Another way to test whether MMPI elevations in battered women are characterological or reactive would be to investigate whether the length of time a woman is abused, the frequency of the abuse, or the severity of the abuse is positively correlated with her elevations on the MMPI. If so, then the elevations are more likely to be reactive than characterological. Rosewater found that the frequency of the abuse correlated positively with elevations on the 1, 2, 6, 7, and 0 scales. Two other studies have used similar methodology, with similar results. n72
n72 Khan et al., supra note 30; Rollstin & Kern, supra note 30.
Khan et al. administered the MMPI-2 to thirty-one battered women in shelter. Each participant also completed a questionnaire regarding the length of time she had been abused, which of nine types of physical abuse she had experienced, and which of nine types of psychological abuse she had suffered. Each participant also was asked to estimate, on a scale of zero to nine, the severity of the psychological abuse and the severity of the physical abuse.
On the mean profile of these battered women, scales 4, 6, and 8 were elevated, thus replicating many of Rosewater's findings. n73 What predicted the overall average T-score was severity of psychological abuse. n74 Based on all their findings, Khan et al. concluded that MMPI elevations are, to some extent at least, reactive, and that clinicians should be careful to avoid misdiagnosing battered women as mentally ill. n75
n73 Id. at 104.
n74 Id. at 107.
n75 Id. at 109.
Rollstin and Kern directly posited as their research question the inquiry first raised peripherally by some previous researchers and directly by Rosewater and by Khan et al.: Are battered women's elevations on the MMPI clinical scales characterological or reactive to the abuse they have suffered? Rollstin and Kern compared MMPIs on currently battered women with MMPIs on women who had been out of the abusive relationships for a year, to see if the longer women are out of abusive situations the more [*104] normal they appear on the MMPI (i.e., the more their profiles resemble the profiles of average, nonabused women).
They reported that the results of their study "provided only mixed support" for the hypothesis that the MMPI-2 elevations of battered women were a reactive state. n76 On the one hand, the elevations of formerly battered women were not significantly different from the elevations of recently battered women. On the other hand, after performing more complex analyses, Rollstin and Kern determined that severity of both physical and psychological abuse were significantly associated with MMPI-2 scores. n77 Concerning the finding that the MMPI-2 elevations were not related to the duration of time since the abusive relationship had ended, Rollstin and Kern queried whether "the period of time these women had been out of their relationships had still not been long enough for them to experience emotional improvement." n78
n76 Rollstein & Kern, supra note 30, at 393.
n77 Id. at 393.
An additional finding is directly applicable to the issue of the use of the MMPI in custody cases involving battered women:
Unexpectedly, larger numbers of children were significantly associated with increased psychologic disturbance. . . . The number of children was not associated with any other measure, including physical abuse . . . or psychological abuse. . . . n79
n79 Id. at 392.
Rollstin and Kern theorized that this finding may reflect the amount of life disruption which occurs after the battered woman has left the abusive relationship as well as the emotional difficulties experienced by the children both during and after the relationship. Experience suggests that it often takes a year or more to stabilize one's life situation and that children exacerbate the extent of that disruption. n80
n80 Id. at 393.
This should be kept in mind by custody evaluators when interpreting MMPIs of battered mothers.