Linda L. Marshall

Development of the severity of violence against women scales

In response to the need for more sensitive assessment instruments, scales were developed applicable to the evaluation of male violence against women. Two versions of the Severity of Violence Against Women Scale (SVAWS) were developed. On 10-point scales, college females (N = 707) rated how serious, aggressive, abusive, violent, and threatening it would be if a man carried out each of 46 acts with a woman. The mean of each act across ratings was calculated and submitted to factor analysis. Nine factors represented symbolic violence: threats of mild, moderate, and serious violence; actual mild, minor, moderate, and serious violence; and sexual violence. Community women (N = 208) rated the acts on seriousness, aggressiveness, and abusiveness. All factors were unidimensional. Second-order factor analysis confirmed the existence of two broader dimensions representing physically threatening acts and actual violence. Ratings of the amount of physical and emotional harm provided the weightings for future research with student (SVAWS-S) and adult (SVAWS) samples.

Distress and symptoms of posttraumatic stress disorder in abused women

Emotional distress and symptoms of posttraumatic stress disorder (PTSD) were examined among women in psychologically abusive relationships. Women (N = 93) were divided into three groups (none, moderate, severe) according to scores on the violence subscale of the Severity of Violence Against Women Scales (Marshall, 1992). All groups reported serious emotional distress on the Symptom Checklist 90-Revised (Derogatis, 1983) dimensions (e.g., more so than 93% of the nonclinical norm for global distress). Psychoticism was the highest subscale for all groups. Most women (56%) suffered PTSD according to a subscale of the SCL90 (Saunders, Arata, & Kilpatrick, 1990). Difficulties with perception, memory, and motor functions (cognitive failure) more consistently predicted intrusive thoughts, PTSD scores, and attempted suicide than did women’s attention to their inner thoughts and feelings (private self-consciousness), which was important for the sample and the subgroup that had sustained severe violence. Limitations and contributions are discussed as are different uses for conservative and inclusive measures of PTSD.

Effects of men's subtle and overt psychological abuse on low-income women.

Asocial influence approach to the psychological abuse of women (Marshall, 1994; 1996) was expanded and tested. Distinctions are made between obvious acts (e.g., verbal aggression, controlling behaviors), overt acts which are easily recognized and described, and subtle acts which are least likely to be recognized as psychologically abusive. Men’s violence and sexual aggression, and overt (dominating acts, indifference, monitoring, discrediting) and subtle (undermining, discounting, isolating) psychological abuse were examined as they related to women’s psychological and emotional state and perceptions of their relationship. Results of regression equations with 834 low-income women in long-term heterosexual relationships are reported. In general, subtle psychological abuse had stronger and more consistent associations with women’s state and relationship perceptions than did their partners’ overt psychological abuse, violence, or sexual aggression. The importance of extending research beyond obvious acts was underscored by findings showing that subtle psychological abuse accounted for a small but significant proportion of the variance in outcome variables even after the effects of violence and sexual aggression (Step 1) and overt psychological abuse (Step 2) were controlled in eight of the nine regression equations. In contrast, when subtle and overt psychological abuse were entered first (in Steps 1 and 2, respectively), violence and sexual aggression (Step 3) made significant contributions in only two of the nine equations.

PTSD symptoms and partner abuse: Low income women at risk.

Aspects of partner abuse (types, severity, chronicity, treatment of injury, fear, etc.) were addressed with low income community women, half of whom scored above the cutoff on the Crime Related PTSD scale. Using this cutoff, 47% of women who sustained moderate violence were high in CR-PTSD symptoms. If their partner also raped them, the rate (63%) was similar to women who sustained severe violence (65%) or severe violence and rape (71%). No ethnic differences were found for rates or severity of CR-PTSD symptoms. A MANCOVA by ethnicity (African Americans, Euro-Americans, Mexican Americans) and CR-PTSD symptoms (low vs. high) identified ethnic differences only on total sexual aggression and recent threats of violence. The high symptom group reported more abuse on all measures. Results from the CR-PTSD and the general lack of ethnic differences support the notion that SES contributes more to womens vulnerability to abuse and stress symptoms than does ethnicity.

Psychological Abuse of Women: Six Distinct Clusters

This study was designed to identify patterns of psychological abuse (abuse) and determine whether different patterns mediate the effects of violence and sexual aggression. Interviews were completed with 578 (80.3%) of 720 women who volunteered for a study of “bad” or “stressful” heterosexual relationships. Cluster analysis with 51 items assessing many types of subtle and overt psychological abuse identified six groups. Cluster 1 sustained the most serious abuse, violence, and sexual aggression, but partners did not denigrate women or control finances. Cluster 2 sustained serious abuse which was more subtle than the dominating-controlling abuse discussed by others and had moderate violence and sexual aggression scores. Cluster 3 scored relatively low on abuse but reported controlling types (e.g., isolation, enforced secrecy) and relatively little violence and sexual aggression. Cluster 4 sustained the least abuse, violence, and sexual aggression but the abuse was overt (e.g., criticism, several types of control). Clusters 5 and 6 were similar with high abuse and moderate violence and sexual aggression scores but they experienced very different types of abuse. Total psychological abuse, threats of violence, acts of violence, and sexual aggression scores made different contributions to womens health, help seeking, and relationship perceptions depending on the pattern of abuse they sustained. Results for each type of harmful act are discussed separately.

Premarital violence: the impact of family of origin violence, stress, and reciprocity

Four hundred and fifty-four premarital undergraduates reported their stress and observation and experience of violence in their families of origin. They also reported the violence they have expressed and received in adult love relationships. Among females 53% to 71% of the variance in their expression and receipt of violence was accounted for by their own receipt and expression of violence (respectively) and by their having been abused as a child. This suggests that what they do is closely related to what is done to them. In contrast, among males less than 1% to 10% of the variance in their expression and receipt of violence was accounted for by the same variables. Neither negative stress ratings or the observation of violence were important explanatory variables. Surprisingly, positive stress contributed to males expressing violence.

Differing Effects of Partner and Nonpartner Sexual Assault on Women's Mental Health:

This study contrasted the effects of intimate partner and nonpartner sexual assault on womens mental health among a sample (N = 835) of low-income, ethnically diverse community women. Compared to sexual assault by a previous partner or by a non-intimate partner, sexual assault by a current partner was the strongest predictor of PTSD, stress, and dissociation. Non–intimate partner sexual assault was only a significant predictor of PTSD and only for African American women. These findings suggest that the victim-offender relationship is important when considering the impact of sexual assault. Specifically, sexual assault perpetrated by an intimate partner may be especially traumatic.

Childhood and adolescent sexual abuse of community women: mediated effects on psychological distress and social relationships.

UNLABELLED Possible mediators of sexual abuse severity were tested on the basis of D. Finkelhor and A. Brownes (1985) traumagenic dynamics model with 178 low-income African American, European American, and Mexican American community women interviewed for Project HOW Health Outcomes of Women. This subsample reported contact sexual abuse before the age of 18 years. Severity was level of force, number of perpetrators, relationship to perpetrator, and age at first assault. As expected, structural equation modeling showed powerlessness, and stigmatization largely mediated the effects of sexual abuse severity on womens psychological distress in adulthood. Powerlessness also mediated the effects of severity on maladaptive social relationships. The expected path from betrayal to relationships was nonsignificant. Overall, the results support extension of D. Finkelhor and A. Brownes model. Possible interventions are addressed.

Toward Ethnically Specific Models of Employment, Public Assistance, and Victimization

Among 836 low-income women, those receiving Aid to Families With Dependent Children (AFDC) or food stamps had experienced more coercive sexual assault, abuse by past partners, psychological abuse by current partners, and types of victimization than women not receiving assistance. The two groups of women were equally likely to endure threats or violence from current partners. African Americans and European Americans were more likely to have been victimized than Mexican Americans. European Americans reported more severe victimization except current partner violence. Multiple regressions on employment and assistance showed victimization predictors that varied by ethnicity. The effects of abuse by current partners were limited and are likely to be indirect.

A unified approach to exploratory factor analysis with missing data, nonnormal data, and in the presence of outliers

Factor analysis is regularly used for analyzing survey data. Missing data, data with outliers and consequently nonnormal data are very common for data obtained through questionnaires. Based on covariance matrix estimates for such nonstandard samples, a unified approach for factor analysis is developed. By generalizing the approach of maximum likelihood under constraints, statistical properties of the estimates for factor loadings and error variances are obtained. A rescaled Bartlett-corrected statistic is proposed for evaluating the number of factors. Equivariance and invariance of parameter estimates and their standard errors for canonical, varimax, and normalized varimax rotations are discussed. Numerical results illustrate the sensitivity of classical methods and advantages of the proposed procedures.