Qualitative Data

An Introduction to Coding and Analysis

Carl Auerbach, Louise B Silverstein, Louise B. Silverstein

• Publisher: NYU Press
• Year: 2003
• Print ISBN: 9780814706954
• eBook ISBN: 9780814707418
• Topic: Psychology
• Subtopic: Research & Methodology in Psychology
• Index: Psychology

A necessary guide through the qualitative research process

Qualitative Data is meant for the novice researcher who needs guidance on what specifically to do when faced with a sea of information. It takes readers through the qualitative research process, beginning with an examination of the basic philosophy of qualitative research, and ending with planning and carrying out a qualitative research study. It provides an explicit, step-by-step procedure that will take the researcher from the raw text of interview data through data analysis and theory construction to the creation of a publishable work.

The volume provides actual examples based on the authors' own work, including two published pieces in the appendix, so that readers can follow examples for each step of the process, from the project's inception to its finished product. The volume also includes an appendix explaining how to implement these data analysis procedures using NVIVO, a qualitative data analysis program.

GETTING INTO QUALITATIVE RESEARCH

Part I

1

Introducing Qualitative Hypothesis-Generating Research
The Yeshiva University Fatherhood Project

RESEARCH BEGINS WITH CURIOSITY about the world. We assume that you are reading this book because you find a particular phenomenon interesting and want to understand it better. For example, you may have questions about trauma, or fathering, or divorce, or immigration, to list some of the topics that our research team has studied. This book will teach you how to use a very powerful research method, qualitative research, to answer your questions and learn more about your topic.

The field of qualitative research is quite diverse. Some methodologies included in this approach are: participant observation, fieldwork, ethnography, unstructured interviews, life histories, textual analysis, discourse analysis, and critical cultural history, and this list is by no means exhaustive. For our purpose, qualitative research can be defined as follows:

Qualitative research is research that involves analyzing and interpreting texts and interviews in order to discover meaningful patterns descriptive of a particular phenomenon.

The qualitative approach to research design leads to studies that are quite different from those designed using the more traditional approach. The traditional approach, often referred to as quantitative research, leads to hypothesis-testing research, whereas the qualitative approach leads to hypothesis-generating research. This chapter will describe the difference between the two approaches, and spell out the conditions under which hypothesis-generating research is an appropriate research strategy. We will illustrate the discussion by describing our own research program, the Yeshiva University Fatherhood Project.

Hypothesis-Generating Research as an Alternative to Hypothesis-Testing Research

Perhaps the best way to explain the difference between quantitative hypothesis-testing research and qualitative hypothesis-generating research is to describe how we became qualitative researchers. This occurred as a result of our research program investigating fathers and fathering, the Yeshiva University Fatherhood Project.

We undertook the project for both personal and professional reasons. Beginning with the personal reasons, we ourselves had intense positive and negative feelings toward our own fathers. In addition, we were both actively involved in raising our children, and were struggling to create families where fathers and mothers had an equal role in child rearing.

Our professional reasons stemmed from the fact that traditional theories of child development had neglected the role of the father. They assumed instead that the mother was the most important figure in the child’s life, and that the father’s role was simply to support the mother. When we began our work, the field had begun to challenge this assumption (e.g., Cath, Gurwitt, & Gunsberg, 1989; Lamb, 1987) and we wanted to contribute to this developing body of research and theory.

We were committed to studying fatherhood from a multicultural diversity perspective. Much of the research on fathering studied only traditional Euro-American families, which we found unnecessarily limiting, both in terms of developing theory and in terms of developing clinically useful knowledge. Thus, we decided to study fathers drawn from the entire spectrum of cultural, ethnic, and sexual orientations. Ultimately, our research included such diverse subcultures as Haitian American fathers, Promise Keeper fathers, divorced fathers, gay fathers, Latino fathers, young unmarried black fathers, stepfathers, and white middle-class fathers in dual-career couples. To date we have completed data collection on over 400 men.

As we thought about designing research to investigate fatherhood from a multicultural diversity perspective, we realized that traditional quantitative hypothesis-testing research wasn’t suitable for our purposes. To explain why, we must describe traditional hypothesis-testing research in a bit more detail. Some of you are already familiar with this material from research design courses, but for those of you who are not, we briefly review it. Hypothesis-testing research may be defined as follows:

Hypothesis testing research investigates a phenomenon in terms of a relationship between an independent and dependent variable, both of which are measurable numerically. This relationship is called a hypothesis. The aim of the research is to test whether the hypothesized relationship is actually true, using statistical methods.

Here is a simplified example of how a hypothesis-testing researcher might design a study of fatherhood. She would begin by choosing a dependent variable to define the phenomenon of fatherhood, such as a father’s affection for his child. To study this variable in research she would have to make it measurable, so she might have the fathers rate their affection for their child on a scale from 1 (the lowest) to 7 (the highest).

Then she would decide on an independent variable, by which is meant a variable likely to have an effect on the dependent variable of affection. She might choose as an independent variable the father’s contact with his child, as measured by the number of minutes the father spends in the same room with his child.

Finally, she would state a hypothesis about the relationship between the independent and the dependent variable. She might hypothesize that the more contact a father has with his child, the greater his affection for that child. This hypothesis could be tested experimentally by seeing whether there is a statistically significant correlation between the independent and dependent variables.

This example, although simplified, illustrates the problems we had in using the hypothesis-testing approach to study fatherhood from a multi­cultural diversity perspective. There were two basic difficulties. First, we didn’t know enough to state meaningful hypotheses, particularly for cultures different from our own. The hypothesis above is plausible for our own white middle-class culture, but it is less likely to be true in other cultures. For example, a middle-class father may be able to make enough money to support his family with one job. For him, choosing to spend time in his child’s presence, rather than in leisure-time pursuits that would exclude his children, may be an accurate reflection of his affection for his child. However, a working-class Latino immigrant father may have to work two full-time jobs to earn enough money to provide tutoring for his children so that they can improve their high school grades. Thus, affection for his children may cause him to spend less, rather than more, time in contact with them.

Moreover, not only did we not know enough to state meaningful hypotheses; we didn’t even know enough to select meaningful independent and dependent variables. For example, when we studied Haitian American fathers, we discovered that their religious belief was an important variable in understanding how they defined good fathering. We would not have expected this based on our experience with our own secular middle-class culture.

The second problem we had with the hypothesis-testing approach is that for clinical and theoretical reasons we were interested in understanding the subjective experience of fathers, and because variables must be defined numerically in hypothesis-testing research, they cannot reflect subjective experience. Even if the study yielded significant results, we would know very little about the fathers’ subjective experience, that is, what they actually felt about their children. In order to understand something meaningful about his affection for his child, we wanted the following and other questions answered.

What does a father’s affection for his child feel like?
What does it mean for a father to be in the presence of his child?
Does he remember times that he was with his own father or mother?
Does he feel nervous being left in charge of an infant, without his wife or another woman present?

In order to address the questions that we were interested in, we searched for a research method that would not require us to state a hy­pothesis prior to beginning our investigation, and that also would allow us to study subjective experience directly. This led us to do hypothesis-generating research using the grounded theory method.

Hypothesis-Generating Research Using the Grounded Theory Method

The grounded theory method allows the researcher to begin a research study without having to test a hypothesis. Instead, it allows her to develop hypotheses by listening to what the research participants say. Because the method involves developing hypotheses after the data are collected, it is called hypothesis-generating research rather than hypothesis-testing research. The grounded theory method uses two basic principles: (1) questioning rather than measuring, and (2) generating hypotheses using theoretical coding.

Questioning Rather Than Measuring

The grounded theory method allows the researcher to acknowledge that she may not know enough to formulate meaningful hypotheses. It uses the research participants as a source of knowledge. After all, they are experts on the phenomenon being studied because they are experiencing it directly. This methodology questions the research participants about their subjective experience and generates hypotheses from their answers. For example, our hypotheses about Haitian fatherhood were developed from what the Haitian fathers said in their interviews with us.

Generating Hypotheses Using Theoretical Coding

The grounded theory method uses a data analysis procedure called theoretical coding to develop hypotheses based on what the research participants say. Grounded theory derives its name from the fact that theoretical coding allows you to ground your hypotheses in what your research participants say.

Our discussion so far has covered the application of grounded theory to interview data, and, as you will see later, our illustrative data will be from group interviews, that is, focus groups. You should be aware, however, that the data for qualitative research can also include observed behavior, participant observation, media accounts, cultural artifacts, among others. Thus, the techniques and illustrations we present here cover an important part of the field of qualitative research, but by no means all of it.

With this qualification in mind, qualitative hypothesis-generating research may be defined as follows.

Qualitative hypothesis-generating research involves collecting interview data from research participants concerning a phenomenon of interest, and then using what they say in order to develop hypotheses. It uses the two principles of (1) questioning rather than measuring and (2) generating hypotheses using theoretical coding.

Describing a systematic method for doing qualitative hypothesis-generating research is the subject of the rest of this book. We will use examples from our own research to illustrate our methodology. We now turn to a broad overview of our research project in order to provide a context within which to understand each example.

The Yeshiva University Fatherhood Project is a large-scale qualitative research study whose researchers have interviewed more than 400 men from many different U.S. subcultures. As we noted previously, these subcultures include Haitian American fathers, Promise Keeper fathers, divorced fathers, gay fathers, Latino fathers, young unmarried black fathers, stepfathers, and white middle-class fathers in dual-career couples.

Each subculture is studied using a sample of approximately 20 fathers who are interviewed in small groups called focus groups, each consisting of 4 to 6 participants. Thus, each study includes 4 or 5 focus groups. The participants are recruited as a convenience sample using a snowball sampling technique. These terms will be defined more precisely later, but basically they mean first interviewing people who are accessible, and then interviewing people known to the original participants. The group interviewer is a graduate student, and is usually but not always a member of the subculture being studied. For example, a Haitian American graduate student interviewed the Haitian American fathers. Three native-born Latinos interviewed 3 of the Latino samples, but the fourth was interviewed by a white Euro-American bilingual school psychologist. One of the gay fathers groups was interviewed by a gay man, but the second group of gay fathers was interviewed by a heterosexual woman.

The focus group interviews are audiotaped or videotaped, the tapes are transcribed, and the transcript is analyzed by a group of four researchers: ourselves, the focus group interviewer, and another graduate student involved with the project. Two consultants from the subculture of the group being studied are also asked to read through the transcripts and comment on them.

After the data are analyzed, a brief synopsis of the findings is brought back to the research participants for discussion. This discussion corrects, broadens, and deepens the researchers’ understanding of the participants’ subjective experience. The data are then organized by the graduate student and written up as that student’s Psy.D. project. At a later date, the findings are reorganized for publication, and a consultant from the subculture is again involved before the final draft of the paper is submitted for publication.

In the chapters that follow, examples from the Haitian American, Promise Keeper, and gay fathers example will be used to illustrate our data analysis procedures. These samples have all been analyzed separately. The articles describing the Haitian American fathers and the Promise Keeper fathers are included as appendixes. The gay fathers have been discussed in a recent publication on gender role strain in U.S. fathers (Silverstein, Auerbach, & Levant, 2002).

PLANNING YOUR FIRST RESEARCH STUDY

Part II

2

Designing Hypothesis-Generating Research
The Haitian Fathers Study

WE WILL NOW CONSIDER how the philosophical differences between hypothesis-generating and hypothesis-testing research translate into research practice. We will contrast a standard hypothesis-testing design with the design of our Haitian Fathers Study. This study was part of our larger research program, the Yeshiva University Fatherhood Project, which we described in chapter 1. The article reporting the results of the Haitian Fathers Study is included as Appendix B (Auerbach, Silverstein, & Zizi, 1997), which you should read at this point.

Hypothesis-Testing Research Design

Research methods courses teach a standard procedure for planning a research study, which goes as follows.

Step 1. Conduct a Literature Review and Identify a Research Problem

Review the research literature to identify an area that we need to learn more about. Pose a small, manageable question about this area. Your research problem is to answer this question.

Step 2. Develop a Research Hypothesis

Propose an answer to your research problem by formulating a hypothesis derived from your review of the literature about how an independent variable influences a dependent variable. This is called your research hypothesis.

Step 3. Operationalize the Variables

Choose numerical scales, tests, or other instruments by which to measure your independent and dependent variables. This is called operationalizing the variables. It allows you to test your research hypothesis statistically.

Step 4. Establish a Random Sampling Technique

Decide on the population you want to study and the procedure you will use to obtain a random sample from this population. Your goal is to use a random sampling procedure so that you can generalize your results.

Step 5. Determine Sample Size

Decide on the number of subjects you will include in your research sample. Your goal is to have a large enough sample so that you...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contents
• Preface
• Acknowledgments
• Part I GETTING INTO QUALITATIVE RESEARCH
• Part II PLANNING YOUR FIRST RESEARCH STUDY
• Part III ANALYZING YOUR FIRST RESEARCH STUDY
• Part IV DESIGNING AND ANALYZING YOUR NEXT RESEARCH STUDY
• Part V FINAL THOUGHTS
• Appendix A Simplifying the Bookkeeping with Qualitative Data Analysis Programs
• Appendix B The Haitian American Fathers Study
• Appendix C The Promise Keepers Study
• References
• Index
• About the Author