Analyzing Focus Group Data

Types of Content Analysis

1. Pragmatical Content Analysis: Classifies signs according to their probable causes and effects; why something is said

2. Semantical Content Analysis: Classifies signs according to their meanings

a. Designation Analysis: Frequency with which nouns are mentioned.

b. Attribution Analysis: Frequency with which certain characterizations or descriptions are used; focus on adjectives adverbs and descriptive phrases.

c. Assertation Analysis: Frequency with which certain descriptors are used with nouns. This is a combination of designation and attribution.

3. Sign Vehicle Analysis: Classifies content according to the psychophysical properties of signs. Analyze the underlying concept in addition to the words. Ex. Respondent talks about his penis a lot but uses different words and descriptors.

Doubt and Disbelief

Consider the pros and cons of various data visualizations

The Scissor-and-Sort Technique

https://www.census.gov/dataviz/visualizations/062/

Reason Versus Reactions

https://www.census.gov/dataviz/visualizations/057/

~ Participants say one thing and do another

~ Focus group moderators and analysts need to be

sensitive to situations in which participants expressions may reflect social desirability influences, pressures to conform to groupthink or the persuasive effects of a dominant group member

1) Read Transcript

2) Id sections that are relevant to the research question

3) Code data

4) Describe groups

a. Conceptually

b. Statistically

~ text items/ word strings are grouped and used as supporting materials

https://www.census.gov/dataviz/

~ Researchers are interested in which marketing strategy is better rather

than why it is better

~ Common in marketing studies the researchers have a strong interest in

separating winning from losing product or advertising concepts

~ Such groups tend to involve a lot of voting and ranking from the

participants and they often pay insufficient attention to the various and often subtle reasons for their evaluations.

Individuals Versus the Group

~ Generalizations to the population based on the sample group can be problematic

~ (alternative approach) view everyone in the group as representing a demographic, lifestyle or attitudinal segment, which encourages a within person rather than across person analysis.

Issues

Computer Assisted Content Analysis

Content Analysis

~ Associations among words are often important determinants of meaning

~ Meaning may be related to the frequency of association of certain words, the distance between associated words or concepts (measured by amount of intervening words) and the number of different associations.

~ Way people use language offers insights into the way they organize information impressions and feelings in memory and how they tend to think

~ Language plays a critical role in how people experience the world (Sapir, 1929).

Issue Order

~ Judgement is required to interpret whether the issues that are raised first, truly represent the

participant’s major concerns or are merely mundane and socially safe topics

Issue Absence or Presence

~ Things that go unsaid or are not raised in the discussion may be equally important

~ Issues that the participants don’t address may simply represent things that are taking for granted

~ Others may represent socially sensitive topics that the individual is trying to avoid

~ Other issues may not materialize because the respondent does not find them important

Time Spent on The Issue

~ Allocate blocks of time to topics that will be covered

~ Amount of time spent of a topic is a good indicator to how much participants care about a particular

issue

Intensity of Expression

~ Intensity as well as the Nature and sources of participant’s emotional reactions are analyzed

~ Customer relationship management (CRM) is a modern marketing strategy

~ The consumer’s emotional connections to products and brands is increasingly seen as a key link in

the relationship.

Content analysis: A diverse domain of techniques designed to explore and describe qualitative verbal, written, and multimedia communications in a systematic, objective, and quantitative manner (303).

~ Any technique (a) for the classification of the sign vehicles, (b) which relies solely upon the judgements which theoretically may range from perceptual discrimination to sheer guesses of an analyst or group of analysts as to which sign vehicles fall into which categories (c) provided that the analysts judgements are regarded as the report of a scientific observer (Janis, 1965)

~ Sign vehicle is anything that may carry meaning though most often it is likely to be a word or set of words in the context of a focus group interview.

~ Content Analysis Instruments: Message Measurement Inventory (R.G. Smith, 1978) and Gottschalk- Gleser Content Analysis Scale (Gottschalk, Winget & Gleser, 1969).

Transcribing the Interview

Qualitative Analytic Approaches

~ Outsourcing is a cheap and effective option

~ Interviews are not always complete

~ Often transcriptions are edited for flow which can hide their true sentiment

~ Nonverbal communication gestures are not reflected in the transcript and the tone with which words are used are important sources of information.

~ Sarcasm is difficult interpret with transcripts

~ Transcripts do not capture body language

Transcribe

~ Less consensus on how to analyze and interpret qualitative data is the result of differences in epistemological orientations and phenomenological foci that characterize the behavioral science disciplines

Epistemological Orientation

Disciplinary Focus

Workbench Issues

1) Social Constructivism: much of reality and the meaning and categories that frame everyday life are essentially social creations.

a. Focus groups analysis that reflect this view tend to emphasize how group members collaborate on some issue or how a consensus is achieved.

2) Phenomelogical Approach: Opposite of social constructivism and emphasis on the subjective idiosyncratic perceptions and motivations of the individual respondent

a. Focus groups in which managers are extremely interested in individual differences.

3) Interpretivism: accept prior perspectives but do not take respondents words at face value; body language and facial expressions are more indicative of underlying thoughts.

1) Hermeneutics: Theory and methodology of literal interpretation

~ Content

2) Semiotics: Theory and methodology of literal interpretation plus verbal expressions and all types of pictures sounds produces and advertisements.

~ Form and Conent

~Part of the analysis begins once the group meeting has begun

~ The moderator must terminate a topic, expand the discussion on one that the group finds involving or introduce an entirely new line of questioning

Stages of Content Analysis

Established Measures of Content Analysis

How Much Analysis?

1) Data making (Data Structuring)

~ Unitizing

a. Sampling Units: Those parts of the larger whole that can be regarded as independent of each other

b. Recording Units descriptive subsets of methods being employed.

c. Context Units: Provide a basis for interpreting a recording unit

~ Sampling: Ensure sample is representative of target population.

~ Recording: use of the defined units of analysis to classify the content of the discussion into categories such that the meaning of the discussion is maintained and explicated

2) Data Reduction

3) Inference

4) Analysis

(Krippendorf, 2004, p. 83f):

1) Frequency with which a symbol or idea appears ; interpreted as importance attention or emphasis

2) Relative balance of favorable and unfavorable attributions regarding a symbol or idea; interpreted as a measure of direction or bias

3) The kinds of qualifications and associations made with respect to a symbol or idea; interpreted as a measure of the intensity of belief or conviction

Coefficient of Agreement

~ The most common analysis of focus group results involve a transcript of the discussion and a summary of the conclusions that can be drawn

~ Results are displayed using descriptive and graphical tactics.

~ Focus groups can be evaluative tools

~ Focus groups can provide a tool for testing the reality of assumptions that go into the design of services programs and products.

Word Cloud in R

Pi (scott, 1955)

Krippendorf’s Alpha (Krippenborf, 1970, 2004)

Kappa (Cohen, 1956)

Textbook : pg 270

0 <- no agreement

1 <- perfect agreement

(Fleiss, 1981)

K > .75 <- Excellent

K = .6 - .5 <- good

K = .4 - .59 <- Fair

K < .4 <- poor

Edward Bernays

Analyzing Focus Group Data

Thomas Foley