Contents - Previous - Next

XXIII. Setting Up Focus Groups

1. Importance of Planning: Significant Amount Required, Appointments, Budget, Time Considerations, Providing Incentives, Location should be "neutral," but "familiar"

2. Number Of Focus Groups: Roughly, 3-4/topical area/identified sub-group

XXIV. Focus Group Participants

1. Number Of Participants Per Group: Range: 412 (Ideal: 6-10), Sampling sometimes random (but rarely), Differences in dynamics for small vs. large groups

2. Participant Characteristics: Homogenous (to avoid conflict, refusals to share opinions), Sex, Age, Other Relevant Characteristics (ethnicity, SES, marital status), Inter-person "comfortableness," No "experts"

XXV. Conducting a Focus Group

1. Importance of Focus: Shared Experience

2. Use of a Focus Group Discussion Guide: Similar to EFG, Open-ended questions, However, usually is completed

3. Role of the Moderator/Facilitator: Should not convey impression of being an expert, To Keep the Focus Group Focused, To Help Less-Participating People Speak, To Keep Dominating People from Speaking Too Much, NOT an Interviewer, Can try and determine consensus around an issue

4. Recording The Focus Group: Use of a Tape Recorder/Videotape, The Human Recorder (Observer)

XXVI. Format ("Flow") of a Focus Group

1. Emphasis on Group Interaction/ Conversation/ Discussion

2. Not a Question-Answer Session

3. First Focus Group(s) Spend More Time on a Particular Topic

4. Some Topics too "Sensitive" for Focus Groups

XXVII. Direct Observations: What to Record

1. Observations Give You Actual Behavior, as Opposed to Reported Behavior

2. What to Record (Focus): Who, Where, When, What, Why, "Key Behaviors," What Does Not Happen, Make Maps/diagrams

XXVIII. Types of Direct Observation:

1. Participant-observation, 2. Unstructured Focused Observations ("Scripting"), 3. Continuous Monitoring Observations, 4. Spot Check Observations, 5. Rating Observations

XXIX. Participant Observation

1. Purpose: Exploration, Rapport-Building

2. Alternate Between Roles of Participant and Observer

3. Strategy That Facilitates Data Collection in the Field

4. Skills of a Good Participant-Observer (Language, Explicit Awareness, Building Memory, Maintaining Naivete, Building Writing Skills, Hanging Out, Objectivity)

5. Usually No Notes Taken During Observation

XXX. Unstructured Focused ("Scripting") Observations

1. Purpose: Exploration

2. Emphasis On Note-Taking/Text

3. Scripting Format (Time, Activities in Sequence)

4. Focal Topic Or Subject (Person, Location, Event)

5. May Have Guides (similar to EFG, FGDG)

XXXI. Analyzing Qualitative Data

1. "Qualitative data analysis is the search for patterns in data and for the ideas that help to explain the existence of these patterns" (Bernard, 1994)

2. Selecting A Mode Of Analysis: Within Case, Within Site; Across Case, Within Site; Across Case, Across Site

3. Selecting An Analytic Foci: Chronological, Key Events, Settings, People, Processes, Issues

4. Examples Of Different Products From Different Modes Of Analysis

XXXII. The Qualitative Data Analysis Process: Generic Style

Review Notes (r) Coding (r) Search and Extraction (r) Pattern Identification (r) Creation of Summarization Figures

XXXIII. Pattern Identification

1. Descriptive:

What terms/key words are used to describe the topic?

What are the different types/categories of response?

How frequently are the different types/categories of response mentioned?

Identify key quotes or observations which illustrate main topic/theme.

2. Interpretive:

What is the relative priority of the different types/categories of response?

How is priority expressed?

What is the level of agreement between respondents?

How do these types/categories of response relate to each other within the analysis?

How do these patterns relate to patterns you have identified in other analyses? Is there synergy?

Do data collected using different techniques appear to validate the patterns you see?

Do the data satisfactorily answer your research question? Is there a next step?

XXXIV. Systematic Data Collection Methods: Characteristics

1. Preceded by unstructured data collection methods to determine coherent cultural domains.

2. Used for exploration of Internal Structure of a coherent cultural /cognitive Domain

3. Structured, easy to administer (including non-expert field workers)

4. Reliability high, permits comparison

5. Validity high, IF: Domains properly identified, Respondents familiar with the domains (must be coherent), Questions are comprehensible and unambiguous.

6. Respondents often chosen through random sampling

XXXV. Types of Systematic Data Collection Methods

1. Item Eliciting Techniques: Free Listing, Sentence Completion, and Fill in the Blank

2. Item Grouping Techniques (Proximities): Pile Sorts, Triadic Comparisons

3. Item Ordering Techniques: Rank ordering, Rating scales

XXXVI. Free-Listing: How to do it

1. Purposes: To determine the items in a cultural domain, to differentiate between items that are Culturally Cognizant or Salient and those which are not

2. Free List Primary Question: Pre-test, Consistency, Question formulation: (Don't Ask: "If..." or "Are there...?," Do Ask: "What are all the different kinds of X you can think of?" or Do Ask: "Name all the Xs you know.")

3. Probing Rules: Need to define levels of probing (Min-Max)

4. The Free List Secondary Question: Optional

Purpose: Clarification, e.g., Junk Foods: Why is ****** a junk food?

Purpose: Additional Information, e.g., Women's Illnesses: What should a woman do if she gets ******?

5. Informant Selection Issues: Knowledgeable, Number (Minimum of 10), Stratification by Significant Subgroup

XXXVII. Free Listing: The Importance of Identifying Salient Items

1. Sequence: Free Listing (r) Analysis: Selection of Salient Items (r) Grouping, Rating and Ranking Using the Salient Items

2. Why Important To Select Items Carefully: Threats to Validity with Grouping, Rating, and Ranking:

- unknown or unfamiliar items, e.g., apple, orange, kiwi

- items from different domains, e.g., cow, apple, car

- items at different levels of contrast, e.g., beagle, dog, pet

XXXVIII. Analysis of Free List Data

1. Forms Of Tabulation: In order of decreasing frequency of mention (most common), Ranked by order of mention (salience), Related items (Proximities(?))

2. Collapsing Items: Generally, avoid the temptation to collapse items

When to collapse items: Singular and plural forms of the same word, Key informants clearly state two items are the same thing

3. Ways Of Selecting Salient/Cognizant Items: Frequency of items: draw frequency chart, Order that respondents give items, Related items

XXXIX. The Pile Sort Method

1. Purpose: To get insider's perspective of how terms related to one another; To determine judged similarities, categories, taxonomies; To explore complex domains (with many items)

2. The Pile Sort Primary Question/Explanation: Pre-test, Consistency, Question/explanation formulation: "Please put these different kinds of X into piles; in whatever piles you think are important"

3. Further Explanation Rules: Need to define what is permissible,

4. The Pile Sort Qualitative Explanation: Essential

Purpose: Identification of Subdomains: "Why did you put a, d, and y in the same pile?"

XXXX. Steps in Doing a "Single/Free" Pile Sort

1. Choosing The (Salient/Cognizant) Items to Sort

2. Assessing Literacy of Potential Respondents

3. Preparing the Pile Sort Cards:

Pictures vs. Words vs. Material Objects

Unique Number on Back

4. Making a Data Collection Sheet

5. Initial Explanations

No Right/Wrong Way, but
Each Card Can Only Go in One Pile
All Cards Cannot Go in One Pile
Cannot Have All Cards in Their Own Separate Pile

6. Doing a Demonstration Sort (Optional) - Concern with Leading

7. Introducing the Pile Sort Cards, Lay Them Out, Re-Sort (Re- Randomize Each Time)

8. Doing the Sort/Recording the Information

9. Important: Asking for and Recording Qualitative Explanations of Groupings by Respondents

XXXXI. Rankings

1. Purpose: Ordering Domain Items in Terms of a Specific Dimension

2. Gets Insider's Perspective

3. Format: Cards vs. Forms, Most to Least, Each item is a Separate Level

4. Ranking Techniques: Standard Ranking, Paired Comparisons, Quick-sorts

XXXXII. Qualitative Data Analysis: Ethnographic Modeling

1. Main Components of Ethnographic Models: Concepts/Terms/Items, Relationship (Semantic) Between Concepts, (causes, stage, part of, kind of, reason for doing, used for)

2. Developing Ethnographic Models:

Why Do We Make Models? A Form of Data Analysis (Especially Summarization), To Identify New Relationships, To Help Plan Further Data Collection, A Tool for Communication, To Assist with Developing Culturally Appropriate Interventions

What Data Do We Use To Create Models? Emic vs. Etic Emphasis, Data Source (Triads, Key Informants, etc.)

Ethnographic Models: Are Tested, Evolve, Grow Increasingly Larger, Should Interact with Others

Steps In Producing A "Box And Arrow" Model

1. Select a topic/cultural domain(s) for the model.

2. Determine which Type of ethnographic model to produce.

3. Extract text "chunks" from qualitative database.

4. Briefly scan through data. Determine Key concepts/terms.

5. Tabulate relationships between Key concepts, counting frequency each relationship is mentioned.

6. Re-order relationships (and associated concepts) in terms of decreasing order.

7. On a clean piece of paper, sketch out the main relationships (most frequently mentioned). Enclose Key concepts inside boxes.

8. Arrows between boxes (concepts) represent relationships. You may wish to draw the nature of the relationship (i.e., causal, progression, etc.) near the arrow. Otherwise, you may choose to indicate the relationship on the border of the model. Relationships can be two-way.

9. You will probably want to indicate the "strength" of the relationship. This can be managed by writing the frequency count (number of informants describing the relationship) next to the arrow. Or else you may want to vary the thickness of the arrow (or make it dashed) to indicate strength.

10. You will probably not wish to indicate relationships mentioned by only 1 or 2 informants in the interest of simplification.

11. You will need to revise this first drawing of the model, to make some relationships more clear to eliminate multiple crossed arrows and so forth.

FORM T1.1
Sample Interview Guide

Opening:
________________________________________________________________

"My name is _____ and I am visiting your village/community/neighborhood to learn about the problems of women in this community, and some of the ways they face and handle these problems. "

Here, it is not only important to learn what problems women face, but what resources they have to deal with those problems, both external (such as clinics) and internal (such as patience, creativity, negotiation).

"Are you able to talk with me now for about 30 minutes?"

If the person appears interested, but does not have time right now, arrange another time. If the respondent appears unwilling, thank her/him kindly and continue on to another informant.

Example of Specific Questions: General Issues Related to Health Problems in the Community
________________________________________________________________

"I would like to learn from you about problems of women in this community. I was wondering if you could share with me some of your insights about this community. "

Some healers may consider some of their knowledge 'proprietary,' thus probing during this interview will need to be done with care and tact.

"What are common problems that women face in this community?" "What do people see as their most important needs?"

"How do you know if somebody is not doing well?" "How does one know a healthy child? A sick child?

A healthy woman? A sick woman?

A healthy man? A sick man?"

"What are ways that women cope with Problem X ?"

"What are problems for women that are most serious in this community?"

Collecting Background Information
________________________________________________________________

Near the end of the interview, it may be appropriate to collect background information on the informant. It is important not to start the interview with these types of questions, as they create a more formal interviewing environment. Answers to many of these questions will probably come up in earlier parts of the interview. It is probably a good idea to double-check some of those earlier responses. For instance,

"Earlier you said that you sell fish in the local market. Is there any other kind of work you do to earn money for your household?" (In order to fill in the space on Occupation)

Closing
________________________________________________________________

"Thank you very much for talking with me today. I know that you are busy and have a lot of work to do - so I am pleased that you can take the time to help me. "

"I would like to come back and talk with you again about other issues as I learn more. Would that be convenient for you? When is the best time for me to come and talk with you?"

Day of the week, time of the day, etc.

It is not possible (or desirable) to cover all of the topics during a single interview. At least one return visit will be necessary to get through the items outlined in the guide below. Probing is the key to good interviews and the interviewer must learn how to probe deeply in key areas. Some tips for good probing are to restate what the informant just said, encourage talking by nodding your head, etc., and try not to ask questions that can be answered by yes or no. In general, do not ask for meaning ask for use. Often the interviewer's goal is to let the respondent "tell the story." It is to be distinguished from badgering, where the interviewer has a specific answer in mind. If they do not understand a question, the interviewer is not using the right terms or phrases, etc. Two examples:

Ineffective Probing:

Interviewer: "What does khamjoori mean?"

Informant:

Interviewer: "Does it mean excessive menstruation?"

Informant: "No"

Effective Probing:

Interviewer: "I was in another village last week and a woman mentioned that she had khamjoori. Is khamjoori in this village also?"

Informant: "Yes, I had it 3 months ago and my daughter had to do all the cooking." Interviewer: "Your daughter had to do all the cooking?"

Informant: "It caused all kinds of problems because she is just learning to cook and my husband could not eat her food. On the second day, I had to lie down near the kitchen to instruct her."

Interviewer: "So with khamjoori, you cannot stand up?"

Informant: "Well usually it is not so bad, but this month the harvest came early and it is much worse. Even the healer's strongest medicine could not cure me..."

Often interviews will last up to an hour. Many researchers find that both the interviewer and the informant become tired if an interview lasts much longer than an hour. Also important data may be lost as the interviewer cannot remember (and cannot write down) all the details in a long interview session.

Contents - Previous - Next