WEEK 1

Required reading:
Chapter 1,2,3 – Section 1: Getting Started section

Lecture 1

Walmart Pickup advertisement: https://www.youtube.com/watch?v=whpJBY5W7xo

Question: What is your perception of this commercial?  Qualitative Research

What do you need to become a Qualitative Researcher?

 Interest in process and meaning
 Be curious what people think and what are the motivation behind their
actions
 Critical approach to life and knowledge
 Why do you think so and how this influences your behavior? How and
why do people say things?
 Eyes and ears
 Observation - listening and observing what is happening
 Reflexivity
 Reflecting on your own role as a researcher - insider/outsider, what is
my point of view
 Good interactional skills
 Communicating with others (you research PEOPLE)
 
What is qualitative research?
 Words as data (written and spoken message)
 About perspective
 Meanings
 Why and how this affects behavior - interpret it
 Rich data
 Not about big numbers of participant but their insights!
 Theory gathering (Table 1.1, p.4)
 Instead of testing theory
 
Two approaches of Qualitative Research

Big Q approach Small Q approach

Paradigm Techniques

Perspective Mixed method

Brought approach more common in CS

 
“Big Q qualitative research: the application of the qualitative methods of data collection and analysis
within a qualitative paradigm, rather than a positivist one. (see Small q qualitative research)”

“Small q qualitative research: the use of qualitative methods of data collection within a positivist or
essentialist paradigm; the occasional use of qualitative questions within a primarily quantitative
method of data collection. (see Big Q qualitative research)”

History QR in CS:
 Interdisciplinary
 Psychology
 Sociology
 Political science
 QR in CS tracks
 
10 fundamentals of QR: Ch. 2
1. Qualitative research is about meaning, not numbers
 Try to really understand people and what they mean, think and their perspective
and why they think what they think?
2. Qualitative research doesn’t provide a single answer
 Every researcher finds different answers from the same data.
 Be creative and interpret it! Can be a bit vague
 Different researcher=different background/perspective= different
theory/structure and interpretation
 Research is subjective!
3. Qualitative research context as important
 People might have different perspectives – know different context.
 Different backgrounds and situations = different reception and perception of
commercial, news etc.
 Subjectivity
 QR recognises the subjectivity of the data we analyse, and the analyses we
produce.
 Two levels of subjectivity:
 subjectivity of the data
 subjectivity of the analysis
“Subjectivity basically refers to the idea that what we see and understand reflects our identities and
experiences – the contexts we’ve existed in, a concept sometimes also referred to as ‘perspectival
subjectivity’ (Kvale, 1996: 212). Qualitative research does not treat this subjectivity as bias to be
eliminated from research, but tends to involve contextualized analysis, which takes this into
account.”

4. Qualitative research can be experiential or critical

 Two different perspectives:
 Experiential – focus on the perspective and experience of customers ->
verstehen = understanding of the people, stepping into their shoes. Used in CS
 Critical - do not use the words of the participants but see what is behind their
answers, behind the language they use, what are the cultural notions - cultural
background etc. - see what is behind the perspective align with your
background, what other hidden meanings are in your words. Sociological
perspective

5. Qualitative research is underpinned by ontological assumptions

 Ontology - if you think reality exists separately from human
understandings- is reality a part of us?
 Realism
 Only one truth out there - with appropriate research we
can reach it and conclude on it. Quantitative approach,
each researcher will come to the same conclusions. One
reality, independent of our subjective ideas.
 Relativism
 Multiple constructive realities, what is true is dependent
on time and context, very relative. Everybody has their
own reality, thus the perspectives are different. Very
difficult to communicate and get research together.
 Critical realism
 There is this real truth but it is behind our subjective
perspective, our knowledge. One reality but is
differentiated by the different perspectives.
 
6. Qualitative research is underpinned by epistemological assumption
 Epistemology -where does knowledge come from
 Positivism
i. Straightforward relationship between knowledge and the
word - qualitative research approach.
 Constructionism
i. Knowledge is constructed of our meanings, experiences
and perspectives. Knowledge is a product of who we come
to understand it, different knowledges from different
perspectives.
 Contextualism
i. There is a truth, but it is local and provisional. It may
change in 5 years or from person to person. Context
shapes how we see the reality and knowledge.
7. Qualitative research involves a qualitative methodology
 Different levels of conducting QR , Level:
1. Paradigm - big, brought perspective on thinking about
reality and knowledge - different perspectives and realities
and knowledge of each person. shaping how you see
yourself as a researcher. how we see the world
2. Methodology – give guidelines in collecting and analyzing
data and methodology, kind of theoretical framework that
guides us when conducting qualitative research
3. Methods/Techniques - the way you actually collect your
data - focus groups, content analysis etc.

8. Qualitative research uses all sorts of data

 We use text, spoken language, photos, movies, commercials etc. When
is this data useful for us?
 Rich data linked to our research - relevant and useful. May be of any
kind - words, texts, photos, movies, ads etc. why do people have this
experience?
 kind of data suitable to answer RQ
9. Qualitative research involves 'thinking qualitatively'.
 Get rid of quantitative thinking. Use your ears - to listen more and
understand the meaning, to deepen into experience, feelings and
 perceptions of interviewers. Find a structure in the pile of data, try to
ask yourself, we generate theory
10. Qualitative research values subjectivity and reflexivity
 Bring your own backgrounds into the research.
 Subjectivity = subjective perception of data and analysis
 Reflexivity - perception and opinions about my subjectivity, what is the
influence of it, how far does it affect my standpoint and research, who
do I sample and how do I get my data, how subjective am I in this
research process?
Personal reflexivity in research is about bringing the researcher into the research, making us
visible as part of the research process – unlike in quantitative research, where the researcher is
typically invisible (that robot).
 
Research question (Ch. 3)
 should focus on how and why people perceive things is a certain way.
 
Sampling (Ch.3)
 Saturation
 When additional data fails to generate new information.
 No exact number of people. Saturation happens when new people do
not provide new insights to ensure that all possible views and
perspectives are collected. Overlap of insights.
 Purposive sampling
 You select people on the basis that they will provide rich data - ones we
know will be relevant and helpful.
 Sampling strategies (box 3.3- chapter 3 ) MCRS: W.6A
 
Cyclical process - during data analysis you go back to data collection or sampling and improve it, re-
sample etc. go back to previous steps and re-do them
 Research question
 Sensitizing concepts
 concepts based on previous research but are very open and guide our
data collection and data analysis; BE OPEN FOR NEW INFORMATION
 Open mind but not empty head - we know things based on our pervious
readings and we just get some guidance from the concepts, but we are
open for the insights of the interviewers.
 Tabula rasa
 Blank mind without any knowledge is impossible
 The mind is open for new insights and can be independent. We cannot
have the blind mind with previous knowledge, but we do not stick to it.
 Purposive sampling/ data collection
 Analyses
 
Ethics in Qualitative research
 Interviewing people - be careful and make sure are anonymous, they know their
rights, they experience no harm etc.
 participants are anonymous
 you do no harm
 participants can stop the interview and withdraw their data
 
 Micro Lectures Chatbot Project

 1) QR Introduction Chatbot Project

Chatbots in customer service: How do customers experience this customer service?
Chatbot is a customer service
 Uses natural language
 Dialogue
 Mimics human to human communication

Qualitative interview study

How do customers experience chatbot communication in a customer service context?

Theoretical background: Human-Machine Communication

 Machines as communicators
 Creation of meaning among humans and machines - sharing a dialogue
 New types of communication  theory development!
 
Possible sensitizing concepts (theoretical concepts form the literature to guide the interview
study)
 Source orientation
 Who or what people think they are communicating with
 Determining the source of a message may be more complex given that
people are simultaneously interacting with multiple human and
technological LAYERS that can be perceived as sources.
 Anthropomorphism
 The assignment of HUMAN traits and characteristics to
computers/media
 Experiencing the machine as human-like
 Anthropomorphic cues: informal language, human name and dialogical
cues (hello)
 mindful/mindless anthropomorphism and social presence

 Motivations
oWhy people use media or how they use media to fulfill specific needs
oUses and gratification approach
 Perceived security and privacy
o Security: To trust a chatbot in customer service, the user needs to be
sure that the security level of the chatbot is sufficient
 BUT what is security for you?
o Privacy: how their personal data from the interaction with the chatbot is
used and stored
 Expectancy violations
 Technology acceptance model: perceived ease of use, etc.
 
Main idea of the study
  To evaluate the experience of customers with the chatbot, introduction of it and
conversational design possibly leading to its improvements

Practical deliverables

YOUR role in this chatbot project

- you as interviewer
- you as analyser and presenter

Theme 1: Source Orientation in chatbot communication in a customer service context

Source orientation
- who or what people think they are communicating with (Guzman, 2019, p. 344)
- determining the source of a message may be more complex given that people are simultaneously
interacting with multiple human and technological “layers” that can be perceived as sources; e.g.,
programmers, applications, hardware, interfaces (Guzman, 2019, p.364)

Qualitative interview study by Guzman (2019)

- voice-based agents in mobile phones
- interview guide and analysis similar to our study
- results:
1. The voice of the machine = communicating with the mobile device
2. The voice in the machine = communicating with an assistant separate from the mobile phone

Our chatbot project

RQ: Who or what do people think they are communicating with when using a chatbot in customer
service?

Theme 2: Anthropomorphism in chatbot communication in a customer service context

Anthropomorphism
- the assignment of human traits and characteristics to computers/media
- experiencing the machine as humanlike

Experiment with chatbots by Araujo (2018)

- Anthropomorphic cues: informal language + human name +dialogical cues (“hello”)
- Scales (1-7 or 1-10) – we do not know why they feel as they feel
Mindful anthropomorphism: humanlike or machinelike, etc.
Mindless anthropomorphism: likeable, sociable, friendly, personal
Social presence: how much did you feel as if you were interacting with an intelligent
being? … as if you were alone
Our chatbot project
RQ: In what ways do customers experience chatbot communication in a customer service context
as humanlike?

4) Theme 3: Motivations in chatbot communication in a customer service context

Motivations
- why people use media, or how people use media to fulfil specific needs
- uses and gratifications approach

Qualitative analysis by Brandtzaeg & Følstad (2017)

- motivations to use chatbots (broader than customer service)
- online questionnaire with open-ended question; qualitative analysis
- results:
1. Productivity
- easy, speed and convenience
- to obtain help and information
2. Entertainment
3. Social and relational purposes
4. Novelty
5. Other motivations

Our chatbot project

RQ: Why do people use chatbot communication in customer service?

Theme 4: Perceived security and privacy in chatbot communication in a customer service context

Perceived security and privacy

- perceived security: to trust a chatbot in customer service, the user needs to be sure that the
security level of the chatbot is sufficient
- perceived privacy: how their personal data from the interaction with the chatbot is used and stored
Qualitative interview study by Følstad, Nordheim & Bjørkli (2018)
- what makes users trust a chatbot for customer service?
- qualitative interview study
- results:
1. factors concerning the chatbot
2. factors concerning the service environment (perceived security and privacy)

Our chatbot project

RQ: How do people perceive security and privacy in chatbot communication in a customer service
context?

WEEK 2
Required reading:
Chapter 4,5,7 - Section 2: Successfully collecting qualitative data

Lecture 2
Data collection methods in Communication Science
 Individual interviews
 Focus groups
 Participant observation (micro lecture)
 Qualitative content analysis (micro lecture)

Interaction between Mary Ann Jolley (Aljazeera journalist) and Najib Razak (interviewee):
- the interviewer has a lot of knowledge on the topic and can build up questions on what Mr. Najib
says
- she cuts him off in meaningful points to keep Mr. Najib on track to get the information she wants
- she is persistent in her questions
- she is friendly but professional
- she does not give up still tries to get information on the topic she is interested
- Mr. Najib stands up because he feels attacked, he does not want to sit  he does not want to give
up (be inconsistent in his statement)
- data from this interview could be richer
 
What are interviews ?
 Semi-structured = participant-led interview
 Interviews used for theory building
 use the flow of interview to get more information

 Pros
 Interviewer structured the interview with topic/guideline list
  Is structured and gives control power to the interviewer of what data is
collected etc.
 A lot of open space for the interviewee to say their story - verstehen
 Super flexible for the researcher to use answers as the basic for the
further question (especially if the interviewer is new to the field)
 Cons
 Super time consuming (at least 1h of the talk and then the transcript)
 Lack of anonymity (you see the interviewee)
 With sensitive issue, not ideal situation - may be a bit awkward (maybe
focus group is better)

 
Interview guide
 Why?
 Structures the interview
 Gives you the control power over what has to be discussed and collected
 Helping tool in times of concern or 'blackout'
 Type of questions
 Opening and closing questions

 Sequencing of question
 In the end it is a conversation so can be a mess as the interviewee does not
know the structure and future topic
 Funnel questions
  Constructing and wording questions
 Prepare open-ended questions so that the interviewee can freely describe
own experience/opinion in own words, when you are nervous or
inexperienced you can prepare backup questions
 Let it open, do not push much into creating questions (except for tips or
background questions for help)

 Probing -> encouraging the interviewee to further talk

 Asking follow-up questions (can you explain? what do you mean by that?)
 Being silent
 Nodding (mhm, mhm)

 Research questions
 What is the goal of your interview? what kind of information from the
interviewee do you need to answer your RQ?
 THE THEORETHICAL RQ is linked to interview goal (what kind of data I need
to collect?) that affects constructing the guide (what are the important
topics/concepts in this situation?) and finally questions
 

Interview guide - example format with most important topics being discussed
 First topic is a warm-up to get the interview started
 Then get straight to the point
What can be prepared? (Ch.4)
Preps per topic: Types of knowledge per topic:
- goal = what is relevant for my research topic? - description of behaviour
- introduction = start with descriptive information - knowledge
- initial question - experiences
- (probes) - opinions
- rounding up - feelings
- motivations

Start superficial but go more in depth later on as we want to focus on the interviewee perception!
Interview guide during the interview: (make notes during interview)
 
Evaluation per topic - reflect on the interview - to improve in future
 Relevance
 Is the information even relevant for me right now? Navigate the
interview as you want to !
 = try to guide the interviewee when one does not give relevant
information
 Clarity
 Do I really understand the answer correctly? Ask for clarity
 Completeness
 Important to have a complete picture of interviewees previous
experience, background etc.
 Validity
 Refers to the truthiness of the answers- is it really based on real
experiences or just lies?
 
 Probing
 
Good questions are…
 Open
  Non-leading
 Singular
 Short
 Clear and precise
 Linguistically appropriate
 Non-assumptive
 Empathetic
Types of questions

Participant issues (Ch. 4, pp. 65-66)

 Struggles with finding specific people and interviewing them etc.
 Issue with interviewing the people you know - ethical issues may arise as you know
the background you are not asking all questions, may be awkward to ask confidential
questions etc.; better  outside from your social network
  Power issue - sometimes participants think you are an expert and think you know it all
- be open and on the same page as the interviewee - say you just started etc. = the
interviewee sees you as an expert  try to build a natural conversation, become a
team
 
Preparations for the interview
 Memorize your interview guide
 Planning
 Location
 Recording and note-taking
 Personal disclosure
 Informed consent
 
What are focus groups? (Ch. 5)

Ted Ed Focus Groups: https://www.youtube.com/watch?v=3TwgVQIZPsw

Focus groups
 Data collected from multiple participants at the same time - gaining insight from
multiple sources - the reasoning or unconscious motivations to perform certain
behavior or even finding new ideas for products and learning what people want
 6-10 people recruited according to the criteria and objectives of the researchers to get
the feelings of group
 2 or more moderators or researcher may be hidden in a group or observed from
behind the glass
 Pros
 Saving time and money
 Gaining insight from multiple people at once
 With sensitive issues it can help if others have the problem as well - you
do not feel ashamed or alone and are more open
 Interaction between people (sometimes you study the interactions
itself!)
 Useful while exploring new topics or ideas
 No power issues -> more power for the participants (outnumber the
moderator) and they structure the conversation - bigger role of
participants
 Cons
 Observer-interference
 Answers given can be influenced by the fact they are observed, wanting
to give a socially desirable answer, being influenced by others ideas and
perceptions or they know they are a part of focus group
 This makes the results hard to generalize - thus the knowledge gained
from focus groups are used in experiments (e.g. how much are they
actually willing to pay)
  Does not allow you to elaborate on specific person in 100% as there
are others in the room but then you can make focus groups selected by
the issue (e.g. cancer patients having sex issues - feel more safe and
empowered to say their experiences and stories)
 May be difficult to moderate and organize

 
Data from a focus group is different than the semi-structured collected data! They provide insight
through brainstorming and their perspectives and ideas about the product or solution to the
issues. - nice to use when exploring new topics. Focus groups are more about participants power
over the conversation whereas the interviews give the power to the interviewer.
 
Participant issues
 What kind of people do I need in my focus group?
 Heterogeneity
 consists of very different kinds of people - e.g. doctors, unemployed,
customers etc.
 Advantage - insight into diverse perspectives thus more diverse
conversation
 Heterogenic groups (people are different from each other – e.g., age, gender,
occupation) = adv: more diverse discussion;
 Homogeneity
 they are like each other - e.g. all suffer from cancer or are in the same
age
 Advantage - greater insight into the specific target group, more
comfortable thus more open conversations
 Homogenic groups (people are similar) = adv: people feel more comfortable with each
other

 Sample size
 Ideal = 8-10 people scheduled, possible cancellations leaving you with 5-
6 people
  
Preparations for focus groups
 Guide

 Recruiting and organizing

 Why do we want this specific kind of people?
 Location
 

 
Difference from semi-structured interview- instead of just asking questions to evoke the reactions
you also show the product/materials to improve and deepen the conversation and reactions
 
Conducting focus groups
1. Introducing ground rules
2. Interviewing skills

3. Group management/moderation skills - controlling for every person having chance to

speak, engage the more shy as well and quiet the ones who constantly talk
 
Transcription (Ch. 7)
 Orthographic/ verbatim
o Everything that is being said is included in transcript! Even hmmm, okay
o When you don’t hear what they say write - inaudible audio!
 Interviewee is NOT interviewer
 Anonymous
 8h per 1h interview
 
People don’t talk in sentences ! - table 7.1 and box 7.1 from chapter 7

 
Find the balance - you need the data for analysis but you only want relevant data !

 
  

WEEK 3
Readings:
 Chapter 8: general info (pages 173-174) + GT (pages 176, 184-187). NOT included:
pages 175, 177-183, 187-198.
 Chapter 9: general info (pages 201, 204-207, 210-211, 218-220) + GT (pages 214- 216).
NOT included: pages 208-209, 212-213, 214 (IPA), 217-218.
 Chapter 10: general info (pages 223-230) + GT (pages 239-242). NOT included: pages
231-238, 242 -245.

Lecture 3
About last weeks…
Week 1 – Getting started Week 2 – Data collection
- what is QR? - semi-structured interviews
- subjectivity and reflexivity - focus groups
- sampling - transcribing
- cyclical process

Week 3 - Data analysis

Data Analysis
 
What is analysing? (Ch.8)
 Descriptive vs interpretative analysis
 We want to hear the voice of the participants- we want to use their
words and how they see the chatbots, we do not want to interpret their
thoughts- we do not want to go 'underneath' their words(interpretative)
 descriptive: we focus on experiences of our interviewees, we give them
voice
 interpretative: we focus on what is hidden below the surface of the
words they use
 Perspectives
 Thematic analysis (TA)
 Interpretative phenomenological analysis (IPA)
 Grounded theory (GT) - only those questions in the exam
 Disclosure analysis (DA)
 
GT: Why this approach? (Ch.8)
 Hugely popular and influential in social sciences
 Aim: develop a theory that is grounded in data - data is the start for coming up with a
theory. We use what we found out from the participants to build the theory from
ground up
 
Brief history

 
GT: How does it work? (Ch.8)
o ‘Full’ GT or ‘GT lite’
 Full GT - how we see reality? what do think that the knowledge comes
from? aim: build a completely new theory; your OWN everything and
building a completely new theory; time and work consuming – highly
demanding
  GT lite - we use some procedure and aspects of building a new theory
from data; we use some basic ideas but overall use existing theories and
studies, procedures, etc.
 Distinctive procedures
 Line by line coding
 Interpret data and summarise it; data reduction; analyse
the transcript line by line
 When you code - you do it line by line and through the
whole document, very precise.
 Constant comparative analysis
 You have to constantly go back to codes/data/steps,
merge them, improve or change them
 Memo writing
 Writing things that you think/decisions/steps during the
research process (logbook)
 Very important for the interpretation of your research and
following it
 Theoretical sampling
 Sampling related to RQ and to first answers from my data
analysis.
 Recruit a sample of participants (e.g. 5) and after
interviewing them you analyze the data, and then you
recruit again the sample of people who either confirm or
disconfirm the data/what you have found, or simply
random people
 Based on your first sample set and data you sample
further - even 4 or 5 rounds.
 Saturation
 Point of the research project when new data do not bring
any new findings
 E.g., select 10 participants and by 8th you reach saturation
(the concepts align)
o Open mind
 Open for the data you collect, and you are open for the collected data -
not an empty mind but have an open mind!
 
How-to guide
1. First steps: open coding (chapter 9)
2. Identifying structure: CIM (chapter 10)
From data to theory
1. First steps: open coding (chapter 9) - what do we do with our data
 Familiarization and coding
 Familiarization - becoming familiar with the data / transcript, reading the transcripts – YOU
KNOW WHAT WAS SAID DURING THE INTERVEIW.
 What is coding? - process identifying important and relevant parts to my RQ, summarizing
the information from the interviewee (key words, codes) to summarize the most important
things!
 Code=label of the most important parts = you add specific key words to summarise big pile
of data
o Selective coding: coding information to answer e.g. only a part of the RQ - more
selective in finding codes for relevance (e.g. for answering sub-RQ)
o Complete coding: you code everything in your transcript - everything can be
relevant for the RQ , even if at first sight do not seem so. (we do this one)
 CAQDAS- software used in coding (here, Atlas.ti)

What makes a good code?

 Data-derived vs. researcher-derived codes
- data-derived codes: codes that are very close to your data (using words of the interviewee)
- researcher-derived codes: you as a researcher interpret and add a label to what is said -
interpretation to specific parts of the interview
 Essential when coding:
1. RQ = mind your RQ when coding - codes must be relevant and can be related to RQ
2. Several codes per relevant segment = a segment is 10-12 lines - paragraph, which cannot
be summarised by only one key code - not possible to summarize with one key code as
you must include background information of the interviewee - thus some segments may
require several codes
3. Be concise = use key words; don’t use long sentences
4. Distinct  merging codes throughout the transcript = go back and forth when coding
5. Inclusivity = include everything discussed in line with interview guide - everything will be
coded as everything can be relevant
6. Different facets = different data (and codes) within groups that are super important in
the qualitative research! Include variety between interviewees, unique perceptions

GT: Coding (Ch.9)

- Initial coding/Open coding = be as open as possible and have an open mind for what is
being said by the interviewee
- Constant comparison = merging, constantly going back and forth in the work, changing
codes constantly
- In vivo/member codes = data-derived codes; you use the language of your interviewee
- Memo writing = important data, why do I code in this way? Why do I use this specific codes?
Why do I code those specific parts?

2. Identifying structure (ch.10): CIM (chapter 10) - focus on summarized data and identifying
patterns/models etc. in our data
 From codes to themes - after the coding, you try to find broader category/perspectives - there
are several codes/facets in one theme - there is some variety in themes e.g. positive feelings
while communicating with a chatbot (familiarity, personal, friendly codes)
 Active process - some themes are obvious; some occur later when coding
 Candidate themes: be prepared to let go = find another theme or skip; in the end only 3 themes
stay!
 Meaningful and important in light of RQ - relevant to the data and RQ, must be linked to RQ
 Faithful to data - embedded in the data, we use data as the basis to build our theory
 
  
GT: Identifying structure
 Open coding - read the whole segment and think: what is my RQ, what is relevant?
 Open coding is no longer productive? shift to Focused coding = way of selective
coding, you only code what is relevant to the RQ, coding with categories in mind and
trying to align them (not relevant) - Axial coding = coding with categories and
structures in mind and trying to connect data
 Developing a model = to present patters/themes found in the data; How we can
relate it to the data? Base for the theory
 
GT: CIM
 Concept indicator Model
 Proposed structure for your analysis
 Why? Focus on variety in your data (including the various views of the
interviewees)
 

 
 
Different levels of the format
Level:
1. Concept is linked to the indicators and the indicators are super close to the data
- Often the RQ or sensitizing concept or part of the RQ
- After coding try to see what the main concept is
2. Dimensions - themes/categories/chapters which vary- they are variables in which the
variety is (e.g. positive and negative feelings while using the chatbots)
3. Indicators- how do people express the dimensions (so positive or negative feelings) ->
with words, how do people express themselves in terms of this specific dimension;
they show the variety how people express a specific dimension

  Example: Meaning of TV in life of elderly people

Proposed structure for your analysis

- Why? Focus on variety in your data

How to develop your CIM

 Choose your concept - what is the most important in the interviews?
 Work towards dimensions first - what topics were the most important/insightful for
us? And then…
 …Re-examine the codes within the dimension to find variation- go back to the
interviews to use the words of the interviewee
 
A good CIM
 Answers part of the RQ or the whole RQ- must be a link to RQ
 Is grounded in data - data is the starting point for the model
 Encompasses all material related to the concept - everything that is being told to us,
is included in the model, showing the variety of expressions of the interviewees
 Includes appropriate labels for the concept, dimensions, indicators - correct
formulation and summary
 Has a logical structure
 
The qualitative research processes EXAM ALERT!!!
 RQ
 Sensitizing concepts
 Sampling
 Data collection (F2F+OML)
 Data analysis (F2F +OML)
 Quality criteria and techniques (OML)
 Reporting your study
 Ethics
3 open questions - including a case, interpreting and creativity.
 
Theoretical for exam:
 Lectures
 Book
 F2f lectures, book, theoretical online micro lectures (5-6-7), see schedule in course
manual
 TedTalk by Tricia Wang

WEEK 4
 
Ted Talk - Tricia Wang The human insights missing form big data
 Big decisions - in ancient times consulting an oracle, nowadays it is big data, "Watson",
'neutral net', 'deep thinking’ to make a good decisions, with some assurance.
 Despite the size, returns on big data are super low because although investing is easy,
using it is not. Over 73% of big data projects aren't profitable - companies are investing
in it but it does not help the employees to make better decisions or come up with
better ideas.
- $122 billion industry
- Investing in big data is easy but using it is hard
- 73% of big data projects are not profitable

Why is having more data not helping us to make good decisions?

 E.g. Nokia case, when in 2009 there was the research about how low-income people
use technology. Tricia did field work of spending time with people – she lived among
low-income people to understand their use and attitude towards technology and
gained knowledge about informal economy.
 This helped her to see the coming change - promise to enter the high-tech life of
iPhone/knock-offs.
 She started putting data together - holistic picture of what was happening - even the
poorest Chinese would die for an iPhone. Although some people said people would
not use them she thought different.
 Big data did not see this coming - their data points did not indicate that people would
but smartphones, and her intel from like 100 people was not enough for them.
 They were sending out surveys assuming that people don't know what a smartphone
is, so of course they were not going to get any data back about people wanting to buy
a smartphone in two years. Those methods have been designed to optimize an
existing business model, and she with her qualitative approach was looking at these
emergent human dynamics that haven't happened yet. They were looking outside of
market dynamics so that companies could use it and get ahead of it.
 Nokia did not use it and fell of the market. Lots of companies do this and they throw
out data when it did not come from a model or did not fit it. But it is the way we use
big data - this is our responsibility.
 Big data's reputation for success comes from quantifying very specific
environments, like electricity power grids or delivery logistics or genetic code, when
we're quantifying in systems that are more or less contained. 
 But human behavior cannot be contained, as new factors emerge all the time - the
conditions are everchanging. This is a never-ending cycle. Big data makes this illusion
that we already know everything but then the changes suddenly happens- relying only
on big data is a mistake.
 Results:
- Experts did not predict success for smartphones because it wasn’t big data
- It’s not big data’s fault, it’s the way we use big data, quantification takes place in
contained systems
- quantification in dynamic system creates a paradox
- Relying on big data alone increases the chances we’ll miss something, while
giving us the illusion we know everything
 The quantification bias - unconscious belief of valuing the measurable over the
immeasurable. It is a huge mistake of using big data. People can be so fixated on
numbers that they do not see what is outside even if you present them with
reasonable data and evidence.
 Although quantifying is sometimes helpful it can cause problems- it is addicting and
sometimes data can be thrown out just because it is not a number. But often the
future we need to predict cannot be found in numbers but in words actually.
 The ancient oracle had helpers that gather empirical, qualitative information and used
it to make the decision. This is why big data needs the qualitative insights - to see the
reasons behind it.
The temple guides = ethnographers, researchers
 - observe emotional state of the oracle
- ask for the reason to get information
Integrate big data + thick data to form a complete picture = the experience
of understanding the human narrative
Thick data. 
 precious data from humans, like stories, emotions and
interactions that cannot be quantified. It's the kind of
data that comes in in the form of a very small sample size
but delivers incredible depth of meaning. 
 What makes it so thick and meaty is the experience of
understanding the human narrative. And that's what helps
to see what's missing in our models. Thick data grounds
our business questions in human questions, and that's
why integrating big and thick data forms a more complete
picture. Big data is able to offer insights at scale and
leverage the best of machine intelligence, whereas thick
data can help us rescue the context loss that comes from
making big data usable and leverage the best of human
intelligence. And when you actually integrate the two,
that's when things get really fun, because then you're no
longer just working with data you've already
collected. You get to also work with data that hasn't been
collected. You get to ask questions about why: Why is this
happening? 
 By integrating big data and thick data, you not only improved your
business, but you can transform how we consume media and make
market gains of it
 Being collectively committed to making better data, better algorithms,
better outputs and better decisions. This is how we'll avoid missing that
something.
= We are collectively committing to creating better data, better
algorithms, better outputs and better decisions 
- e.g. Netflix recommendation algorithm = ethnographer discovered
that people loved binge-watching  Netflix started offering more of
the same show to encourage binge-watching

WEEK 5
Required reading:
Chapter 12

Micro Lecture 5 – Quality Criteria

What are they? Why important?
- Reliability - Conducting your own study
 - Validity - Evaluating published studies
- Generalisability
How can I ensure that my study is good?
- 8 techniques

Can we apply quantitative quality criteria to qualitative research? (Ch. 12)

Reliability
Different researchers generate the same results when they use the same measures

Quantitative Qualitative
 Trustworthiness of the process
 Procedural reliability being able to follow
 Replicability - finding the same results
the same steps and understand the steps
every time (not possible in qualitative)
taken by researcher
 Dependability

Validity
The study shows what it aims to show

Quantitative Qualitative
 Ecological validity studying the participant
with use of the methods as close as
possible to their daily life to obtain the
most possible natural results
 Construct validity - measuring  Credibility - the findings need to make
the aimed thing sense and be credible
 In qualitative ecological validity should
be good as the researcher goes to the
participants field and his setting.

Generalisability
The extent to which the results can be applied to a wider or different population

Quantitative Qualitative
 Statistical generalisability  Transferability to what extent the
findings also apply to other
contexts/societies/courtiers etc.;
 Aim is not to make it as high as
possible but to make it in such a
detailed way that the reader can
assess for themselves if the
research will also be applicable to
their setting
 “Transferability: the extent to which
qualitative research results can be
‘transferred’ to other groups of people
or contexts. (see Generalisability)”

Why is it important?
 Conducting your own study
 Evaluating published studies

How can it be done?

8 techniques used to enhance reliability, validity and/or generalisability

Reliability Validity Generalisability

Trustworthiness Credibility Transferability
you show your
findings to your
respondents and ask
Member checking
them for feedback by
(respondent
showing them results
validation)
and adjusting the
findings to maximize
the validity
combining research method, e.g. combining
individual interviews for individual insights and
Triangulation
use of focus groups for the group dynamics -
(combination)
more generalizable as increases credibility and
transferability
getting feedback from
Peer debriefing
peer researchers
writing the notes during research process,
Memo writing
eases the credibility and reliability
good solid description of participants and
Thick description phenomenon studied (context and background)
for readers to assess the transferability
good and detailed
method section to
ensure
trustworthiness by good and detailed good and detailed
Method section
following clear and method section method section
detailed description of
steps (assess
transferability)
Prolonged understanding the
engagement participants'
perspective through
 extensive contact with
them - a researcher is
deeply engaged in the
participants
environment
Computer assisted
qualitative data
analysis software-
used for data, coding,
CAQDAS structuring it and
memo writing. use
computer program to
keep track of the
research process

WEEK 6
Micro Lecture 6 – Participant observation
Participant observation
Defninition: field strategy that combines several data collation methods.
1. Participation - doing the same as participants, researcher does what the participants
do
2. Observation - see and listen to what participants do, do not take action; the
researcher goes to the participants = field observation
 3. Informal conversation - engaging into small talk with people around to gain
understanding
4. Interviews - asking people to give a short interview with guides, transcripts, more
formal
5. Document analysis - analyses documents
6. Introspection -reflecting on myself in the field, how my context can impact the
research
 
Purposive sampling - several levels
 Level 1: Selecting cases/settings/organizations - who do I want to investigate?
 Level 2: Select individuals - decisions about which individuals documents I want to
investigate
 Level 3: Select documents to analyse

 Types of data
- Field notes - during observations
- Transcripts of interviews
- Documents
- Audio recordings
- Video recordings
 
Your role in the field
 Open vs. closed field
- Open field: easy access, e.g., music festival, bar – easier access
- Closed field: e.g., newsroom
 Covert vs. overt
- Covert: not saying that you are a researcher, lying, not very ethical, pros: very good
insight, treat you as one of them
- Overt: saying you are a researcher but the question of what you say your research is
about still remains
 Professional stranger
You are a stranger, but you come as a researcher – not a part of the group but a
professional researcher
 
Techniques in participant observation:
1. Triangulation
combination of data collection methods, using interviews, documents, audio etc.
2. Memo writing
writing field notes
3. Thick description
detailed description of phenomena and participants - as you spend a long time in the
field, you should be able to provide the detailed insights
4. Prolonged engagement
 Extensive engagement with the phenomena that you are studying – as you are a
part of the field / participants' group/setting, you are expected to provide extensive
description
 

Possible disadvantage: One negative site  possibility of going native = if you spend too much time
in the field, you can lose your perspective, forget that you are a researcher and professional stranger

Micro Lecture 7 - Qualitative content analysis 

Qualitative content analysis
a method for systematically describing the meanings of media materials (with focus on ComScience)
in search for a specific phenomenon or meanings
 Research questions start with HOW? as you focus on specific phenomena/meaning
 Very flexible method: You can use various media materials: text, audio, visual (in com
science Twits, blocks on specific phenomena, movies to study specific behaviors,
frames etc.)
 Materials used to identify manifest meanings - things you can seeing the text or
picture and for latent meanings - those are a bit covered or in the background
QCA- sources: media materials - purposive sampling
 Selection of relevant materials: which is relevant, which is important for my analysis ?
 RQ is a starting point
 Selection of sampling is changing during the process, it is not fixed beforehand
 
Techniques in QCA:

1. Method section - important why and how you selected materials, how you developed
coding and how you used codes for framing and interpretation
2. Prolonged engagement and member checking not relevant for QCA!!