total work hours (wh) and sleep hours (sh) for the 24 hour period
before the casualty. Determine the Fatigue Index Score using the formula.
3. If the Fatigue Index Score is greater than 50, assume that fatigue was a contributing cause
of the casualty. Our research has shown that this formula will produce correct results 80% of
the time.
Example: At 0130, the F/V SEA MONKEY ran aground while returning to port. The mate on
watch reported that he had slept 3.2 hours and worked 18.6 hours in the 24 hours preceding the
casualty. He also said that while on watch before the casualty, he had difficulty keeping his
eyes opened and felt distracted. The equation for this casualty would then read:
s(21.4) + wh(6.1) - sh(4.5) = Fatigue Index Score
2(21.4) + 18.6(6.1) - 3.2(4.5) = Fatigue Index Score
141.9 = Fatigue Index Score
Since the results are greater than 50, fatigue is assumed as a contributing factor to the
casualty.
Page 119 of 141
Houston, Texas
Instructions for Investigating Communications Problems in Marine Casualties

These instructions provide an aid in using the Communications Problems Screening and Investigation Procedures to
investigate communications problems in vessel and personnel injury casualties.
Background
These procedures were developed as part of a
Coast Guard study of how best to investigate and
report on communications problems. As part of
that study, a general model of communications
problems was developed, shown in the adjacent
figure. This model divides communications into
four Communications Processes (prepare and
send message, message transmission, receive
and interpret message, and act on message) and
four corresponding Communications Problem
Areas.
The model further identifies seven
Contributing Factor Areas that can cause or
contribute to communications problems.
Basis
Contributing Factor Areas

Knowledge or
Experience
Procedures
Performance
Assumptions
Communications Processes
Prepare
&
Send
Message
Message
Transmission
Receive
&
Interpret
Message
Vessel
Operations
Act
Problems Preparing
And Sending
Problems with
Message
Problems Receiving
And Interpreting
Problems
Acting on
Transmission
Messages
Messages
Messages
Investigation procedures based on this model
were developed and then applied by Investigating
Communications Problem Areas
Officers as part of the study. During the study,
investigators screened casualties to identify those
that required effective communications to support
Environment
Communication
Management &
safe operations. Of those casualties identified as
Government
Equipment
Regulations
requiring effective communications, 76 percent
Contributing
Factor
Areas
were
subsequently
found
to
have
a
communications problem that contributed to the
casualty. Following their initial screening of cases, investigators conducted in-depth investigations and analyses of
selected casualties to identify specific communications problems and contributing factors. Investigating Officers
were able to use the procedures to reliably identify communications problem areas and specific factors contributing
to the casualties. Overall, the study found that 18 percent of critical vessel casualties and 28 percent of critical
personnel injuries had a communications problem that contributed to the casualty.
Instructions
Step 1 is conducted to identify if there was a potential for a communications problem to have contributed to the
casualty. This step identifies casualties where there is a 76 percent probability that ineffective, inappropriate, or a
lack of communications contributed to the casualty, according to the results of the research study.
Step 1: Review the five conditions, check any that apply, and identify the type(s) of communications that should be
further analyzed (vessel-vessel, bridge-pilot, vessel-shore authority, crew-crew, and vessel-shore worker).
The remaining steps call for a further investigation of the specific communications causes that contributed to the
casualty. Complete Step 2 to identify the specific communications causes, if any. Complete Step 3 to document
your conclusions regarding the type of communications that contributed to the casualty. Use Step 4 as an aid in
investigating and reporting any communication types identified in Step 3.
Step 2: For each communication type identified in Step 1, consider the actions in which ineffective, inappropriate,
or a lack of needed communications could have contributed to the casualty.
Step 3: Check the types of communications that likely contributed to this casualty and complete Step 4 for each
type checked.
Step 4: For this step, it will typically be necessary to contact individuals involved in the casualty to determine the
events leading up to the casualty, specific communications problems that occurred, and the factors that
contributed to these problems.
Page 120 of 141
Houston, Texas
Communications Problem Screening and Investigation Procedures

Please refer to the Instructions for Investigating Communications Problems in Marine Casualties for a
summary of the background and basis for these procedures, as well as general instructions for their use.
Step 1: Was there a potential for a communications problem contributing to the casualty?
all that apply, and note the corresponding communication
Review the following casualty conditions, check
type(s) for further review in Step 2. If no conditions apply, communications were likely not required in the
situation.
q
q
q
q
q
Casualty Condition
Two or more vessels were involved in this casualty.
There was a pilot (other than a member of the vessels crew) responsible for
navigation of the ship.
The vessel was navigating in an area under the supervision of a VTS operator,
a bridge tender, a lockmaster, or a light operator.
Two or more crewmembers who were directly involved in this casualty were
working together, or this casualty could have been prevented if someone had
shared additional information with another crewmember.
The casualty occurred during coordination of activities between the vessel and
shore-based personnel (e.g., dock worker, crane operator, or vessel agent).
Communication Type
Vessel-Vessel
Bridge-Pilot
Vessel-Shore Authority
Crew-Crew
Vessel-Shore Workers
Step 2: What specific communications actions contributed to the casualty?
Check all actions in which ineffective, inappropriate, or a lack of needed communications may have contributed
to the casualty. Note any other causes not listed. If any potential causes are identified, continue with Steps 3 and
4.
Vessel-Vessel Communication Problems
Vessel communication using a VHF radio system
Vessel communication using visual signals
Vessel communication using sound signals
Vessel communication using some other means
Other:
Bridge-Pilot Communication Problems
Pilot request for vessel and situation information
Pilot brief to bridge crew on operating conditions
Bridge crew warned pilot of equipment malfunction
Pilot update to bridge crew on change in plans
Pilot brief to bridge crew on navigation plan
Crew update to pilot of change in situation
Other:
Vessel-Shore Authority Communication Problems
Vessel call to shore authority
Vessel statement of intentions to shore authority
Shore authority advisory to vessel of situation
Shore authority acknowledgement of vsl
intentions
Other:
Crew-Crew Communication Problems
Use of direct and verbal conversation
Use of communications devices
Use of hand signals
Use of written communications
Other:
Vessel-Shore Worker Communication Problems
Use of direct and verbal conversation
Use of communications devices
Use of hand signals
Use of written communications
Other:
No Potential Communication Problems Identified
Further investigation failed to support communications as a causal factor
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
q
Step 3: Which of the following types of communication contributed to this casualty?
Based on the response to Step 2, check

the types of communication, if any, that likely contributed to this
casualty and complete Step 4 for each type checked.
Page 121 of 141
Houston, Texas
q
q
q
q
q
q
Vessel-Vessel Communications
Bridge-Pilot Communications
Vessel-Shore Authority Communications
Crew-Crew Communications
Vessel-Shore Worker Communications
N/A--no communication problems identified
Step 4: What specific communications problems and factors contributed to this casualty?
For each type of communication checked in Step 3, check

all communications problems that contributed to the
casualty. For each problem identified below, list at least one contributing factor from the list below by indicating
its corresponding identification number (#1-41). For example, Did not request information3 , 15 , 28.
Communications Process
Communications Problem
Prepare & Send Message
Did not communicate ................................................................................... ___
___
___
(includes
spoken
and
written
communications,
hand and sound signals)
Communicated ambiguous, incorrect, or incomplete information ............... ___

Did not question others actions or assert own interpretation of situation ... ___
___
___
___
___
Did not request information .......................................................................... ___

___
Did not send information in a timely manner ...............................................
___
Sent different information than intended.......................................................
___
___
___
___
___
___
Message was not transmitted ...................................................................... ___
___
___
Message was interrupted ............................................................................. ___

Message was incomprehensible ................................................................. ___
___
___
___
___
Did not monitor communications ................................................................. ___
___
___
Did not listen to complete message ............................................................. ___

Did not acknowledge information reception ................................................ ___
___
Did not interpret the information correctly ....................................................
___
Did not verify the validity or accuracy of the information..............................
___
___
___
___
___
___
___
___
Took no action............................................................................................... ___
___
___
Action was not in accordance with agreement ............................................ ___
___
___
___
___
___
Message Transmission
Receive
Message
&
Act on Message
Interpret
Others:
Page 122 of 141
Contributing Factor
(see 1 41 below)
Houston, Texas

Knowledge or Experience
Assumptions
1. Improper use of signaling techniques (hand, light, flag)
22. Assumed that there was no need to communicate
2. Improper use of standard marine technical vocabulary
23. Assumed lack of response as implicit (silent) confirmation
3. Inadequate knowledge of company procedures or policies
24. Assumed incorrectly that other party knew the information
4. Inadequate knowledge of correct communications protocol
25. Assumed that individual in charge recognized the problem
5. Inadequate knowledge of regulatory requirements
26. Confusion regarding who was communicating
6. Limited English skills or knowledge
27. Confusion regarding who was in charge of situation
7.
28. Incorrect interpretation of the situation
Language difficulty (e.g., enunciation, strong accent)
8. Lack of common language
29. Other:
9. Other:
Environment
Procedures
30. Excessive ambient noise
10. Did not carry communications equipment on person
31. Excessive electronic or atmospheric disruption of signal
11. Did not operate the communications equipment correctly
32. Excessive traffic (i.e., too many users, too lengthy) on the
assigned communications channel
12. Selected incorrect communications channel or frequency
33. Other:
13. Selected incorrect communications device
Communications Equipment
14. Other:
34. Communications equipment malfunction
Performance
15. Distracted or interrupted by other tasks (e.g., high workload)
16. Forgot information or intended actions
17. Tired or sleepy
18. Individual not at work station
19. Not willing to challenge authority
20. Not willing to communicate
21. Other:
35. Communications equipment not available

36. Communications equipment turned off
37. Other:
Management and Government Regulations
38. No regulatory requirement to communicate
39. Not part of individuals job description or responsibilities
40. Inadequate Standard Operating Procedures
41. Other:
Page 123 of 141
Houston, Texas
COMMUNICATIONS REPORTING FORM

Section 1. Reference Information
1. Name of Investigator
3. Case Number
2. Office
4. Vessel Name
Section 2. Individuals Contacted

5. Name & Job Position of Individual
Name
Job
6. Individual contacted

Name
Job

Name
Job
7. Individuals Fluency in English Language
By phone
By deposition
In person
Unable to contact
a. No English
b. Standard maritime phrases
c. Limited conversation
d. Fluent
e. Unknown
By phone
By deposition
In person
Unable to contact
a. No English
d. Fluent
e. Unknown
By phone
By deposition
a. No English
In person
Unable to contact
d. Fluent
e. Unknown
Section 3. Communications Contribution

This form relates to which type of communications problem? (check only one use separate form for each type)
bridge-pilot
vessel-vessel
crew-crew
vessel-shore workers
vessel-shore authority
14. Were communications or coordination between (the groups checked above) advisable during the events leading up to the casualty?
Yes
No
If Yes, briefly describe activities (e.g., course change, passing, line handling, cargo transfer, locking) or situation requiring
communications:
15. Describe any needed communications that were either not done or done ineffectively.
16. In the Investigators opinion, were communications a contributing factor to this casualty?
Yes. Describe the specific type of communications problems checked above (i.e., Who? What? How?).
No. Indicate the apparent cause(s) of the casualty

If No is checked, proceed with Section 4, but skip Section 5 (Communications Analysis), otherwise complete both sections 4 & 5.
Section 4. Conclusions and Comments

17. On a scale of 1 to 5, do you feel that the individual(s) gave you true and accurate information?
Not at all true
& accurate
N/A since nobody
was contacted
Individuals initials ______

Extremely true
& accurate
1.
2.
3.
4.
5.
1.
2.
3.
4.
5.
1.
2.
3.
4.
5.
18. To what extent was there a discrepancy in the information received from the various individuals contacted?
N/A since only 1 individual was
contacted
Complete
disagreement
1.
Complete
agreement
2.
3.
4.
5.
19. Safety recommendations to prevent similar communications-related casualties
20. Additional Comments
Page 124 of 141
Houston, Texas

Section 5. Communications Analysis
Please complete this section if bridge-pilot communications is a contributing factor to this casualty. Check all bridgepilot
communications process problems that apply. For each process problem identified, list at least one contributing factor from the
list below by indicating its corresponding identification number (#1-41). (e.g., Did not request information 6 , 8 , 11)
Contributing Factors
21. Process Problems

Communications Process
Process Problems
(see 1 41 below)
Prepare & Send Message
Did not communicate .............................................................................
___
___
___
___
(includes spoken and written

communications)
Communicated ambiguous, incorrect, or incomplete information ...........
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
___
Did not question others actions or assert own interpretation of situation.

Did not request information ....................................................................
Did not send information in a timely manner ..........................................
Sent different information than intended .................................................
Message Transmission
Message was not transmitted .................................................................

Message was interrupted .......................................................................
Message was incomprehensible ............................................................
Receive & Interpret Message
Did not monitor communications ............................................................

Did not listen to complete message ........................................................
Did not acknowledge information reception ............................................
Did not interpret the information correctly ...............................................
Did not verify the validity or accuracy of the information .........................
Act on Message
Took no action .......................................................................................

Action was not in accordance with agreement ........................................
Others:
Contributing Factors
Knowledge or Experience
Assumptions
1. Improper use of signaling techniques (hand, light, flag, Morse)
22. Assumed that there was no need to communicate
2. Improper use of standard marine technical vocabulary
23. Assumed lack of response as implicit (silent) confirmation
3. Inadequate knowledge of company procedures or policies
24. Assumed incorrectly that other party knew the information
4. Inadequate knowledge of correct communications protocol
25. Assumed that individual in charge recognized the problem
5. Inadequate knowledge of regulatory requirements
26. Confusion regarding who was communicating
6. Limited English skills or knowledge
27. Confusion regarding who was in charge of situation
7.
28. Incorrect interpretation of the situation
Language difficulty (e.g., enunciation, strong accent)
8. Lack of common language
29. Other: ____________________________________________
9. Other:_____________________________________________
Environment
Procedures
30. Excessive ambient noise
10. Did not carry communications equipment on person
31. Excessive electronic or atmospheric disruption of signal
11. Did not operate the communications equipment correctly
32. Excessive traffic (i.e., too many users, too lengthy) on the
assigned communications channel
12. Selected incorrect communications channel or frequency
33. Other: ____________________________________________
13. Selected incorrect communications device

14. Other: ____________________________________________
Communications Equipment
34. Communications equipment malfunction
Performance
15. Distracted or interrupted by other tasks (e.g., high workload)
35. Communications equipment not available

36. Communications equipment turned off
16. Forgot information or intended actions
37. Other: ____________________________________________
17. Tired or sleepy
Management and Government Regulations
18. Individual not at work station
38. No regulatory requirement to communicate
19. Not willing to challenge authority
41. Not part of individuals job description or responsibilities
20. Not willing to communicate

21. Other: ____________________________________________
42. Inadequate Standard Operating Procedures

41. Other: ____________________________________________
Page 125 of 141
Houston, Texas
MARINER SKILL & KNOWLEDGE LIMITATIONS INVESTIGATION SCREENING

Step 1: Human Factors Involvement
Did at least one persons actions, inaction, or decisions directly contributed to the casualty or its severity?
<(6 Go to Step 2.
12 Human factors are likely not involved and further investigation of mariner skill and knowledge limitations is unwarranted.
Step 2: Contributing Individuals
List the names and job positions of up to three persons whose actions, inaction, or decisions most directly contributed to the
casualty. For each person, identify the general area(s) of vessel operations that contributed to the casualty, then go to Step 3.
Mariners Name
Job Position
Vessel Operations Contributing to Casualty

(Bridge, Deck, Engineering, Safety & Emergency)
1.
2.
3.
Step 3: Contributing Activities

Briefly describe each persons actions, inaction, and/or decisions that contributed to the casualty.
Mariner 1:
Mariner 2:
Mariner 3:
Step 4: Potential for Skill and Knowledge Limitations

For each mariner, respond to the following questions. If possible, interview the
mariner(s) in-person or by telephone to address these questions.
a.
Did this persons action or inaction result in their knowing violation of an applicable
law, rule, policy or standard operating procedure?
Mariner 1
Mariner 2
Mariner 3
<(6
12
<(6
12
<(6
12
<(6
12
<(6
12
<(6
12
If NO for any involved mariner, go to b.

If YES for all involved mariners, end report. This casualty is likely a violation, not
the result of skill and knowledge limitations.
b.
Has each person successfully demonstrated the contributing activities many times
before under similar circumstances and within the last five years?
If NO for any involved mariner, go to Step 5.
If YES for all involved mariners, end report. This casualty is likely either a slip or a
lapse, not the result of skill and knowledge limitations.
Step 5: Completion of Operations Form(s)

Complete applicable operational area investigation form(s) for Bridge, Deck, Engineering, and/or Safety &
Emergency Operations, for each mariner with NO answers to Questions 4a and 4b.
Page 126 of 141
Houston, Texas

Bridge Operations Mariner Skill & Knowledge Limitations
Please complete a separate copy of this form for each person whose bridge activities contributed to the casualty.
Step 5.1: Maritime Work History of Contributing Mariner

1. Mariners name:
2. Job position at time of casualty:
3. _____ years _____ months in this industry.
4. _____ years _____ months with this company.
5. _____ years _____ months in present position.
6. _____ years _____ months on present vessel or facility.
7. _____ years _____ months on present route.
8. Current licenses/documents (N/A if not applicable):
Step 5.2: Mariners Actions, Inaction, or Decisions Contributing to the Casualty

9.
Briefly describe how this persons specific bridge actions, inaction, or decisions contributed to the casualty:
Now, check
all bridge activities (10-17) that directly contributed to the casualty.
10. Did changing bridge watch activities contribute?
YES Check all activities that apply.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
Check and acknowledge passage plan, orders, and

special information
Assess traffic and weather conditions
11. Did visual monitoring and lookout activities contribute?
Instruct Lookout as to duties and ensure Lookout is
prepared to assume the watch
Maintain lookout to detect objects, traffic, or
navigational aids and assess visibility
12. Did collision avoidance activities contribute?
q a.
q b.
Adjust and operate radar/ARPA

Monitor radar/ARPA and radar contacts
13. Did grounding avoidance and navigation contribute?
q a.
qb
Establish a passage plan based on navigation

information and knowledge of are
Determine vessel position using available systems
14. Did shiphandling activities contribute?
q a.
q b.
Maneuver in accordance with sea/river/weather

conditions
Maneuver in accordance with vessel and/or tow
handling characteristics
15. Did bridge communications contribute?
q c.
q d.
q a.
q b.
Communicate and coordinate effectively among the

vessels crew (Bridge, Engine, and Deck)
Communicate and coordinate between the bridge
watch team and the federal/state pilot
16. Did port or anchor watch activities contribute?
Inspect for leaks, loose or weak mooring lines, and
smoke or fire
17. Did a bridge activity not listed above contribute?
NO Go to 15.
Maneuver vessel in accordance with conning orders

Maneuver vessel during docking, anchoring, and
mooring
Manage and coordinate assist vessels
q a.
q b.
NO Go to 14.
Calculate course changes based on navigation

information, local conditions, and local regulation
Check and update navigation charts and publications
q c.
q d.
q e.
NO Go to 13.
Assess collision threat and determine avoidance

maneuver
Recognize and apply COLREGS
q c.
q d.
NO Go to 12.
Determine type, aspect, and relative motion of other

vessels
Receive and verify reports of visual contact
q c.
q d.
NO Go to 11.
Check status of ships equipment

Ensure that watch is relieved
NO Go to 16.
Interpret and reply to signals (flag signals, flashing

light, and ships whistle)
Establish and maintain VHF radio communications with
other vessels and appropriate shore authorities
NO Go to 17.
Monitor vessel position to determine if anchor is

dragging
YES Briefly describe activity below.
Page 127 of 141
NO Go to 18.
Houston, Texas

(Bridge Operations Mariner Skill & Knowledge Limitations, cont.)
Step 5.3: Training and Procedures
Write the identification numbers of up to three bridge activities checked in Step 5.2 that most contributed to the
casualty. (Example: Activity 1: 10a, Activity 2: 13b, Activity 3: 16a.) Then, complete the remaining items under each
listed activity.
18. Activity 1:
23. Activity 2:
28. Activity 3:
19. What training has the mariner had to

prepare for Activity 1?
(Check all that apply.)

29. What training has the mariner had

to prepare for Activity 3?
q
q
q
q
q
q
q
No training of any kind

Informal on-the-job training
Formal, structured on-the-job
training and supervision
Coast Guard-approved course
Maritime trade school training
Maritime college or academy
training
Other training
q
q
q
q
q
q
q
q
q
q
q
q
q
q

training
Other training

training
Other training
20. Briefly describe the mariners most

relevant training for Activity 1.


21. How long has it been since the

mariner received this Activity 1
training?

training?

training?
_____ years and _____ months
22. Briefly describe established vessel

procedures, regulations, or common
practices that guide mariner
performance of Activity 1.


Are procedures adequate?
<(6 12
<(6 12
<(6 12
Step 5.4: Conclusions and Recommendations

Respond to items 33-34 after completing Step 5.3.
33. If the mariner lacks skill or knowledge in any activity
(1 to 3), complete 33 and 34; otherwise, end this
report (NOT skill or knowledge related).
Activity 1
Activity 2
Activity 3
a.
The mariner most likely lacks skill in this

activity.
758( )$/6(
758( )$/6(
758( )$/6(
b.
The mariner most likely lacks knowledge in

this activity.
758( )$/6(
758( )$/6(
758( )$/6(
34. What could be done to improve this mariners skill and/or knowledge, or to improve established procedures and reduce
casualties?
Minimum:
Ideal:
Page 128 of 141
Houston, Texas
Deck Operations Mariner Skill & Knowledge Limitations

1. Mariners name:

9.
Briefly describe how this persons specific deck actions, inaction, or decisions contributed to the casualty:
Now, check
all deck activities (10-21) ) that directly contributed to the casualty.
10. Did vessel stability and integrity management activities

contribute?
q a.
Load and unload a vessel taking into account load

lines, stability, trim, and stress principles and
calculations
11. Did deck equipment operations activities contribute?
q b.
q c.
q d.
Adjust ballast as required to maintain stability

Operate vessel in compliance with Stability Letter
Ensure vessels water tight integrity
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
q d.
Board pilot
Conduct docking, anchoring, and mooring operations
12. Did container cargo operations activities contribute?
Establish container stowage plan
Load and unload containers
13. Did bulk cargo operations contribute?
Set, retrieve, and handle fishing gear

Bring aboard and load catch
17. Did deck communications activities contribute?
q a.
q b.
Perform deck, hull, and surface chipping, painting

Maintain deck equipment
19. Did general activities on deck activities contribute?
q a.
q b.
Embarking or disembarking vessel

Moving around the vessel
20. Did passenger safety activities contribute?
q a.
q b.
NO Go to 19.
Work in confined spaces

Perform hot work
q c.
q d.
NO Go to 18.
Coordinate between deck and assist vessels

Coordinate between deck and dock crew
NO Go to 20.
General off-duty activities onboard vessel
Ensure the safety of passengers during embarkation

and disembarkation
Ensure the safety of passengers when underway and
during ship operations
21. Did a deck activity not listed above contribute?
NO Go to 17.
Unload or transfer catch

Process catch
q c.
NO Go to 16.
Check tow for water and pump barges

Conduct locking and lock assist operations
q c.
q d.
NO Go to 15.
Clean petroleum cargo tanks

Conduct inert gas and gas-free operations
Communicate effectively between deck and bridge

Communicate effectively among deck crew
18. Did deck maintenance activities contribute?
NO Go to 14.
Monitor and maintain cargo security

Handle dangerous and hazardous cargo
Establish a tow diagram

Make up, check, and tighten towlines and headwires
16. Did fishing operations activities contribute?
NO Go to 13.
Lash all containers

Monitor and maintain cargo security
Operate pumping equipment

Monitor piping and pumping systems
15. Did towing and fleeting operations contribute?
NO Go to 12.
Assist in tug/escort vessel tie-up operations

Prepare and stow cargo handling equipment
Establish bulk cargo loading plan

Load and unload bulk cargo
14. Did petroleum cargo activities contribute?
NO Go to 11.
NO Go to 21.
Inspect passenger spaces for hazards and take

appropriate action
Confine passenger access to safe vessel spaces only
Page 129 of 141
NO Go to 22.
Houston, Texas
(Deck Operations Mariner Skill & Knowledge Limitations, cont.)

Write the identification numbers of up to three deck activities checked in Step 5.2 that most contributed to the casualty.
(Example: Activity 1: 10a, Activity 2: 13b, Activity 3: 16a.) Then, complete the remaining items under each listed activity.
22. Activity 1:
27. Activity 2:
32. Activity 3:



q
q
q
q
q
q
q

training
Other training
q
q
q
q
q
q
q
q
q
q
q
q
q
q

training
Other training

training
Other training




training?

training?

training?



<(6 12
<(6 12
<(6 12

Activity 1
Activity 2
Activity 3
a.

activity.
758( )$/6(
758( )$/6(
758( )$/6(
b.

this activity.
758( )$/6(
758( )$/6(
758( )$/6(
casualties?
Minimum:
Ideal:
Page 130 of 141
Houston, Texas
Engineering Operations Mariner Skill & Knowledge Limitations

Please complete this form separately for each person whose engineering activities contributed to the casualty.

1. Mariners name:

9.
Briefly describe how this persons specific engineering actions, inaction, or decisions contributed to the casualty:
Now, check
all engineering activities (10-16) ) that directly contributed to the casualty.
10. Did changing engineering watch activities contribute to

casualty?
q a.
q b.
Check and acknowledge standing orders, night orders,

and special information
Check status of ships equipment
11. Did engineering systems operations activities contribute
to casualty?
q a.
q b.
q c.
Operate main propulsion system (engines, boilers, fuel

and steering
Operate generating and electrical systems
Operate motors, pumps, and lubrication systems
12. Did engineering systems inspection and testing
activities contribute to casualty?
q a.
q b.
Inspect and test main propulsion system (engines,

boilers, fuel, and steering)
Inspect and test generating and electrical systems
13. Did routine, scheduled, and preventive maintenance

q a.
q b.
Maintain main propulsion system (engines, boilers,

fuel, and steering)
Maintain generating and electrical systems
14. Did unscheduled, corrective repair activities contribute
to casualty?
q a.
q b.
Repair main propulsion system (engines, boilers, fuel,

and steering)
Repair generating and electrical systems
15. Did engineering communications activities contribute to
casualty?
q a.

vessels crew (Bridge, Engine, and Deck)
16. Did engineering activity not listed above contribute to
casualty?
q c.
q d.
Assess traffic and weather conditions

Ensure that watch is relieved
q d.
q e.
NO Go to 15.
Repair motors, pumps, and lubrication systems

Repair service equipment (evaporators, refrigeration,
heating, AC, sewage, and garbage treatment)
q b.
NO Go to 14.
Maintain motors, pumps, and lubrication systems

Maintain service equipment (evaporators, refrigeration,
q c.
q d.
NO Go to 13.
Inspect and test motors, pumps, and lubrication

systems
Inspect and test service equipment (evaporators,
refrigeration, heating, AC, sewage, and garbage
treatment)
q c.
q d.
NO Go to 12.
Operate service equipment (evaporators, refrigeration,

Load, discharge, or transfer fuel between tanks
q c.
q d.
NO Go to 11.
NO Go to 16.

engineering crew
Page 131 of 141
NO Go to 17.
Houston, Texas
(Engineering Operations Mariner Skill & Knowledge Limitations, cont.)

Write the identification numbers of up to three engineering activities checked in Step 5.2 that most contributed to the casualty.
(Example: Activity 1: 10a, Activity 2: 13b, Activity 3: 16a.) Then, complete the remaining items under each listed activity.
17. Activity 1:
22. Activity 2:
27. Activity 3:



q
q
q
q
q
q
q

training
Other training
q
q
q
q
q
q
q
q
q
q
q
q
q
q

training
Other training

training
Other training




training?

training?

training?



<(6 12
<(6 12
<(6 12

Activity 1
Activity 2
Activity 3
a.

activity.
758( )$/6(
758( )$/6(
758( )$/6(
b.

this activity.
758( )$/6(
758( )$/6(
758( )$/6(
casualties?
Minimum:
Ideal:
Page 132 of 141
Houston, Texas
Safety & Emergency Operations Mariner Skill & Knowledge Limitations

Please complete this form separately for each person whose safety and emergency activities contributed to the casualty.

1. Mariners name:

9. Briefly describe how this persons specific safety and emergency actions, inaction, or decisions contributed to the casualty:
Now, check
all safety and emergency activities (10-18) ) that directly contributed to the casualty.
10. Did general safety activities contribute to casualty?
q a.
q b.
Embark and disembark vessel safely

Walk about vessel safely
11. Did safety equipment inspection and service activities
contribute to casualty?
q a.
q b.
q c.
Inspect and service fire detection equipment

Inspect and service fire extinguishing equipment
Inspect and service lifesaving equipment, locating
devices, and flotation devices
12. Did controlling and fighting fires activities contribute to
casualty?
q a.
q b.
q c.
Establish and maintain a Fire Safety Plan

Organize and conduct fire drills
Identify the type of fire
13. Did confined space rescue activities contribute to
casualty?
q c.
q d.
q e.
q c.
q d.
q a.
q b.
q c.
q d.
q a.
q b.
q c.
Locate individual(s)
Establish a rescue plan
14. Did person overboard procedures contribute?
Don survival suits and personal flotation devices

Launch, load, and maneuver lifeboats and life rafts
16. Did emergency medical and life-saving procedures
q a.
q b.
q c.
q d.
q e.
NO Go to 16.
Employ locating devices properly
q b.
Establish and maintain communications with crew

Establish and maintain communications with
passengers
Establish and maintain emergency communications
with other vessels
18. Did safety and emergency activity not listed above
contribute to casualty?
NO Go to 15.
Maneuver vessel
Bring person aboard
a. Use medical chest and first aid items

17. Did emergency communications activities contribute?
NO Go to 14.
Use breathing apparatus and other required equipment

Maintain back-up personnel and escape routes
Initiate warning
Locate person overboard
15. Did abandon vessel operations contribute to casualty?
NO Go to 13.
Use fire-fighting equipment and procedures

Maintain escape routes
q a.
q b.
NO Go to 12.
Inspect and service survival craft

Inspect and service emergency generator, batteries,
etc.
q d.
q e.
NO Go to 11.
Perform off-duty activities safely
NO Go to 17.
Apply First Aid/CPR

NO Go to 18.
Establish emergency communications with shore

authorities
Monitor GMDSS and other emergency frequencies as
required
Page 133 of 141
NO Go to 19.
Houston, Texas
(Safety & Emergency Operations Mariner Skill & Knowledge Limitations, cont.)
Write the identification numbers of up to three safety and emergency activities checked in Step 5.2 that most
contributed to the casualty. (Example: Activity 1: 10a, Activity 2: 13b, Activity 3: 16a.) Then, complete the items
under each listed activity.
19. Activity 1:
24. Activity 2:
29. Activity 3:



q
q
q
q
q
q
q

training
Other training
q
q
q
q
q
q
q
q
q
q
q
q
q
q

training
Other training

training
Other training




training?

training?

training?



<(6 12
<(6 12
<(6 12

Activity 1
Activity 2
Activity 3
a.

activity.
758( )$/6(
758( )$/6(
758( )$/6(
b.

this activity.
758( )$/6(
758( )$/6(
758( )$/6(
casualties?
Minimum:
Ideal:
Page 134 of 141
Houston, Texas
Appendix H. Human Factors Investigation Questions

The following is taken from Role of the Human Element in Maritime Casualties, by the Joint
ILO/IMO adhoc Working Group on Investigation of Human Factors in Maritime Casualties,
Maritime Safety Committee, MSC 69/13/1 ([MSC], 1998, Annex 4, Appendix 2). These are
questions designed to help the casualty investigator ask questions appropriate to discovering
human errors which may have contributed to an incident. While no list of questions can be
comprehensive, this one provides a good set of initial inquiries into the events and underlying
causes. Investigators can use these to explore many facets surrounding the incident, choosing
relevant areas for follow-up. Even though the questions below assume an accident to a
commercial ship, the questions can easily be tailored to investigate incidents on any type of
facility or vessel. The text of the MSC document is provided below.
-------------------Note: The following questions are designed to aid the investigator while investigating for
human factors, particularly fatigue. Skilful [sic] questioning can help the investigator eliminate
irrelevant lines of inquiry and focus on areas of greater potential significance.
The order in which the questions should be asked will depend on who is being
interviewed and on his or her willingness and ability to describe personal behaviour and
personal impressions. Also, it may be necessary to verify, cross-check or augment information
received from one person by interviewing others on the same points.
These areas of inquiry can be used in training investigators as well as in planning
interviews. The following questions are not intended as a checklist, and some may not be
relevant in the investigation of a particular accident.
1
Safety Policy
.1
.2
.3
.4
.5
.6
.7
Does the company have a written safety policy?

Is there a designated person for shipboard safety matters in the company?
When did a company Representative last visit the vessel or when were you last in
contact with the company?
When were you last provided safety training? What was the training and how was it
provided?
When was the last emergency drill (e.g., fire, abandon ship, man-overboard,
pollution response, etc) and what did you do during the drill?
Was appropriate personal protective equipment provided and did you use it?
Are you aware of any personal accidents which occurred on board in the period
prior to the accident?
Activities prior to incident

.1
.2
(If the ship was leaving port at the time of the accident) In general, how did you
spend your time while the ship was in port?
(If the ship was approaching port or at sea at the time of the accident) How long
has the ship been on passage since its last port or terminal operation?
Page 135 of 141
Houston, Texas
.3
.4
.5
.6
.7
Duties at the time of accident

.1
.2
.3
.2
.3
.4
.5
How long have you been assigned to this ship? Have you requested that your
assignment be lengthened or shortened?
How long have you filled your crew position? What other crew positions have you
held on this ship?
How long have you held the certificate indicating your qualifications?
Before being assigned to this ship, did you work on other ships? If so, what crew
positions have you held?
What is the longest time you have been to sea in a single voyage? How long have
you been at sea on this passage? What was your longest single passage?
Physical condition
.1
.2
.3
Precisely where were you located at the time of the accident?

What specific task were you performing at the time of the accident?
Had you at any time since reporting for duty found that you could not concentrate
(focus your attention/keep your mind) on a task you were trying to perform?
Training/Education/Certification/Professional Experience
.1
Where were you on the ship when the accident occurred?

What specific job or duty were you assigned at the time? By whom? Did you
understand your assignment? Did you receive any conflicting orders?
How often have you performed this job in the past (on the specific ship involved in
the accident)?
Actual behaviour at time of accident

.1
.2
.3
What were you doing immediately prior to coming on watch or reporting for duty,
and for how long? Recreational activity? Physical exercise? Sleeping? Reading?
Watching T.V.? Eating? Paperwork? Travelling to vessel?
Specifically what were you doing approximately 4 hours ........ 1 hour ........ 30
minutes ....... before the accident?
What evolution was the ship involved in when the accident occurred? What was
your role during that evolution?
Immediately prior to the accident, what were you thinking about?
At any time before the accident, did you have any indication that anyone was tired
or unable to perform their duty?
Were you feeling ill or sick at any time in the 24 hours immediately before the
accident? If so, what symptoms did you have? Did you have a fever, vomit, feel
dizzy, other? Also, did you tell anyone? What do you believe the cause was?
When was the last meal you had prior [to] the accident? What did you eat? Was it
adequate?
Do you exercise regularly while onboard? When did you last exercise (before the
accident)? How long was the session?
Psychological, emotional, mental condition and employment conditions
Page 136 of 141
Houston, Texas
.1
.2
.3
.4
.5
.6
.7
8
Workload/Complexity of Tasks
.1
.2
.3
.4
.5
What is the shipboard organization?

Is the shipboard organization effective?
What is your position in the shipboard organization (i.e., who do you work for, report
to or assign duties to)?
What is the nature of your work? Sedentary? Physically demanding?
Was anyone involved in the accident impaired due to heavy workload?
Work-period/rest-period/recreation pattern
.1
.2
.3
.4
.5
.6
.7
.8
10
When was the last time you felt cheerful or elated onboard the ship, and what were
the circumstances that generated this emotion?
When was the last time you were sad or depressed or dejected, on board the ship?
Why? Did you talk about it with anyone else?
Have you had to make any difficult personal decisions recently? Have you had any
financial or family worries on your mind recently?
Have you been criticized for how you are doing your work lately? By whom? Was it
justified?
What was the most stressful situation you had to deal with on the voyage (prior to
the accident)? When did the situation occur? How was it resolved?
What are the contractual arrangements for all crewmembers?
Have there been any complaints or industrial action in the last (12) months?
What is your normal duty schedule?

Are you a day worker or a watchstander?
What was your duty schedule on the day before the accident and during the week
before the accident?
Were you on overtime at the time of the accident?
How long had you been on duty, or awake performing other work, at the time of the
accident?
When was your lat period of sleep? How long did it last? How often did you
awaken during your last sleep period? Did you awaken refreshed? If not, what
would have made your sleep period more restful?
How do you normally spend your off-duty time while on board? Play cars? Read?
Listen to music? Watch T.V.? Other?
When was your last extended period of off duty time when you were able to rest?
Relationship with other crewmembers and superiors/subordinates

.1
.2
.3
.4
.5
.6
Who among the crew would you consider to be a friend?

Do you find any members of the crew unpleasant to be with?
Do you have difficulty talking with any of the crewmembers because of language
barriers?
Have any new crewmembers recently joined the ship? Have you had a chance to
get acquainted with them?
Did you have any argument recently with another crewmember?
In an emergency, would you trust your fellow crewmembers to come to your
assistance?
Page 137 of 141
Houston, Texas
.7
.8
.9
11
Living conditions and shipboard environment

.1
.2
12
.3
Are there written standing orders to the whole crew complement from the Master?
Did the Master/Chief Engineer provide written or verbal standing orders to the
watchkeeping personnel?
Were the orders in conflict with the company safety policy?
Level of automation/reliability of equipment

.1
.2
.3
15
Is the manning level sufficient in your opinion for the operation of the vessel?
Masters standing orders

.1
.2
14
Do you consider your personal area on board the ship to be comfortable? If not,
how would you like it to [be] improved?
Prior to the accident, did you have any difficulty resting as a result of sever weather,
noise levels, heat/cold, ships motion, etc.?
Manning levels
.1
13
Has another crewmember ever offered to take your place on watch or perform a
duty for you to let you get some extra rest?
What was the subject of your last conversation with another crewmember before
reporting for duty (when the accident occurred)?
Have you talked with any other crewmembers since the accident? If so, what was
the subject of your conversation? Have you talked with anyone else about the
accident prior to being interviewed?
In your opinion, was the system reliable?

Were there earlier failures in the system?
Were the failures repaired by the crew or shore-based workers?
Ship design, motion/cargo characteristics

.1
Did you observe anything out of the ordinary on this passage concerning the ship
design, or motion or cargo characteristics?
Questions 16-24 [25] are EXAMPLES OF QUESTIONS THAT MAY BE DIRECTED TO

SHORESIDE MANAGEMENT
16
Scheduling of work and rest periods

.1
17
Manning level
.1
18
What is the companys work schedule and relief policy?
How is the manning level determined for your fleet?
Watchkeeping practices
.1
Do you require the Master to stand watch?
Page 138 of 141
Houston, Texas
.2
19
Assignment of duties
.1
20
Are welfare/recreational services and facilities provided on board?
Contractual and/or industrial arrangements and agreements

.1
.2
25
How does the Master plan the voyages?
Recreational facilities
.1
24
Does the company have a written safety policy?
Voyage planning and port call schedules

.1
23
How do you support the vessels Master?
Management policies
.1
22
Do you leave this matter to the Master?
Shore-ship-shore support and communications

.1
21
Do you leave the watchkeeping practices to the discretion of the Master?
What are the contractual agreements for all crewmembers?

Have there been any complaints or industrial action in the last (12) months?
National/international requirements
.1
Are the management/Master complying with the requirements and recommendations of the applicable international conventions and Flag State regulations?
Page 139 of 141
Houston, Texas
Appendix I. Norwegian Petroleum Directorate Near-Miss System
In 2000, Norway instituted a near-miss database to assess risk levels in the continental shelf
offshore industry. The database contains reports on major hazards, occupational injuries,
occupational disease, cultural risk factors, and perceived risks. The focus is on preventing risk
to personnel on offshore facilities, such as production installations, mobile drilling units, and
flotels. A report, Trends in Risk Levels on the Norwegian Continental Shelf, gives an overview
of the near-miss system and findings to date. The report can be found on the Norwegian
Petroleum Directorate (NPD) web site (http://www.npd.no ; click on the British flag for English,
then go to the Health, Environment & Safety (HSE) page). This appendix gives a brief
presentation of the near-miss system and a couple examples of the data, courtesy of Professor
Jan Erik Vinnem.
The near-miss database is limited to incidents that may have the potential to cause major
accidents, if multiple barrier failures occur. Some other incidents that are essential for
emergency preparedness planning are also covered. The operators on the Norwegian
Continental Shelf have a duty (enforced by NPD through regulations) to notify NPD about
injuries, accidents and near-misses within a short time after the occurrence of these events.
(For the most serious incidents, further investigation reports will be required for submission.)
The cut-off limits for which incidents to report are somewhat loosely defined, and there are
significant differences between the companies, with respect to reporting practices. The
database is based on a subset of these mandatory reports, and includes those which have been
extensively reviewed and verified to ensure consistency and which have established exposure
data (activity levels).
The NPD classifies major hazards into eleven types (or DFUs), shown below.
presents a specific, potentially-serious hazardous situation.
DFU
1
2
3
4
5
6
7
8
9
10
11
Each DFU
Event Scenario
Unignited hydrocarbon leak
Ignited hydrocarbon leak
Kick/loss of well control
Fire/explosion, excluding DFU#2
Vessel on collision course
Drifting object/vessel on collision course
collision with field related traffic
Structural damage
Leak from subsea installation
Damage to subsea installation
Evacuation (precautionary/emergency)
Page 140 of 141
Houston, Texas
Although the near-miss database was not formally brought on-line until 2000, the NPD had
incident data from several prior years. The NPD has analyzed yearly incident data in terms of
the DFUs, as shown in the trend analysis presented in Figure I-1 (see Sec. 6.3 for information
on trend analysis). It is clear from the figure that unignited hydrocarbon leaks are a primary
source of potential accidents. These data have been further analyzed by size of leak in Figure
I-2. The NPD system demonstrates how near-miss data can be used to identify potential
hazardous situations so that the industry can seek safety solutions before a significant accident
results.
120
Number of DFU occurences
Helicopt incid, f ield

Evac/muster
100
Dam. subs ins t

Subs equipm leak
80
Struct. damage
60
Coll. attend vess.

Drif ting object
40
Ship on coll course

Other fire/expl
20
Kick
Ign HC leak
Unign HC leak
1996
1997
1998
1999
2000
2001
Figure I-1. Norwegian Petroleum Directorate Near-Miss Data. Six years of

incident data for all major hazard DFUs, representing near-misses at
all continental shelf offshore installations.
Page 141 of 141
Houston, Texas
50
45
Number of leaks
40
35
30
>10 kg/s
25
1-10 kg/s
20
0,1-1 kg/s
15
10
5
0
1996
1997
1998
1999
2000
2001
Figure I-2. NPD Data on Unignited Hydrocarbon Leaks. Data are segmented
by size of the leak.
Page 142 of 141
Houston, Texas

Human Factors in Incident Investigation and Analysis

Uploaded by

Copyright:

Available Formats

Human Factors in Incident Investigation and Analysis

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Human Factors in Incident Investigation and Analysis

Uploaded by

Copyright:

Available Formats

Human Factors in Incident Investigation and Analysis

Capt. David Wheal and Mr. Stuart Withington

Dr. Scott A. Shappell

Dr. Douglas A. Wiegmann

Mr. William Boehm

Dr. Marc Chaderjian

April 8 10, 2002

Working Group 1 - HFW2002

Report Documentation Page

4. TITLE AND SUBTITLE

5a. CONTRACT NUMBER