(p.163) Appendix: Study Methods
(p.163) Appendix: Study Methods
The study reported here began several years ago. The School of Social Welfare at the University of California at Berkeley was granted funding from the Smith Richardson Foundation and the Walter S. Johnson Foundation to examine the social implications of welfare reform for families. Neil Gilbert, Marcia Meyers, and I looked at the experiences of more than four hundred women and and their children in the GAIN program in California and also at child-care patterns and changes in family life.1
Our study consisted of three structured telephone interviews with AFDC mothers. We spoke with them before they began the GAIN program, three months into the program, and a full year after our original interview. Each interview lasted between forty-five minutes and an hour. We probed for information about the women’s experience in the GAIN program, and in the process, they told us about their poverty and their experience with AFDC. One of the challenges in conducting this quantitative study was in finding patterns of similar responses, because these women’s experiences with poverty were very different.
The later study undertaken for this book analyzed these women’s different experiences of living on AFDC. Unlike the earlier study, this project relied solely on qualitative data. I used secondary data from large-scale national studies to clarify the aggregate experience of welfare.
Unlike quantitative studies whose sample of participants is statistically representative of the population at large, the samples for qualitative studies are selected somewhat differently. The sample of participants that I selected for this book was fairly limited in scope and size. I looked (p.164) for AFDC recipients that were typical of the various types of women on welfare. Some of the characteristics that I looked for were differences in education, race, marital status, and work history.2
Other factors may be, of course, relevant to selecting the participants in a qualitative study. Some researchers look for respondents who are somewhat “trustworthy, observant, reflective, articulate, and good storytellers.”3 Bernard observed that the selection of a sample is based on “luck, intuition, and hard work by both parties to achieve a working relationship based on trust.”4 The respondents in my study contained some or all of these qualities.
Four of the five women selected for this project had participated in the earlier GAIN study. In accordance with the design of the earlier study, I excluded short-term AFDC recipients. (Women who use AFDC for a short period of time are not likely to enter the GAIN program.) Therefore, Ana had not participated in the original study. In order to find an appropriate respondent to illustrate the experience of the short-term user, I interviewed fifty-four women by telephone or in person, and I asked the staff at local AFDC offices to identify short-term users. In addition, I placed ads for study candidates in newspapers. I interviewed eleven women in depth at least twice before I finally selected Ana.
Besides the other four study participants, additional respondents initially participated in the study. But because of family crises, some of the women dropped out of the study for a time and then reentered later. Three women who participated at the beginning of the study had to drop out entirely; one, for example, moved several hundred miles away when her son was seriously injured in an accident, and she understandably found time and distance to be insurmountable obstacles to continuing to participate.
All the names and identifying information of the women have been changed, but all of the women are real; none is a composite. Women on welfare live in constant fear of being “found out” by welfare officials, and therefore I felt obligated to hide their identities.
I conducted this study using qualitative field strategies including in-depth interviews, informal participant observation, and key informants.5 By means of these methods, a researcher can obtain direct knowledge of experiences, events, and the emotions they elicit in real-life settings.6
(p.165) These women welcomed me into their homes and allowed me to sit back and unobtrusively observe their family life. At other times I actively participated in family activities. In addition to these informal gatherings, some of our meetings were unstructured interviews, usually focusing on a specific theme.7 Central to the interviewing process was the development of an oral history for each woman as she reexamined and told her life story. Wiener and Rosenwald noted that individual life stories often take place as a series of episodes, some of which the respondents remember more clearly than others.8 Therefore, part of my analysis of the notes I took when eliciting these life stories was find their meaning to their teller.9 Accordingly, I hope that I have been true to these women’s understanding of their lives as I tried to capture their perspective in each chapter.
Each meeting with the women and their families lasted between two and six hours. I collected data for between three and eighteen months, intending to approximate an informal conversation between friends whenever possible.10 In some cases, our relationship developed into a friendship, making it difficult for me to make a completely unbiased analysis.11 Other factors also were relevant to my data. My own personal characteristics, such as my race, class, and age, were influential at first in gaining my respondents’ trust. In some cases, we needed to talk through their perceptions of who I was or, more important, the institutions I represented.12 For example, some of Cora’s family members initially responded to me with some skepticism, if not hostility, as a “researcher.” Their responses are not unusual.13 But one way to overcome fear or hostility is to become a familiar person in the family landscape. When I brought birthday gifts for the children or helped run errands with Cora, her children began to see me as a friendly visitor and eventually welcomed me into their home.
Analyzing the Data
I audiotaped all my meetings with the women and their families, and at the end of each meeting I also took notes about any tone or manner that might not be discernible on the tape. I then had each tape fully transcribed.
My data analysis included identifying and labeling those themes that were consistently repeated in each meeting.14 I also found patterns in the women’s oral histories that related back to the initial themes, and I was able to tie some of my initial analyses to my interviews.15 Those issues and themes that needed further exploration I often discussed (p.166) with the respondents at a later meeting, either informally or in an unstructured interview. Before beginning my final analysis and writing, I had to put some distance between myself and the families, in order to regain my objectivity.16 I communicated the data by means of two methods: narrations and augmentations.17 The narrations are those portions of each chapter provided directly by the respondents, and the augmentations are my descriptions of or notations regarding striking events or experiences. Throughout the book I attempted to tie the women’s and children’s individual life experiences to aggregate experiences of welfare use, gleaned from large-scale national studies.
All the women discussed in this book reside in California, a state with relatively high welfare benefits, an extremely high cost of living, and a recession that has lasted far longer than have those in other states. These women’s experiences should not be seen as “average.” Indeed, a sample of five women cannot be generalized to the larger population; they and their experiences are offered only as a complement to existing quantitative studies and are meant as an illustration of the effects on women and children of living in poverty.
(1) . N. Gilbert, J. D. Berrick, and M. Meyers, GAIN Family Life and Child Care Study (Berkeley, CA: Family Welfare Research Group, 1992).
(2) . Regarding the validity of ethnographic samples, Mead writes, “The validity of the sample depends not so much upon the number of cases as upon the proper specification of the informant, so that he or she can be accurately placed, in terms of a very large number of variables—age, sex, order of birth, family background, life-experience, temperamental tendencies, political and religious position, etc…. Within this extensive degree of specification, each informant is studied as a perfect example, an organic representation of his complete cultural experience.” M. Mead, quoted in J. C. Johnson, Selecting Ethnographic Informants (Newbury Park, CA: Sage, 1990), p. 22.
(3) . D. L. Jorgensen, Participant Observation: A Methodology for Human Studies (Newbury Park, CA: Sage, 1989), p. 30.
(4) . H. R. Bernard, Research Methods in Cultural Anthropology (Newbury Park, CA: Sage, 1988), p. 177.
(5) . M. Burawoy, “The Extended Case Method,” in M. Burawoy et al., eds., Ethnography Unbound: Power and Resistance in the Modern Metropolis (Berkeley and Los Angeles: University of California Press, 1991), pp. 271–90; R. F. Ellen, ed., Ethnographic Research: A Guide to General Conduct (New York: Academic Press, 1984); O. Werner and G. M. Schoepfle, Systematic Fieldwork (Newbury Park, CA: Sage, 1987).
(6) . M. Bulmer, “The Value of Qualitative Methods,” in M. Bulmer, with K. G. Banting, S. S. Blume, M. Carley, and C. Weiss, eds., Social Science and Social Policy (Boston: Allen & Unwin, 1986), pp. 80–204.
(7) . D. M. Fetterman, Ethnography: Step by Step (Newbury Park, CA: Sage, 1989).
(8) . W. J. Wiener and G. C. Rosenwald, “A Moment’s Monument,” in R. Josselson and A. Lieblich, eds., The Narrative Study of Lives (Newbury Park, CA: Sage, 1993), pp. 30–58.
(9) . Jorgensen, Participant Observation.
(10) . J. P. Spradley, The Ethnographic Interview (New York: Holt, Rinehart & Winston, 1979).
(11) . S. M. Miller, “The Participant Observer and ‘Over-rapport’,” in G. J. McCall and J. L. Simmons, eds., Issues in Participant Observation (Reading, MA: Addison-Wesley, 1969), pp. 87–89.
(12) . E. Sawyer, “Methodological Problems in Studying So-Called ‘Deviant’ Communities,” in J. Ladner, ed., The Death of White Sociology (New York: Random House, 1973), pp. 361–79; Spradley, The Ethnographic Interview; R. H. Wax, “Gender and Age in Fieldwork and Fieldwork Education: No Good Thing Is Done by Any Man Alone,” Social Problems 26 (1979): 509–22.
(13) . Jorgensen, Participant Observation.
(14) . G. Rosenthal, “Reconstruction of Life Stories,” in R. Josellson and A. Lieblich, eds., The Narrative Study of Lives (Newbury Park, CA: Sage, 1989), pp. 59–91.
(15) . Jorgensen, Participant Observation.
(17) . Rosenthal, “Reconstruction of Life Stories.”