Violence Against Women http://vaw.sagepub.com/
Effects of Work on Hitting and Hurting Violence Against Women 2003 9: 1213 DOI: 10.1177/1077801203255848 The online version of this article can be found at: http://vaw.sagepub.com/content/9/10/1213
Published by: http://www.sagepublications.com
Additional services and information for Violence Against Women can be found at: Email Alerts: http://vaw.sagepub.com/cgi/alerts Subscriptions: http://vaw.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://vaw.sagepub.com/content/9/10/1213.refs.html
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
VIOLENCE ARTICLE AGAINST WOMEN / October Brush / 2003 EFFECTS OF 10.1177/1077801203255848 WORK ON VIOLENCE
Effects of Work on Hitting and Hurting LISA D. BRUSH University of Pittsburgh
Researchers have documented the ways battering can impede work and impoverish women. This project looked at abuse, work, and poverty the other way around, asking, Does women’s working affect abuse or post-traumatic stress disorder (PTSD)? This article reports 162 welfare recipients’ subjective assessments of the effects their working had on physical violence, work-related abuse, and PTSD symptoms. For many respondents, work aggravated abuse, in which case their average wages and weeks worked were lower. Many respondents also reported that working ameliorated their symptoms. These findings explain contradictory results of previous studies and form the basis for research, policy, and practice recommendations. Keywords: battering; poverty; welfare; work
To the extent that battering—that is, physical, emotional, financial, and sexual abuse through which men exert power and control over their current or former wives or girlfriends—interferes with work, it may also trap women in poverty (Davis, 1999; Raphael, 1999, 2000). Moreover, when welfare reforms focus narrowly on employment, battering can also keep women from complying AUTHOR’S NOTE: This project was supported by Grant No. 2000-WT-VX-0009 awarded by the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. Points of view in this document are those of the author and do not necessarily represent the official position or policies of the U.S. Department of Justice. The research received expedited review and approval from the University of Pittsburgh Institutional Review Board (No. 001097). The Pennsylvania Department of Public Welfare’s Bureau of Program Evaluation authorized site access and release of findings. This project would have been impossible without Bob Reynolds and Paula Hustwit. S. Laurel Weldon made a key methodological suggestion early on. Danielle Ficco assisted with data collection and analysis. Amelia Haviland improved the analysis and argument. Karen Drescher Flora, Jody Raphael, and two anonymous reviewers offered valuable editorial suggestions. Individuals and institutions providing funds, approval, or support do not necessarily agree with the findings or interpretations presented herein, which are solely those of the author. Thanks to the participants who made this research possible. VIOLENCE AGAINST WOMEN, Vol. 9 No. 10, October 2003 DOI: 10.1177/1077801203255848 © 2003 Sage Publications
1213-1230
1213
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1214
VIOLENCE AGAINST WOMEN / October 2003
with job search and work requirements (Brush, 2000). Abuse may lead to sanctions or otherwise deepen women’s poverty. When women must rely on currently or formerly abusive partners for work supports, such as child care or transportation, welfare reform can deepen what Scott, London, and Myers (2002) called “dangerous dependencies.” Research on the connections between poverty and violence against women started with anecdotal descriptions (Raphael, 1996, 2000). Advocates and researchers quickly sought to estimate the prevalence of battering among poor women in general and welfare recipients in particular (Olsen & Pavetti, 1996). “Barrier” studies have found a wide range of prevalence rates for physical abuse that might interfere with women’s labor force participation and earnings (see, e.g., Olsen & Pavetti, 1996; Tolman & Raphael, 2000). Unfortunately, even those barrier studies that include items on physical violence by an intimate partner seldom include items measuring work-related abuse, control, and sabotage. The findings reported here come from research designed to increase what we know about poverty and violence against women by including measures of the ways men interfere directly with women’s employment. Investigations of the relationships among poverty, work, welfare, and battering generally use longitudinal data to assess the degree to which battering obstructs women’s employment and transition from welfare to work. Although anecdotal evidence suggests it might be important (Raphael, 1999), no published research has assessed the extent to which employment precipitates or aggravates battering or battering’s consequences. One way to measure the effects of work on hitting and hurting is to ask women whether their working changed the onset, frequency, or severity of abuse. My research contributes to knowledge about poverty and abuse by analyzing the subjective assessments welfare recipients (N = 162) gave of the effects their working had on abuse and the symptoms of traumatic stress. The findings provide empirical evidence that reconciles the sometimes contradictory results of previous research and forms the basis for recommendations for research, practice, and policy.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1215
BATTERING, WORK, POVERTY: PREVALENCE AND BEYOND Overall, recent data indicate that domestic violence affects a considerable proportion of welfare recipients and other poor women. For example, one in seven welfare recipients in the Women’s Employment Study (conducted in Michigan) reported “recent severe domestic violence” in the past year (Danziger, 2001, p. 1; see also Tolman, Danziger, & Rosen, 2002). Longitudinal research with welfare recipients in Kern and Stanislaus Counties (in California) found that between one in six and one in five of the sample reported experiencing domestic violence (CalWORKs Project, 2002a, 2002b). The evidence for associations between battering and women’s labor force participation is less definitive than are the prevalence findings. Studies with groups of poor women in Chicago (Lloyd & Taluc, 1999); Worcester, Massachusetts (Browne, Salomon, & Bassuk, 1999); and Allegheny County, Pennsylvania (Brush, 2000) found widespread abuse. Using the Work/School Abuse Scale with a group from a battered women’s shelter, researchers found that several forms of work-specific interference, threats, and control were associated with higher levels of unemployment, school absence, missed work days, and other outcomes related to earnings capacity and, therefore, poverty (Riger, Ahrens, & Blickenstaff, 2000). Tolman et al. (2002) reported that persistent, severe, recent physical violence lowers women’s economic wellbeing and increases the hardships associated with poverty. At the same time, in all these studies, past or recent physical violence or work-related control is inconsistently associated with specific hardships, such as food insecurity or trouble paying bills and utilities, with work outcomes, such as earnings or weeks or hours worked, and with participation in welfare-to-work programs (for a detailed review, see Brush, 2002). The mechanisms through which abuse might trap women in poverty by obstructing work are complex, and so are the realities of women’s individual circumstances. Some abused women are able to use employment to spend time away from their abusers, to build supportive networks and resources, and to distract themselves
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1216
VIOLENCE AGAINST WOMEN / October 2003
from symptoms of posttraumatic stress. For others, injuries, distress, and sabotage make working or complying with the requirements of welfare reform extremely difficult. There is empirical evidence to support both accounts. Surely, both are true for different women or for the same woman at different times. Reconciling these contradictions is a central goal of my research. The results I report here extend and clarify researchbased knowledge of the complicated connections between poverty and abuse by looking at them the other way around. For welfare policies and services to respond adequately to the experiences and needs of battered women, we need to know more than just the ways battering impoverishes women by obstructing work. We also need to understand the effects of welfare, poverty, and work on battering. To that end, this research asked the following: What are the effects of poor women’s waged work on physical violence, work-related control and sabotage, and trauma symptoms? Does going to work precipitate, aggravate, or ameliorate abuse and symptoms? The optional provisions for battered welfare recipients in the U.S. welfare reforms of 1996 (known as the Family Violence Option) were based on the assumption that going to work aggravates battering and posttraumatic stress symptoms. If men use physical, emotional, sexual, and other abuse to control women, and if women’s employment represents a threat to men’s power and control, then in abusive relationships, women’s working will trigger the onset or escalation of battering. Moreover, the pressures of welfare sanctions, work requirements, and conditions in the generally low-wage jobs many former welfare recipients obtain could aggravate posttraumatic stress symptoms. Similarly, advocates for the poor and for welfare recipients assume that women who experience an increase in abuse or in related distress when they work will have lower wages and shorter or interrupted work histories, both of which contribute to poverty. This article describes welfare recipients’ subjective assessments of the effects of their working on physical violence, work-related abuse, and posttraumatic stress symptoms. I compare employment outcomes for women with different subjective assessments. The goal is to bring empirical data to bear on these two assumptions about work, poverty, and abuse.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1217
MEASURES AND METHODS RESEARCH SITE AND INTERVIEW ADMINISTRATION
During May to June 1998 and June 2001, researchers conducted structured interviews with welfare recipients enrolled in “work first” (job search, employment skills, and paid work experience) activities in Allegheny County, Pennsylvania. Enrollees were eligible if they were women, receiving Temporary Aid to Needy Families (the cash benefit for poor mothers and children), not pregnant, and at least 18 years old. Out of the 165 eligible enrollees in the two cohorts (1998 and 2001), 162 participated in the interviews. Respondents were released from program responsibilities for the duration of the interview, which took place in private at the program sites. Interviews ranged from 20 to 30 minutes for those conducted in 1998 and from 24 to 103 minutes for the extensive retrospective interviews administered in 2001. Interviewers compensated each respondent with a voucher from a regional supermarket chain. Interviewers offered all respondents small cards with information on safety planning, where and how to file for protective orders, and where to turn if they needed shelter from battering. The interviewers also arranged to make referrals to a local counselor specializing in abuse issues. The characteristics of the two cohorts of welfare recipients who participated in the study corresponded closely to the population of welfare recipients in the county. In the pooled data set (combining data from the 1998 and 2001 studies), the average respondent was 30 years old, was the only adult in the household, and had worked 14 weeks during the past year (including those who had not worked at all in the past year). Nearly 1 in 5 had dropped out of high school; 38% had attained only a high school diploma or equivalent. Employment was typically in predominately female, low-paid occupations. Earnings averaged $7.02 per hour in the most recent job. (For those who had any recorded earnings, the mean hourly wage excluding one outlier was $6.81.) Three quarters self-identified as Black. Of the women in the two cohorts, 60% had their first birth before age 20, and 65% of the fathers of their first child were older than 20 at the time of the first birth. One in 7 had an infant living in her household at the time of the interview.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1218
VIOLENCE AGAINST WOMEN / October 2003
And 3% were currently married, although roughly 1 in 4 were divorced or separated. The remainder (70%) had never married. The respondents faced considerable hardships associated with poverty. For example, 18% had been evicted or forced to move because they could not afford rent or utilities. MEASURES OF ABUSE AND ITS CONSEQUENCES
The two studies shared a common set of measures of physical violence, work-related interference and sabotage, and symptoms related to posttraumatic stress. Violence and Work-Related Control, Abuse, and Sabotage
The interview included a series of questions about whether the woman’s current or former boyfriend or husband ever engaged in physical or sexual violence against her, caused her serious injury, or used weapons against her (see Table 1). Interviewers asked if the partner in question had ever done any of the listed acts (coded 1 if yes, 0 otherwise). Interviewers also obtained a conservative measure of battering by asking if the respondent had ever filed for a Protection From Abuse order (PFA) (the civil restraining order available in the county) against that intimate partner. Interviewers also administered eight work-related control items and seven threat and interference items using the Work-Related Control, Abuse, and Sabotage Checklist (WORCASC) (for measurement analysis, see Brush, 2002). For each relationship, interviewers asked if the respondent’s partner had done any of the things listed in Table 2. For both the violence/injury items and the WORCASC items, an additional follow-up question elicited the respondent’s subjective assessment of whether that behavior “seemed to start, get worse, get better, or stay about the same” when she was working. Post-Traumatic Stress Symptoms
After the control, sabotage, and violence questions about each relationship, the interview protocol turned to a number of indicators of distress, some of which constitute criteria for the cognitive and emotional problems associated with post-traumatic stress
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1219
TABLE 1 Physical Violence Items (N = 162)
Abuse Indicator
Frequency
Hit, kick, or throw something at you (yes = 25) Threaten you with a knife or gun or use a knife or gun to hurt you (yes = 7) Demand to have sex with you or force you to have sex (yes = 12) Cut, bruise, choke, or seriously injure you (yes = 13)
Start or Worsen Get Better No Effect (%) (%) (%)
15
40
20
40
4
57
0
43
7
58
8
33
8
54
0
46
NOTE: Total physical violence items, M = 1, SD = 1.4. Final three columns may not sum to 100% due to rounding.
disorder (PTSD). PTSD is a characteristic set of symptoms associated with surviving a single serious trauma (in the case of natural disaster or accident victims, for example), with combat, with rape, and with the sorts of persistent physical, mental, and emotional abuse associated with battering as well as with childhood physical and sexual abuse (Herman, 1992; Root, 1992; Rothschild, 2000; for more extensive discussion, see Brush, 2003). The three diagnostic categories for PTSD are reliving the trauma, avoidance or numbing, and increased arousal or hypervigilance (Foa, Cashman, Jaycox, & Perry, 1997; Herman, 1992; Rigley, 1985; Root, 1992). In this study, questions about intrusive memories and nightmares measured reexperiencing the trauma. Questions about depression and diminished energy indicated avoidance or numbing. Questions about insomnia and irritability measured hypervigilance (see Table 3). Interviewers also asked respondents about other measures of physical and emotional distress that could present barriers to safety and solvency through waged work. These included problems with disordered eating (anorexia, bulimia, binge and purge eating), problem drinking, and pain from injuries caused by past or recent family violence (Herman, 1992; Thompson, 1994). For all of these possible consequences of abuse, respondents provided subjective assessments of the effects of work on the symptoms they experienced (“seemed to start, get worse, get better, or stay about the same”).
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1220
VIOLENCE AGAINST WOMEN / October 2003 TABLE 2 Work-Related Abuse (N = 162)
Abuse Indicator
Frequency
Work-related control Promise to provide child care and does not (yes = 29) Withhold the car keys or a promised ride (yes = 16) Pick fights when you need to leave for work, study, or just be alone (yes = 37) Take or wreck your books, homework assignments, or other materials for school, job training, or work (yes = 8) Take or wreck your work clothes or other important possessions, such as your glasses or dental appliance (yes = 9) Keep you up late at night or interrupt your sleep (yes = 36) Want or demand sex when you need to leave for work, study, sleep, take care of your family, or just be alone (yes = 20) Need help because of being drunk, high, or in trouble (yes = 17) Threats and interference Threaten to withhold money or gifts from you or your children if you work (yes = 12) Threaten to hurt you or your children or threaten to leave you if you work (yes = 17) Seem jealous that you might meet someone new at work or in job training (yes = 43) Tell you that you could never keep a job, learn, or accomplish things in life (yes = 20) Tell you that women should not work outside the home or that women who work outside the home are bad mothers (yes = 12) Tell you that you can only work outside the home if you keep up with the housework (yes = 7) Threaten to bother you at work or call or visit you at the training site or at work when it is not allowed (yes = 5)
Start or Worsen Get Better (%) (%)
No Effect (%)
18
62
24
14
10
38
19
44
23
76
3
22
5
50
0
50
6
56
0
44
22
61
11
28
12
55
20
25
11
59
6
35
7
67
0
33
11
71
0
29
27
67
7
26
12
50
25
25
7
84
8
8
4
86
0
14
3
60
0
40
NOTE: Total Work-Related Control, Abuse, and Sabotage Checklist items, M = 3.5, SD = 3.3.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1221
TABLE 3 Posttraumatic Stress Symptoms (N = 162)
Symptom
Frequency
Intrusive memories (yes = 41) Depression (yes = 41) Irritable or easily angered (yes = 34) Difficulty concentrating (yes = 33) Trouble falling or staying asleep (yes = 22) Easily startled (yes = 17) Nightmares (yes = 6) Want to make changes but do not have the energy (yes = 23) Dissociation (yes = 21) Other problems Drink to get drunk (yes = 9) Drink or get high to manage pain (yes = 7) Disordered eating (yes = 10)
Start or Worsen Get Better (%) (%)
No Effect (%)
25 25
2 24
83 61
15 15
21 20
27 21
50 64
24 15
14 11 4
36 24 17
46 47 33
18 29 50
14 13
35 52
39 29
26 19
6
33
11
56
4 6
43 30
14 60
43 10
NOTE: Total symptoms, M = 1.5, SD = 1.8.
EMPLOYMENT MEASURES
In addition to questions about relationships (including measures of interference, sabotage, and violence and the distress they may cause), the interviews recorded measures of work experience. In both sets of interviews, wage and occupation data referred to the current or most recent job. Interviewers also asked respondents how many weeks they had worked in the past year. FINDINGS In this section, I report women’s subjective assessments of the effects of their working on battering and the associations between subjective assessments and employment outcomes. DOES WORKING PRECIPITATE OR AGGRAVATE PHYSICAL VIOLENCE AND INJURY?
Recall that a central assumption of the provisions for battered welfare recipients is that working triggers or escalates battering.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1222
VIOLENCE AGAINST WOMEN / October 2003
Table 1 displays results for the items about physical violence and injury required to check this assumption. The first column gives the prevalence (the proportion of respondents who answered yes) of each item. About 1 in 6 respondents reported at least one of the four physical abuse items. Of the 162 respondents, 15% reported that their current or most recent partner hit, kicked, or threw something at them. The middle two columns give the percentage of those respondents who reported any instance of an item who said their working precipitated, aggravated, or ameliorated the abuse. Reading across the columns, of the 25 women who were hit, kicked, or had something thrown at them, 40% said this abuse started or became worse when they were working. And 20% said that being battered in this way lessened or was better when they were working. In the final column, the data appear for abuse that stayed about the same. The remaining 40% said that this type of physical abuse happened whether they were working or not and did not seem to be related to their employment. For the other three physical violence and injury items, the subjective assessments were even more clearly bimodal. At least half the women who reported being physically abused specifically said their working precipitated or aggravated the abuse. Most of the rest said it happened whether or not they worked. Only a very small fraction, or none, said their working improved the situation. Compare this finding with the fact that 1 in 5 women whose partners hit, kicked, or threw something at them said the abuse was less frequent or less severe when they worked. What researchers count as physical violence may determine the observed association between abuse and labor force participation. Between 33% and 46% of physically battered respondents said their working seemed not to be related to the onset, frequency, or severity of physical violence and injury. The women in this category did not experience their partners as abusing them in response to, or to prevent, their working. This finding is consistent with the notion that although batterers often blame women for provoking violence (“If you would just do x, I wouldn’t hurt you”), women’s actions seldom precipitate physical abuse. Approximately 4 out of 5 of the respondents who reported being physically battered said the abuse was the same or even worse when they worked. For up to half of the respondents in this category, working may not be any more risky than not working
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1223
when it comes to physical abuse. However, whether their working precipitated or aggravated physical battering or not, battered women may also need urgent accommodation to the ways battering can interfere with work, so that they are not sanctioned for having been hit or hurt. DOES WORKING AGGRAVATE WORK-RELATED CONTROL AND SABOTAGE?
The first column of data in Table 2 shows the results of asking respondents each of the eight control and seven threat and interference items of the WORCASC. Rates ranged from less than 5% to more than 25%, and the respondents averaged three items total. One quarter of the respondents reported four or more items on the checklist. The middle two columns in Table 2 give the proportion of those respondents who reported yes on an item who said, in response to a follow-up question, that working seemed to precipitate, aggravate, or ameliorate that type of abuse. In contrast to the physical abuse items, for 14 out of the 15 WORCASC items, at least half the abused respondents (and sometimes as many as four fifths) said that their going to work made the abuse start or increase. For 6 items, not a single woman said her working lessened the abuse. In about 1 in 4 cases, it appears that going to work deterred abusers from some forms of sabotage, such as verbally disrespecting or discouraging the woman or failing to provide promised child care. One in 5 women who reported unwanted or forced sexual demands said going to work lessened that abuse, and roughly the same proportion who reported sabotage related to transportation also said the situation improved when they worked. Finally, in the last column, the proportion who said their working seemed to have no effect on work-related control, abuse, and sabotage varied across the items from a low of 8% to a high of 44%. It appears (for most of these items) that going to work precipitated or aggravated the abuse rather than ameliorating it (by getting women out of the house, for example) or having no effect. Women’s subjective assessments of the temporal and causal relationship between their going to work and the behaviors measured by the checklist reinforce the notion that “work-related” control, abuse, and sabotage are in fact related to work. Abused women
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1224
VIOLENCE AGAINST WOMEN / October 2003
perceived that men were acting instrumentally when they engaged in these forms of control and sabotage, unlike when they were physically violent. If these women are right, then it is important to assess not only physical violence but also sabotage and interference with work to avoid sanctioning the women whose partners are most likely to disrupt their transition from welfare to work. DOES WORKING PRECIPITATE OR AGGRAVATE SYMPTOMS OF POSTTRAUMATIC STRESS?
The stunning pattern in the final three columns of Table 3 tells a very different story from the patterns of physical abuse and workrelated control and sabotage. The vast majority of women who reported intrusive memories, depression, and the hyperarousal symptoms (including hair trigger temper, difficulty concentrating, and sleep disturbance) indicated that their symptoms were better when they were working. Reported rates of posttraumatic stress symptoms ranged from 1 in 25 to 1 in 4 respondents (see Table 3). Just more than half (54%) the respondents reported at least one symptom. Relatively few women reported experiencing the other problems frequently related to traumatic stress and physical injuries caused by abuse. For four of the nine PTSD symptoms, at least half (and sometimes as many as four fifths) of the respondents who reported the symptoms said that when they went to work, they had some relief. Only women with nightmares said that work had no effect in such high rates, and only women reporting dissociation (a severe form of the numbing effects of trauma) more frequently reported that their working started or worsened that symptom than otherwise. The pattern of responses on these items is especially useful for interpreting the complicated links among battering, work, and poverty found in earlier research. People often develop eating disorders or use alcohol or other drugs to manage pain and trauma symptoms. Such behaviors are especially likely to interfere with learning, employment, and compliance with the requirements of welfare reform (Brush, 2003; Horsman, 2000). It is welfare recipients with these clusters of abuse-related symptoms and barriers who advocates have been most concerned to protect from sanctions and other pressures that could aggravate their
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1225
trauma symptoms and other obstacles to work. At the same time, these data support the assertion of some occupational and vocational therapists (see, e.g., Murphy, 1993) that a large proportion of women with PTSD symptoms find that the routines, social contact, and other aspects of waged work alleviate their symptoms, while earned income may increase their ability to leave abusive men (see also Davis, 1999; Raphael, 1999, 2000). The apparently contradictory effects of battering on women’s labor force participation may be explained at least in part by some abused women having relatively strong work outcomes because employment ameliorates their trauma symptoms. This finding reinforces the importance of screening not only for physical battering and workspecific abuse but also for PTSD symptoms. Battered welfare recipients must not be allowed to fall through the cracks if trauma symptoms obstruct employment, or if they might benefit from work in unforeseen ways. ARE THERE ECONOMIC RAMIFICATIONS WHEN EMPLOYMENT PRECIPITATES OR AGGRAVATES ABUSE?
For the subgroups of women who reported any (valid) wages in their most recent job (69% of the 162 respondents) and any weeks worked in the past year (66%), Table 4 arrays comparisons of these employment outcomes. The comparison here is within the group that reported any items on the WORCASC, physical abuse, and PTSD symptom lists and between those who said that going to work precipitated or aggravated the abuse or symptom (or who reported filing a civil restraining order) and those who said that was not the case. Table 4 presents means, the number of cases, and standard deviations. For each comparison, Table 4 also shows measures of the association between the dependent variables (wages and weeks worked, both continuous variables) and the independent variables (all ordinal categorical variables). As predicted by the assertions of advocates for battered women, the mean wages earned and weeks worked were generally lower for respondents who said their working made the abuse start or get worse than for those reporting abuse for whom that was not the case. The “wage penalty” for this pattern of abuse was 88 cents per hour in the case of WORCASC items, 90 cents per hour in the case of PTSD symptoms, and 35 cents per hour in the
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1226
VIOLENCE AGAINST WOMEN / October 2003
TABLE 4 Associations Between Work-Related Outcomes and Effects of Working on Abuse M
n
SD
Hourly wages in last job (nonzero) If WORCASC start or worsen If WORCASC otherwise If physical abuse start or worsen If physical abuse otherwise If PTSD symptoms start or worsen If PTSD symptoms otherwise If PFA filed If no PFA filed
$6.81 $6.57 $7.45 $7.79 $6.72 $6.38 $7.28 $6.55 $6.90
110 48 13 9 11 48 9 28 82
$1.80 $1.72 $1.94 $2.20 $1.56 $1.72 $1.93 $1.80 $1.80
Weeks worked in past year (nonzero) If WORCASC start or worsen If WORCASC otherwise If physical abuse start or worsen If physical abuse otherwise If PTSD symptoms start or worsen If PTSD symptoms otherwise If PFA filed If no PFA filed
22 23 27 21 26 22 32 24 21
107 48 11 9 8 50 6 31 76
14 15 17 14 17 15 17 15 14
a
Eta
Eta
.20
.04
.29
.08
.19
.03
.08
.00
.69
.41
.18
.03
.20
.03
.11
.01
2
NOTE: WORCASC = Work-Related Control, Abuse, and Sabotage Checklist; PTSD = post-traumatic stress disorder; PFA = Protection From Abuse order. a. One case reporting wages of $30 per hour was excluded. Excluding this single outlier decreased the mean by $0.21 and reduced the standard deviation by $1.00.
case of filing a PFA. For women concentrated in the low-wage occupations where many former welfare recipients find work, these are sizable wage differences, even if the variability in these small cohorts means the differences between groups were not statistically significant. The average number of weeks worked in the past year ranged from 4 to 10 between groups, except if a PFA was filed, where the difference was 3 weeks and in the opposite direction (presumably, because for filers the legal action is an effective way of stopping abuse that might otherwise obstruct employment). Eta is a measure of association equivalent to a correlation coefficient for one continuous and one ordinal variable. Eta2 is the equivalent of R2 and can similarly be read (moving the decimal point appropriately) as the percentage of variance in the continuous dependent variable accounted for by the ordinal independent variable. The WORCASC, physical abuse, and PTSD items were
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1227
all modestly associated with hourly wage, although none of them accounted for even as little as 10% of the variance in wages. The largest eta and eta-square values were for the relationship between the WORCASC items and weeks worked in the past year. The physical abuse and PTSD symptom items showed weak associations with weeks worked. CONCLUSIONS The findings of this research partially supported advocates’ assumptions. For many women, battering is aggravated by going to work. Certainly, for the majority of those who report either physical abuse or work-related interference and control, going to work either precipitates or aggravates the abuse, or seems to have no effect. Only a minority (at most, 25%) reported that working made the abuse slacken or stop. In addition, for the respondents in this study, work interruption and the pay penalty are generally, although not uniformly, worse for those women who report that working precipitates or aggravates battering or PTSD symptoms. RECOMMENDATIONS FOR RESEARCH, PRACTICE, AND POLICY
These findings suggest it is important for researchers to differentiate among measures of battering and its consequences. Using measures of physical violence alone can result in contradictory and sometimes counterintuitive research findings. One clarifying difference is between battering measures and PTSD symptom measures. Between 1 in 3 and 4 in 5 respondents with PTSD symptoms report that working brought some relief. They may be distracted from intrusive memories or supported through social contacts. Battered women may have better work-related outcomes than expected because, in some cases, the ameliorative effects of working on symptoms outweigh the aggravating effects on battering. My results also suggest that researchers and service providers should ask specifically about the relation battered women observe between their going to work and their being abused or suffering from symptoms. Breaking the links between poverty and abuse means mandating work only when it is unlikely to precipitate or
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1228
VIOLENCE AGAINST WOMEN / October 2003
aggravate abuse and PTSD symptoms. Asking women directly— not only in research settings but also in service provision and practice settings—is an excellent first step in the sort of risk-safety analysis that will allow for effective referrals and appropriate exemptions, and prevent damaging sanctions. Another practice and policy recommendation is universal screening for, and discussion of, work-specific abuse and trauma symptoms as well as physical battering. Screening for and discussing battering and trauma could be useful in the context of life skills courses, employment training, job placement, and other programs to monitor and promote women’s transitions from welfare to work. At the same time, the variations in women’s experiences suggest that service providers and policy makers should consider employment one part of a multidimensional strategy to enable women to escape the dual traps of poverty and abuse. Some women, those who experience an increase in abuse or symptoms when they work, will benefit from exemptions from welfare time limits and work requirements, such as those provided in the Family Violence Option. Others will benefit from work supports, especially transportation and child care, that meaningfully reduce their dependence on currently or formerly abusive partners. Most will benefit from discussing safety planning and the effects of work on relationships and trauma symptoms. Welfare policy and practice can most safely encourage work on a case-by-case basis depending in part on women’s perceptions of the relationship between their working and their being abused or experiencing symptoms. Examining the effects of work on hitting and hurting confirms the importance of listening carefully to women’s accounts of the risks, dangers, and possibilities in their own lives. Reauthorization of welfare reform legislation will more adequately meet the needs of poor women, and more effectively support battered women’s leaving poverty through waged work, when it takes these findings into account. REFERENCES Browne, A., Salomon, A., & Bassuk, S. S. (1999). The impact of recent partner violence on poor women’s capacity to maintain work. Violence Against Women, 5, 393-426.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
Brush / EFFECTS OF WORK ON VIOLENCE
1229
Brush, L. D. (2000). Battering, traumatic stress, and welfare-to-work transition. Violence Against Women, 6, 1039-1065. Brush, L. D. (2002). Work-related abuse: A replication, new items, and persistent questions. Violence and Victims, 17, 743-757. Brush, L. D. (2003). “That’s why I’m on Prozac”: Battered women, traumatic stress, and education in the context of welfare reform. In V. Adair & S. Dahlberg (Eds.), Reclaiming class: Women, welfare, and the promise of higher education in America (pp. 215-239). Philadelphia: Temple University Press. CalWORKs Project. (2002a). Impact of alcohol and other drugs, mental health problems and domestic violence on employment and welfare tenure (Policy and Practice Brief No. 2). Sacramento: California Institute for Mental Health. CalWORKs Project. (2002b). Welfare reform: Personal stories of four women who have faced alcohol and other drug, mental health and domestic violence issues (Policy and Practice Brief No. 4). Sacramento: California Institute for Mental Health. Danziger, S. K. (2001). Why some women fail to achieve economic security: Low job skills and mental health problems are key barriers. FORUM, 4, 1-4. Davis, M. F. (1999). The economics of abuse: How violence perpetuates women’s poverty. In R. A. Brandwein (Ed.), Battered women, children, and welfare reform: The ties that bind (pp. 17-30). Thousand Oaks, CA: Sage. Foa, E. B., Cashman, L., Jaycox, L., & Perry, K. (1997). The validation of a self-report measure of posttraumatic stress disorder: The Posttraumatic Diagnostic Scale. Psychological Assessment, 9, 445-451. Herman, J. L. (1992). Trauma and recovery: The aftermath of violence from domestic abuse to political terror. New York: Basic Books. Horsman, J. (2000). Too scared to learn: Women, violence, and education. Mahwah, NJ: Lawrence Erlbaum. Lloyd, S., & Taluc, N. (1999). The effects of male violence on female employment. Violence Against Women, 5, 370-392. Murphy, P. A. (1993). Making the connections: Women, work, and abuse. Orlando, FL: PMD Press. Olsen, K., & Pavetti, L. (1996). Personal and family challenges to the successful transition from welfare to work: How prevalent are these potential barriers to employment? Washington, DC: Urban Institute. Personal Responsibility and Work Opportunity Reconciliation Act of 1996, Pub. L. No. 104193 (1996). Raphael, J. (1996). Domestic violence and welfare receipt: Toward a new feminist theory of welfare dependency. Harvard Women’s Law Journal, 19, 201-227. Raphael, J. (1999). Keeping women poor: How domestic violence prevents women from leaving welfare and entering the world of work. In R. W. Brandwein (Ed.), Battered women, children, and welfare reform: The ties that bind (pp. 31-43). Thousand Oaks, CA: Sage. Raphael, J. (2000). Saving Bernice: Battered women, welfare, and poverty. Boston: Northeastern University Press. Riger, S., Ahrens, C., & Blickenstaff, A. (2000). Measuring interference with employment and education reported by women with abusive partners: Preliminary data. Violence and Victims, 15, 161-172. Rigley, C. R. (1985). Trauma and its wake: The study and treatment of post-traumatic stress disorder. New York: Brunner/Mazel. Root, M. P. P. (1992). Reconstructing the impact of trauma on personality. In L. S. Brown & M. Ballou (Eds.), Personality and psychopathology: Feminist reappraisals (pp. 229-265). New York: Guilford.
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
1230
VIOLENCE AGAINST WOMEN / October 2003
Rothschild, B. (2000). The body remembers: The psychophysiology of trauma and trauma treatment. New York: Norton. Scott, E. K., London, A. S., & Myers, N. A. (2002). Dangerous dependencies: The intersection of welfare reform and domestic violence. Gender & Society, 16, 878-897. Thompson, B. W. (1994). A hunger so wide and so deep: A multiracial view of women’s eating problems. Minneapolis: University of Minnesota Press. Tolman, R. M., Danziger, S. K., & Rosen, D. (2002). Domestic violence and economic well-being of current and former welfare recipients (Working Paper 304). Ann Arbor, MI: Joint Center for Poverty Research. Tolman, R., & Raphael, J. (2000). A review of research on welfare and domestic violence. Journal of Social Issues, 56, 655-682.
Lisa D. Brush is an associate professor and director of graduate studies in sociology at the University of Pittsburgh. She directs the Family Violence & Self-Sufficiency Project, a research program dedicated to converting policy analysis into feminist intelligence. Results from her research on battering and work in the context of welfare reform may be found in the Journal of Poverty, Journal of Sociology and Social Welfare, Violence Against Women, and Violence and Victims. She is the author of Gender and Governance, a book in the Gender Lens Series published by AltaMira Press (2003).
Downloaded from vaw.sagepub.com at UNIV OF WISCONSIN MADISON on May 15, 2011
