Controlling is a four-step process of establishing performance standards based on the firm’s

objectives, measuring and reporting actual performance, comparing the two, and taking
corrective or preventive action as necessary. This article will give you some information about
controlling function of management.
Performance standards come from the planning function. No matter how difficult, standards
should be established for every important task. Although the temptation may be great, lowering
standards to what has been attained is not a solution to performance problems. On the other hand,
a manager does need to lower standards when they are found to be unattainable due to resource
limitations and factors external to the business.
Corrective action is necessary when performance is below standards. If performance is
anticipated to be below standards, preventive action must be taken to ensure that the problem
does not recur. If performance is greater than or equal to standards, it is useful to reinforce
behaviors that led to the acceptable performance.
Characteristics of the Control Process
The control process is cyclical which means it is never finished. Controlling leads to
identification of new problems that in turn need to be addressed through establishment of
performance standards, measuring performance etc.
Employees often view controlling negatively. By its very nature, controlling often leads to
management expecting employee behavior to change. No matter how positive the changes may
be for the organization, employees may still view them negatively.
Control is both anticipatory and retrospective. The process anticipates problems and takes
preventive action. With corrective action, the process also follows up on problems.
Ideally, each person in the business views control as his or her responsibility. The organizational
culture should prevent a person walking away from a small, easily solvable problem because
“that isn’t my responsibility.” In customer driven businesses, each employee cares about each
customer. In quality driven dairy farms, for example, each employee cares about the welfare of
each animal and the wear and tear on each piece of equipment.
Controlling is related to each of the other functions of management. Controlling builds on
planning, organizing and leading.
Management Control Strategies
Managers can use one or a combination of three control strategies or styles: market, bureaucracy
and clan. Each serves a different purpose. External forces make up market control. Without
external forces to bring about needed control, managers can turn to internal bureaucratic or clan
control. The first relies primarily on budgets and rules. The second relies on employees wanting
to satisfy their social needs through feeling a valued part of the business.
Self-control, sometimes called adhocracy control, is complementary to market, bureaucratic and
clan control. By training and encouraging individuals to take initiative in addressing problems on
their own, there can be a resulting sense of individual empowerment. This empowerment plays
out as self-control. The self-control then benefits the organization and increases the sense of
worth to the business in the individual.
Designing Effective Control Systems
Effective control systems have the following characteristics:
1. Control at all levels in the business (Figure 19.1)
2. Acceptability to those who will enforce decisions
3. Flexibility
4. Accuracy
5. Timeliness
6. Cost effectiveness
7. Understandability
8. Balance between objectivity and subjectivity
9. Coordinated with planning, organizing and leading
Dysfunctional Consequences of Control
Managers expect people in an organization to change their behavior in response to control.
However, employee resistance can easily make control efforts dysfunctional. The following
behaviors demonstrate means by which the manager’s control efforts can be frustrated:
1. Game playing–> control is something to be beaten, a game between the “boss and me and I
want to win.”
2. Resisting control–> a “blue flu” reaction to too much control
3. Providing inaccurate information –> a lack of understanding of why the information is needed
and important leading to “you want numbers, we will give you numbers.”
4. Following rules to the letter–> people following dumb and unprofitable rules in reaction to “do
as I say.”
5. Sabotaging –> stealing, discrediting other workers, chasing customers away, gossiping about
the firm to people in the community
6. Playing one manager off against another –> exploiting lack of communication among
managers, asking a second manager if don’t like the answer from the first manager.

Control is one of the managerial functions like planning, organizing, staffing and directing. It is
an important function because it helps to check the errors and to take the corrective action so that
deviation from standards are minimized and stated goals of the organization are achieved in
desired manner.
According to modern concepts, control is a foreseeing action whereas earlier concept of control
was used only when errors were detected. Control in management means setting standards,
measuring actual performance and taking corrective action. Thus, control comprises these three
main activities.
Definitions
According to Henri Fayol,
Control of an undertaking consists of seeing that everything is being carried out in accordance
with the plan which has been adopted, the orders which have been given, and the principles
which have been laid down. Its object is to point out mistakes in order that they may be rectified
and prevented from recurring.
According to EFL Breach,
Control is checking current performance against pre-determined standards contained in the
plans, with a view to ensure adequate progress and satisfactory performance.
According to Harold Koontz,
Controlling is the measurement and correction of performance in order to make sure that
enterprise objectives and the plans devised to attain them are accomplished.
According to Stafford Beer,
Management is the profession of control.
In 1916, Henri Fayol formulated one of the first definitions of control as it pertains to
management:
Control consists of verifying whether everything occurs in conformity with the plan adopted, the
instructions issued, and principles established. It ['s] object [is] to point out weaknesses and
errors in order to rectify [them] and prevent recurrence.[1]
Robert J. Mockler presented a more comprehensive definition of managerial control:
Management control can be defined as a systematic effort by business management to compare
performance to predetermined standards, plans, or objectives in order to determine whether
performance is in line with these standards and presumably in order to take any remedial action
required to see that human and other corporate resources are being used in the most effective
and efficient way possible in achieving corporate objectives.[2]
Also control can be defined as "that function of the system that adjusts operations as needed to
achieve the plan, or to maintain variations from system objectives within allowable limits". The
control subsystem functions in close harmony with the operating system. The degree to which
they interact depends on the nature of the operating system and its objectives. Stability concerns
a system's ability to maintain a pattern of output without wide fluctuations. Rapidity of response
pertains to the speed with which a system can correct variations and return to expected output.[3]

A political election can illustrate the concept of control and the importance of feedback. Each
party organizes a campaign to get its candidate selected and outlines a plan to inform the public
about both the candidate's credentials and the party's platform. As the election nears, opinion
polls furnish feedback about the effectiveness of the campaign and about each candidate's
chances to win. Depending on the nature of this feedback, certain adjustments in strategy and/or
tactics can be made in an attempt to achieve the desired result.

From these definitions it can be stated that there is close link between planning and controlling.
Planning is a process by which an organisation's objectives and the methods to achieve the
objectives are established, and controlling is a process which measures and directs the actual
performance against the planned objectives of the organisation. Thus, planning and control are
often referred to as siamese twins of management. controlling is the managerial function of
management and correction of performance in order to make sure that enterpriseobjectives and
the plans devised to attain them being accomplished.

[edit] Characteristics of Control

 Control is a continuous process
 Control is a management process
 Control is embedded in each level of organizational hierarchy
 Control is forward looking
 Control is closely linked with planning
 Control is a tool for achieving organizational activities

[edit] The elements of control

The four basic elements in a control system — (1) the characteristic or condition to be
controlled, (2) the sensor, (3) the comparator , and (4) the activator — occur in the same
sequence and maintain a consistent relationship to each other in every system.[3]
The first element is the characteristic or condition of the operating system which is to be
measured. We select a specific characteristic because a correlation exists between it and how the
system is performing. The characteristic may be the output of the system during any stage of
processing or it may be a condition that has resulted from the output of the system. For example,
it may be the heat energy produced by the furnace or the temperature in the room which has
changed because of the heat generated by the furnace. In an elementary school system, the hours
a teacher works or the gain in knowledge demonstrated by the students on a national examination
are examples of characteristics that may be selected for measurement, or control. The second
element of control, the sensor, is a means for measuring the characteristic or condition. The
control subsystem must be designed to include a sensory device or method of measurement. In a
home heating system this device would be the thermostat, and in a quality-control system this
measurement might be performed by a visual inspection of the product.

The third element of control, the comparator, determines the need for correction by comparing
what is occurring with what has been planned. Some deviation from plan is usual and expected,
but when variations are beyond those considered acceptable, corrective action is required. It is
often possible to identify trends in performance and to take action before an unacceptable
variation from the norm occurs. This sort of preventative action indicates that good control is
being achieved.

The fourth element of control, the activator, is the corrective action taken to return the system to
expected output. The actual person, device, or method used to direct corrective inputs into the
operating system may take a variety of forms. It may be a hydraulic controller positioned by a
solenoid or electric motor in response to an electronic error signal, an employee directed to
rework the parts that failed to pass quality inspection, or a school principal who decides to buy
additional books to provide for an increased number of students. As long as a plan is performed
within allowable limits, corrective action is not necessary; this seldom occurs in practice,
however.

Information is the medium of control, because the flow of sensory data and later the flow of
corrective information allow a characteristic or condition of the system to be controlled. To
illustrate how information flow facilitates control, let us review the elements of control in the
context of information.[4]

[edit] Relationship between the elements of control and information

Controlled Characteristic or. Condition The primary requirement of a control system is that it
maintain the level and kind of output necessary to achieve the system's objectives.[5] It is usually
impractical to control every feature and condition associated with the system's output. Therefore,
the choice of the controlled item (and appropriate information about it) is extremely important.
There should be a direct correlation between the controlled item and the system's operation. In
other words, control of the selected characteristic should have a direct relationship to the goal or
objective of the system.

Sensor
After the characteristic is sensed, or measured, information pertinent to control is fed back.
Exactly what information needs to be transmitted and also the language that will best facilitate
the communication process and reduce the possibility of distortion in transmission must be
carefully considered. Information that is to be compared with the standard, or plan, should be
expressed in the same terms or language as in the original plan to facilitate decision making.
Using machine methods (computers) may require extensive translation of the information. Since
optimal languages for computation and for human review are not always the same, the relative
ease of translation may be a significant factor in selecting the units of measurement or the
language unit in the sensing element.

In many instances, the measurement may be sampled rather than providing a complete and
continuous feedback of information about the operation. A sampling procedure suggests
measuring some segment or portion of the operation that will represent the total.[2]

Comparison with Standard

In a social system, the norms of acceptable behavior become the standard against which so-
called deviant behavior may be judged. Regulations and laws provide a more formal collection
of information for society. Social norms change, but very slowly. In contrast, the standards
outlined by a formal law can be changed from one day to the next through revision,
discontinuation, or replacement by another. Information about deviant behavior becomes the
basis for controlling social activity. Output information is compared with the standard or norm
and significant deviations are noted. In an industrial example, frequency distribution (a
tabulation of the number of times a given characteristic occurs within the sample of products
being checked) may be used to show the average quality, the spread, and the comparison of
output with a standard.

If there is a significant and uncorrectable difference between output and plan, the system is "out
of control." This means that the objectives of the system are not feasible in relation to the
capabilities of the present design. Either the objectives must be reevaluated or the system
redesigned to add new capacity or capability. For example, the traffic in drugs has been
increasing in some cities at an alarming rate. The citizens must decide whether to revise the
police system so as to regain control, or whether to modify the law to reflect a different norm of
acceptable behavior.

Implimentor

The activator unit responds to the information received from the comparator and initiates
corrective action. If the system is a machine-to-machine system, the corrective inputs (decision
rules) are designed into the network. When the control relates to a man-to-machine or man-to-
man system, however, the individual(s) in charge must evaluate (1) the accuracy of the feedback
information, (2) the significance of the variation, and (3) what corrective inputs will restore the
system to a reasonable degree of stability. Once the decision has been made to direct new inputs
into the system, the actual process may be relatively easy. A small amount of energy can change
the operation of jet airplanes, automatic steel mills, and hydroelectric power plants. The pilot
presses a button, and the landing gear of the airplane goes up or down; the operator of a steel mill
pushes a lever, and a ribbon of white-hot steel races through the plant; a worker at a control
board directs the flow of electrical energy throughout a regional network of stations and
substations. It takes but a small amount of control energy to release or stop large quantities of
input.[4]

The comparator may be located far from the operating system, although at least some of the
elements must be in close proximity to operations. For example, the measurement (the sensory
element) is usually at the point of operations. The measurement information can be transmitted
to a distant point for comparison with the standard (comparator), and when deviations occur, the
correcting input can be released from the distant point. However, the input (activator) will be
located at the operating system. This ability to control from afar means that aircraft can be flown
by remote control, dangerous manufacturing processes can be operated from a safe distance, and
national organizations can be directed from centralized headquarters.

[edit] Process of Controlling

 Setting performance standards.
 Measurement of actual performance.
 Comparing actual performance with standards.
 Analysing deviations.
 Correcting deviations.

[edit] Kinds of control

Control may be grouped according to three general classifications: (1) the nature of the
information flow designed into the system (that is, open- or closed-loop control), (2) the kind of
components included in the design (that is man or machine control systems), and (3) the
relationship of control to the decision process (that is, organizational or operational control).[3]

[edit] Open- and Closed-Loop Control

A street-lighting system controlled by a timing device is an example of an open-loop system. At

a certain time each evening, a mechanical device closes the circuit and energy flows through the
electric lines to light the lamps. Note, however, that the timing mechanism is an independent unit
and is not measuring the objective function of the lighting system. If the lights should be needed
on a dark, stormy day the timing device would not recognize this need and therefore would not
activate energy inputs. Corrective properties may sometimes be built into the controller (for
example, to modify the time the lights are turned on as the days grow shorter or longer), but this
would not close the loop. In another instance, the sensing, comparison, or adjustment may be
made through action taken by an individual who is not part of the system. For example, the lights
may be turned on by someone who happens to pass by and recognizes the need for additional
light.

If control is exercised as a result of the operation rather than because of outside or predetermined
arrangements, it is a closed-loop system. The home thermostat is the classic example of a control
device in a closed-loop system. When the room temperature drops below the desired point, the
control mechanism closes the circuit to start the furnace and the temperature rises. The furnace-
activating circuit is turned off as the temperature reaches the preselected level. The significant
difference between this type of system and an open-loop system is that the control device is an
element of the system it serves and measures the performance of the system. In other words, all
four control elements are integral to the specific system.

An essential part of a closed-loop system is feedback; that is, the output of the system is
measured continually through the item controlled, and the input is modified to reduce any
difference or error toward zero. Many of the patterns of information flow in organizations are
found to have the nature of closed loops, which use feedback. The reason for such a condition is
apparent when one recognizes that any system, if it is to achieve a predetermined goal, must have
available to it at all times an indication of its degree of attainment. In general, every goal-seeking
system employs feedback.[3]' ==

[edit] Man and Machine Control

The elements of control are easy to identify in machine systems. For example, the characteristic
to be controlled might be some variable like speed or temperature, and the sensing device could
be a speedometer or a thermometer. An expectation of precision exists because the characteristic
is quantifiable and the standard and the normal variation to be expected can be described in exact
terms. In automatic machine systems, inputs of information are used in a process of continual
adjustment to achieve output specifications. When even a small variation from the standard
occurs, the correction process begins. The automatic system is highly structured, designed to
accept certain kinds of input and produce specific output, and programmed to regulate the
transformation of inputs within a narrow range of variation.[6]

For an illustration of mechanical control, as the load on a steam engine increases and the engine
starts to slow down, the regulator reacts by opening a valve that releases additional inputs of
steam energy. This new input returns the engine to the desired number of revolutions per minute.
This type of mechanical control is crude in comparison to the more sophisticated electronic
control systems in everyday use. Consider the complex missile-guidance systems that measure
the actual course according to predetermined mathematical calculations and make almost
instantaneous corrections to direct the missile to its target.

Machine systems can be complex because of the sophisticated technology, whereas control of
people is complex because the elements of control are difficult to determine. In human control
systems, the relationship between objectives and associated characteristics is often vague; the
measurement of the characteristic may be extremely subjective; the expected standard is difficult
to define; and the amount of new inputs required is impossible to quantify. To illustrate, let us
refer once more to a formalized social system in which deviant behavior is controlled through a
process of observed violation of the existing law (sensing), court hearings and trials (comparison
with standard), incarceration when the accused is found guilty (correction), and release from
custody after rehabilitation of the prisoner has occurred.[6]
The speed limit established for freeway driving is one standard of performance that is
quantifiable, but even in this instance, the degree of permissible variation and the amount of the
actual variation are often a subject of disagreement between the patrolman and the suspected
violator. The complexity of our society is reflected in many of our laws and regulations, which
establish the general standards for economic, political, and social operations. A citizen may not
know or understand the law and consequently would not know whether or not he was guilty of a
violation.

Most organized systems are some combination of man and machine; some elements of control
may be performed by machine whereas others are accomplished by man. In addition, some
standards may be precisely structured whereas others may be little more than general guidelines
with wide variations expected in output. Man must act as the controller when measurement is
subjective and judgment is required. Machines such as computers are incapable of making
exceptions from the specified control criteria regardless of how much a particular case might
warrant special consideration. A pilot acts in conjunction with computers and automatic pilots to
fly large jets. In the event of unexpected weather changes, or possible collision with another
plane, he must intercede and assume direct control.[4]

Organizational and Operational Control

The concept of organizational control is implicit in the bureaucratic theory of Max Weber.
Associated with this theory are such concepts as "span of control", "closeness of supervision",
and "hierarchical authority". Weber's view tends to include all levels or types of organizational
control as being the same. More recently, writers have tended to differentiate the control process
between that which emphasizes the nature of the organizational or systems design and that which
deals with daily operations. To illustrate the difference, we "evaluate" the performance of a
system to see how effective and efficient the design proved to be or to discover why it failed. In
contrast, we operate and "control" the system with respect to the daily inputs of material,
information, and energy. In both instances, the elements of feedback are present, but
organizational control tends to review and evaluate the nature and arrangement of components in
the system, whereas operational control tends to adjust the daily inputs.

The direction for organizational control comes from the goals and strategic plans of the
organization. General plans are translated into specific performance measures such as share of
the market, earnings, return on investment, and budgets. The process of organizational control is
to review and evaluate the performance of the system against these established norms. Rewards
for meeting or exceeding standards may range from special recognition to salary increases or
promotions. On the other hand, a failure to meet expectations may signal the need to reorganize
or redesign.[7]

In organizational control, the approach used in the program of review and evaluation depends
on the reason for the evaluation — that is, is it because the system is not effective (accomplishing
its objectives)? Is the system failing to achieve an expected standard of efficiency? Is the
evaluation being conducted because of a breakdown or failure in operations? Is it merely a
periodic audit-and-review process?
When a system has failed or is in great difficulty, special diagnostic techniques may be required
to isolate the trouble areas and to identify the causes of the difficulty. It is appropriate to
investigate areas that have been troublesome before or areas where some measure of
performance can be quickly identified. For example, if an organization's output backlog builds
rapidly, it is logical to check first to see if the problem is due to such readily obtainable measures
as increased demand or to a drop in available man hours. When a more detailed analysis is
necessary, a systematic procedure should be followed.[7]

In contrast to organizational control, operational control serves to regulate the day-to-day

output relative to schedules, specifications, and costs. Is the output of product or service the
proper quality and is it available as scheduled? Are inventories of raw materials, goods-in-
process, and finished products being purchased and produced in the desired quantities? Are the
costs associated with the transformation process in line with cost estimates? Is the information
needed in the transformation process available in the right form and at the right time? Is the
energy resource being utilized efficiently?

The most difficult task of management concerns monitoring the behavior of individuals,
comparing performance to some standard, and providing rewards or punishment as indicated.
Sometimes this control over people relates entirely to their output. For example, a manager might
not be concerned with the behavior of a salesman as long as sales were as high as expected. In
other instances, close supervision of the salesman might be appropriate if achieving customer
satisfaction were one of the sales organization's main objectives.

The larger the unit, the more likely that the control characteristic will be related to some output
goal. It also follows that if it is difficult or impossible to identify the actual output of individuals,
it is better to measure the performance of the entire group. This means that individuals' levels of
motivation and the measurement of their performance become subjective judgments made by the
supervisor. Controlling output also suggests the difficulty of controlling individuals' performance
and relating this to the total system's objectives.[7]

Problems of control
The perfect plan could be outlined if every possible variation of input could be anticipated and if
the system would operate as predicted. This kind of planning is neither realistic, economical, nor
feasible for most business systems. If it were feasible, planning requirements would be so
complex that the system would be out of date before it could be operated. Therefore, we design
control into systems. This requires more thought in the systems design but allows more
flexibility of operations and makes it possible to operate a system using unpredictable
components and undetermined input. Still, the design and effective operation of control are not
without problems.

The objective of the system is to perform some specified function. The purpose of organizational
control is to see that the specified function is achieved; the objective of operational control is to
ensure that variations in daily output are maintained within prescribed limits. It is one thing to
design a system that contains all of the elements of control, and quite another to make it operate
true to the best objectives of design. Operating "in control" or "with plan" does not guarantee
optimum performance. For example, the plan may not make the best use of the inputs of
materials, energy, or information — in other words, the system may not be designed to operate
efficiently. Some of the more typical problems relating to control include the difficulty of
measurement, the problem of timing information flow, and the setting of proper standards.[7]

Measurement of Output

When objectives are not limited to quantitative output, the measurement of system effectiveness
is difficult to make and subsequently perplexing to evaluate. Many of the characteristics
pertaining to output do not lend themselves to quantitative measurement. This is true particularly
when inputs of human energy cannot be related directly to output. The same situation applies to
machines and other equipment associated with human involvement, when output is not in
specific units. In evaluating man-machine or human-oriented systems, psychological and
sociological factors obviously do not easily translate into quantifiable terms. For example, how
does mental fatigue affect the quality or quantity of output? And, if it does, is mental fatigue a
function of the lack of a challenging assignment or the fear of a potential injury?

Subjective inputs may be transferred into numerical data, but there is always the danger of an
incorrect appraisal and transfer, and the danger that the analyst may assume undue confidence in
such data after they have been quantified. Let us suppose, for example, that the decisions made
by an executive are rated from 1 to 10, 10 being the perfect decision. After determining the
ranking for each decision, adding these, and dividing by the total number of decisions made, the
average ranking would indicate a particular executive's score in his decision-making role. On the
basis of this score, judgments — which could be quite erroneous — might be made about his
decision-making effectiveness. One executive with a ranking of 6.75 might be considered more
effective than another who had a ranking of 6.25, and yet the two managers may have made
decisions under different circumstances and conditions. External factors over which neither
executive had any control may have influenced the difference in "effectiveness".[7]

Quantifying human behavior, despite its extreme difficulty, subjectivity, and imprecision in
relation to measuring physical characteristics is the most prevalent and important measurement
made in large systems. The behavior of individuals ultimately dictates the success or failure of
every man-made system.

Information Flow
Oscillation and Feedback

Another problem of control relates to the improper timing of information introduced into the
feedback channel. Improper timing can occur in both computerized and human control systems,
either by mistakes in measurement or in judgment. The more rapid the system's response to an
error signal, the more likely it is that the system could overadjust; yet the need for prompt action
is important because any delay in providing corrective input could also be crucial. A system
generating feedback inconsistent with current need will tend to fluctuate and will not adjust in
the desired manner.

The most serious problem in information flow arises when the delay in feedback is exactly one-
half cycle, for then the corrective action is superimposed on a variation from norm which, at that
moment, is in the same direction as that of the correction. This causes the system to overcorrect,
and then if the reverse adjustment is made out of cycle, to correct too much in the other direction,
and so on until the system fluctuates ("oscillates") out of control. This phenomenon is illustrated
in Figure 1. “Oscillation and Feedback”. If, at Point A, the trend below standard is recognized
and new inputs are added, but not until Point B, the system will overreact and go beyond the
allowable limits. Again, if this is recognized at Point C, but inputs are not withdrawn until Point
D, it will cause the system to drop below the lower limit of allowable variation.[3]
One solution to this problem rests in anticipation, which involves measuring not only the change
but also the rate of change. The correction is outlined as a factor of the type and rate of the error.
The difficulty also might be overcome by reducing the time lag between the measurement of the
output and the adjustment to input. If a trend can be indicated, a time lead can be introduced to
compensate for the time lag, bringing about consistency between the need for correction and the
type and magnitude of the indicated action. It is usually more effective for an organization to
maintain continuous measurement of its performance and to make small adjustments in
operations constantly (this assumes a highly sensitive control system). Information feedback,
consequently, should be timely and correct to be effective. That is, the information should
provide an accurate indication of the status of the system.[3]

Setting Standards
Setting the proper standards or control limits is a problem in many systems. Parents are
confronted with this dilemma in expressing what they expect of their children, and business
managers face the same issue in establishing standards that will be acceptable to employees.
Some theorists have proposed that workers be allowed to set their own standards, on the
assumption that when people establish their own goals, they are more apt to accept and achieve
them.

Standards should be as precise as possible and communicated to all persons concerned.

Moreover, communication alone is not sufficient; understanding is necessary. In human systems,
standards tend to be poorly defined and the allowable range of deviation from standard also
indefinite. For example, how many hours each day should a professor be expected to be available
for student consultation? Or, what kind of behavior should be expected by students in the
classroom? Discretion and personal judgment play a large part in such systems, to determine
whether corrective action should be taken.

Perhaps the most difficult problem in human systems is the unresponsiveness of individuals to
indicated correction. This may take the form of opposition and subversion to control, or it may
be related to the lack of defined responsibility or authority to take action. Leadership and positive
motivation then become vital ingredients in achieving the proper response to input requirements.

Most control problems relate to design; thus the solution to these problems must start at that
point. Automatic control systems, provided that human intervention is possible to handle
exceptions, offer the greatest promise. There is a danger, however, that we may measure
characteristics that do not represent effective performance (as in the case of the speaker who
requested that all of the people who could not hear what he was saying should raise their hands),
or that improper information may be communicated.[3]