Recursiveness in Qualitative Research: The Story about the Story

by Kenneth Stewart & La Nae Valentine

The Qualitative Report, Volume 1, Number 1, Summer, 1990

The idea that therapy is a conversational art concerning itself with the recording and careful widening of the narrative accounts of clients has begun to receive more attention in both the field of family therapy (Anderson & Goolishian, 1988; Hoffman, 1990; White & Epston 1990) and individual psychology (Sarbin, 1986). In a recent discussion with Bradford Keeney, we discovered we had common interests in the ways ethnography seemed to be informing new directions in family therapy. He described his interests in a narrative approach to therapy-- especially his current work in creating galleries and frames within which to frame discourse in therapy (Keeney, 1991). In discussing how these ideas might be applied to qualitative research, he shared with us an article that he and Monte Bobele had written on discourse in family violence (Keeney & Bobele, 1989). That article offered an analysis of the words professionals used to describe their clients who were involved in situations of domestic violence. These words were clustered into categories with a brief discussion about the meaning of these categories. We found their brief research project to be quite interesting and, when asked by Keeney if we would be interested in conducting a follow up study, we readily agreed.

We drew up a protocol for a project on developing a "lexicon of family violence."# However, instead of just asking the professionals, we thought it would be interesting to ask the "male batters" and the "battered women" themselves. We would then compare their responses with each other and with those of the professionals. The questions we chose which extend the Keeney and Bobele study were:

"What words do you use to describe yourself, your situation and your relationship with your partner ?
What words would you use to describe the process of getting help ?"

For the professionals working with these men and women we asked:

"How would you describe the women/men? Their situation? Their relationships?"
"What words would you use to describe what you do with them in treatment?"
"Are there any words you feel uncomfortable using?"

We began setting up meetings with groups of "battered" women and "battering" men. The women were interviewed at a local shelter for battered women and a rape and abuse crisis counseling center. The men were seen at a halfway house for ex-offenders that conducted court-ordered psycho-educational groups for men who batter women. We met with them in small groups. Data was gathered from the women and men in group settings by setting up an easel with a large newsprint pad while one of the authors would then solicit specific words/phrases that described self, relationship, situation, and the process of getting help. For example, we would ask, "What words would you use to describe yourself?" and would begin writing down exactly their responses. These sessions usually began cautiously but responses from the group quickly picked up steam, with words and phrases flowing rapidly and with kind of cathartic energy.

Recursiveness within the Data Gathering Process

During the course of our interviews, a number of events happened that redirected our inquiry and ultimately led to a richer, more complex final set of data. The first unplanned event during the data gathering process occurred during one of the interviews with the battered women. While describing their parents, one woman angrily said she wanted to describe her mother-in-law as well. The other women quickly agreed. Wanting to be responsive, we allowed for a deviation from our protocol and invited them to share what was on their minds. They described their mothers-in-law as: "controlling, demanding, a witch, stupid with their sons, enablers of their son's behavior and deniers of their sons as abusive and alcoholic." The intensity of their anger seemed to suggest that they were blaming the mothers- in-law for their partners' behavior. Although they described their fathers-in-law as: "alcoholic, abusive, flirty, a womanizer, judgmental, overpowering, a wuss, and ball-less," they did not show the same intensity of anger as they did toward their mothers-in-law. These two sets of descriptions revealed a cynical and bleak picture of a context of abuse, evasions--misplacement of responsibility, and imbalances--abuses of power across at least two generations.

After gathering these spontaneous outpourings, it was decided that before gathering words from the men, we would again deviate from our original protocol and ask them to give us the words the would use to describe their own mothers and fathers in order to discover how these words compared with the women's descriptions of their in-laws. Not surprisingly, the men described their own mothers in fairly positive terms: "loving, caring, giving, supportive, beautiful, understanding, sympathetic, etc." They described their fathers in both positive as well as negative terms: "drunk, abusive, angry, hateful, ignorant, selfish, (as well as) loving and caring, a good role model, changed in his old age, etc."). We concluded that these two diverse sets of descriptions dramatically pointed out the differences in perspective, depending on whether one creates distinctions from a cared-for perspective or from an abused perspective. That is, everyone understood by the battered women as contributing to her own abuse--her partner and his father and mother, received only negative descriptions.

Counting in the spirit of recursively organized research, at the end of a data gathering session with the men, they were asked if there was anything else they wanted to describe. They said they wanted to describe society. They offered words representing imbalance such as: "little regard for human life," "greedy," "unequal," and "divided into normal and not normal." Other words described blindness to injustice such as "blind to what's really happening" and "short sighted." To describe how they thought society treated them they used: "they got your number," "gestapo," "laws," "limitations on individuals," and "railroad."

This diversion from the protocol revealed a picture somewhat similar to that of the women when asked to describe their relationships with their partners: blind, unaware, unequal, powerless, discounted, and used. In other words, the men believed that society viewed them as inadequate, while they in turn felt degraded and controlled.

While the women felt victimized by the larger family systems, the men felt victimized by the larger social system. Interestingly, no one commented on this isomorphism. Without these diversions, important directions for the clinical application of these findings might not have been revealed:

1. Both may feel victimized, degraded and discounted by larger systems--he by the larger social system and she by him and the extended family system. However, with his greater physical strength, he may often displace his anger onto her in an abusive manner, but where does she go with her frustration: To the children? To herself?
2. There are important differences in these two sets of descriptions. When victimization happens in intimate settings a person ends up with cuts, bruises, broken bones and a pervasive feeling of terror and mistrust. When victimization happens in the less intimate setting of work or the social system, the person ends up with feelings of inadequacy, incompetence, and frustration. While both of them may have come from families with situations of abuse, the consequences to her in the present context are quite personal, physical and intimate. The consequences to him, on the other hand, may be personal, but are more social and economic.
3. At the end of the interviews with the battered women, one woman said, "Someone should write a book about what we've just told you. I think it would be helpful to others in similar situations." In response, the second author asked: "If you were to go ahead and write such a book, what advice would you give to women in violent situations similar to your own?" They said things like, "You can't tell them what to do or to get out, they won't listen; instead, be supportive, tell them they can call anytime to talk. I'd let them know I understand, that I've been there...[and] this is how I did it..." "I would encourage them to get a job and to think about the future."

Reflections on Recursiveness and "Significance"

A number of observations and reflections became apparent at the close of this process. In analyzing the data, we considered a number of approaches, including domain analysis and end linkage analysis. We consulted with an anthropologist/linguist from the Anthropology Department on campus about various approaches to the analysis. In addition to suggesting a type of domain analysis, he pointed out that the words might represent the complementary ideals of the person. For example, if some was called "insecure" it would imply the complement- -someone secure or strong, since the words would reflect the values of the person uttering them. We finally decided to cluster the words into categories and examine those categories for recursive patterns of symmetry, complementarity, and intergenerational dynamics. We were so impressed with the power and intensity of the words that we wanted to ensure that any analysis conducted would not remove us too far from the experience and emotion that the words seemed to evoke.

At the beginning of writing this brief report, the first author told a former graduate student from the Child Development and Family Studies Department (who also happens to be the Program Director of the local YWCA shelter for battered women and who assisted in the project), that that he was writing an article on "recursively organized research," she cynically remarked, "Well, why wouldn't you ask people what's important to them and let them talk about it?" It was as if she was almost sarcastically saying, "Gee, its nice research can be responsive to people." She then went on to add: "I can understand why people wouldn't do that kind of research because they wouldn't have control over it, and couldn't predict the outcome. It takes the researcher out of his more important role... and is more subject directed."

In reflecting on the clinical utility of this project, we were again reminded of Anderson and Goolishian's (1988) premises that psychotherapy is essentially "linguistic event" where "new descriptions arise, new meanings are generated, and therefore, new social organization will occur around different narratives" (p. 384). Interventions can even take place at the level of the individual descriptors people employ to construct meaning and inform their actions. Beginning with just the descriptive words could be seen as fundamental to this re-languaging art. The reauthoring wounded narratives begins with the careful introduction of alternatives to rigidly-held constructions in order that words and frames can be offered that not only open dialogical space between people but also begin the healing.


Anderson, H., & Goolishian, H. A. (1988). Human systems as linguistic systems: Preliminary and evolving ideas about the implications for clinical theory. Family Process, 27, 371-393.

Hoffman, L. (1990). Constructing realities: An art of lenses. Family Process, 29, 1-12.

Keeney, B. P., & Bobele, M. (1989). A brief note on family violence. Australian and New Zealand Journal of Family Therapy, 10(2), 93-95.

Keeney, B. P. (1991). Improvisational therapy. New York: Guilford.

Sarbin, T. (1986). (Ed.) Narrative psychology: The storied nature of human conduct. New York: Praeger.

White, M., & Epston, D. (1990). Narrative means to therapeutic ends. New York: W. W. Norton.

Author Note

# This study will be published in 1991 as:

Stewart, K., & Valentine, L. (1991). Metaphors in domestic violence. In B. K. Keeney, W. L. Madsen, & B. F. Nolan (Eds.), The Systemic Therapist (Vol. 2) (pp. 60-77). St. Paul: The Systemic Therapy Press.

Kenneth Stewart, Ph.D., is Assistant Professor and Director of the Family Therapy Program, Department of Child Development and Family Science, North Dakota State University, Fargo, North Dakota, USA.

La Nae Valentine, Ph.D., is Assistant Professor and Acting Director of the Family Therapy Program, Department of Human Development and Family Studies, Colorado State University, Fort Collins, Colorado, USA.

Return to the Table of Contents