Voices of Strength and Resistance: A Contextual and Longitudinal Analysis of Women's Responses to Battering.

A Contextual and Longitudinal Analysis of Women's Responses to Battering

A recent national probability study of women estimates that at least 4.4 million women are physically assaulted in the United States each year (Plichta, 1996). There is substantial evidence demonstrating that battered women are seriously at risk for homicide as well as a wide range of physical and mental health problems (Campbell, 1995; Eby, Campbell, Sullivan, & Davidson, 1995; McCauley et al., 1996; Plichta, 1996; Ratner, 1993). In spite of this knowledge, there is a paucity of information about women's responses to battering over time that integrates physical, emotional, social, and behavioral reactions within complex familial and environmental contexts. A study was undertaken to examine women's responses at three points in time over a 3 1/2- to 4-year period, using both in-depth interviews and standardized instruments. This article reports the component of the study combining qualitative and quantitative data analysis to illustrate the complexity of women's decision making in terms of the relationship status and continuing or cessation of abuse.

Literature Review

The traditional picture of battered women as passive victims has been counteracted at least partially in recent years by a variety of studies using both qualitative and quantitative data. Gondolf, Fisher, and McFerron (1988) and Okun (1986) both used large-record reviews from wife-abuse shelters to demonstrate survival strategies of battered women in the face of serious terrorism. Hoff (1990), in an ethnographic analysis, documented the importance of women's social support systems--especially family, friends, and shelters--in helping them to survive and escape the abuse. Okun (1986) and Campbell, Miller, Cardwell, and Belknap (1994) found that the majority of battered women do leave an abusive relationship but generally leave several times before they leave permanently. In the recent Jacobson, Gottman, Gorner, Berns, and Wu Shortt (1996) study, 42% left or achieved nonviolence. It also has been demonstrated that women's risk for abuse continues after they leave (Kurz, 1996; Wilson, Johnson, & Daly, 1995).

Several investigators have used qualitative approaches to examine these processes. Merritt-Gray and Wuest (1995) found women actively counteracted abuse from the first appearance of violence. Women first relinquished parts of themselves, minimized the abuse, and fortified their defenses as the first part of their process of breaking free. Landenburger (1989) and Mills (1985) described abused women as losing a sense of self or identity, either early in the relationship or during what Landenburger calls an "entrapment" phase. Her study showed subsequent stages of disengaging and recovery. Lempert (1996) developed a central concept of agency to describe the strategies and processes women used to "halt, change or cope with their partner's violence." Her study redefined seemingly passive actions as active survival strategies and seemingly self-erasures as face-saving strategies to maintain invisibility. These studies have been instrumental in demonstrating the interactive and complex processes that characterize abusive relationships and women's responses to them. All of these qualitative studies, however, were cross sectional accounts and included primarily well-educated White women, most of whom had left the abusive relationships and retrospectively were recounting the events and their responses. The current study addresses these gaps by interviewing primarily African American women over three points in time.

METHOD

The study was conducted over 3 years, with interviews conducted every 8 to 12 months. A total of 164 women were recruited at Time 1, and 96 women returned for the third interview. Although 38% of the overall sample was lost by Time 3 due to safety issues and transience, there were no significant differences in demographic or major predictor and outcome variables (abuse, stress, self-care agency, self-esteem, depression, health symptoms) between those who returned and those who did not. The data for this analysis are from a subset of 32 of the women randomly selected from the 96 battered women that were followed at all three points in time.

Women who "had serious problems in an intimate relationship with a man" were recruited with newspaper advertisements and bulletin board postings in places frequented by women, such as child care centers, community centers, laundromats, and university women's restrooms in the metropolitan area of a large Midwestern city. Because prior research has shown that committed relationships are more likely to be characterized by violence (Fagan & Browne, 1994), only women who had been in a relationship for at least 1 year were included.

Women were screened for battering using the Conflict Tactics Scale, developed by Straus and Gelles (1990). The scale assessed conflict tactics directed toward the woman only and was modified by adding the question, "Has your partner ever forced you into sex that you did not wish to participate in?" This question originally was used by Russell (1982) in her study of marital rape and subsequently used by Campbell (1989) to successfully determine forced sex in battering relationships. Battering was defined in this study as repeated physical and/or sexual assault within a context of coercive control (Campbell & Humphreys, 1993). A woman was categorized as battered if there were more than one instance of minor violence or at least one major violence tactic or forced sex perpetrated by her partner against her during the past year. The coercive control aspects of the relationship were determined by responses on the Index of Spouse Abuse (ISA) (Hudson & McIntosh, 1981). Coercive control includes emotional abuse, threats, and intimidation, as well as financial control and social isolation (Pence & Paymer, 1993). Identifying self as battered was not used as an inclusion criteria, as many women subjected to violence in intimate relationships do not put those labels on themselves (Schechter & Jones, 1992).

The methodology was enacted as feminist action research as described across disciplines by such scholars as Reinharz (1992), Small (1995), Lather (1991), and Henderson (1995). The term can be said to encompass a group of evolving methodologies that are striving to create change as the result of the research (action research) and/or have participants both be changed and create change as part of the research process (participatory research). The feminist specificity of this particular strand of action research adds that it is research for women, is reflexive, represents human diversity, strives to equalize power relationships both within research teams and between researchers and participants, critiques prior scholarship (especially for andocentric and ethnocentric bias), uses both qualitative and quantitative data, and portrays women's strengths (Bunting & Campbell, 1994).

Data were collected with in-depth structured, semistructured, and open-ended interviews that used a dialogic approach (give-and-take discussion rather than traditional question and answer). Each interview lasted approximately 2 hours. Interview topics included relationship history, relationship problems and problem-solving strategies, attributions, and cultural norms about abuse. They also included safety planning and referrals as requested by the participants. Because of safety concerns, many women were only contacted through work or a family member or friends and were never called at home.

Instruments

Severity of physical and emotional abuse were assessed using the modified version of the ISA (Hudson & McIntosh, 1981) and interview questions about abuse experience by the women. The ISA (Hudson & McIntosh, 1981) is a …