Measuring gender differences in partner violence: implications from research on other forms of violent and socially undesirable behavior

Sex Roles: A Journal of Research,  June, 2005  by Sherry L. Hamby

• E-mail
• Print
• Link

Focus groups held prior to the Statistics Canada Violence Against Women survey (Johnson & Sacco, 1995) found that many respondents were confused by items on the original CTS that referred to throwing things or being hit with things, and they added the phrase "that could hurt" to emphasize that physical play should not be included (a modification subsequently made to the revised CTS, Straus et al., 1996). In another focus group study, participants said terms that are often used interchangeably on sexual assault questionnaires, such as "unwanted," "involuntary," and "forced," had different meanings, not all of which connoted violence (Hamby & Koss, 2003). In a survey of incarcerated men, an item on "demanded sex whether she wanted to or not" caused so many questions it had to be dropped (Cook, 2002, p. 552).

Similarly, a review of narratives about physical force found that 2.4% of female and 1.6% of male college students provided accounts that were entirely about playful episodes such as pillow fights (Hamby & Tsang, 1996). Similar numbers provided accounts that began as physical play but degenerated into violence, suggesting that it is not just reporting errors but also true ambiguity of some events. This was less common in a group of older, primarily married women, in which only 1% provided accounts that began as play but degenerated into violence, and none reported entirely playful episodes. In two studies that recorded queries during closed-ended surveys, one found that 7% of respondents (gender not recorded) asked whether they should report incidents of physical play, and in the other, 5% of women and 0% of men asked similar questions (Hamby & Tsang, 1996).

Aggression researchers tend to assume that the phenomenon of violence is clear-cut, but the interpretation of any action is socially constructed (Muehlenhard & Kimes, 1999). For example, one can be pushed or shoved not only during physical play, but also in accidents, team sports, or to be taken out of harm's way. There are also settings, such as subway stations during rush hour (Levine, 1976), where pushing is commonplace. By and large, such incidents are not socially constructed as violence, which implies behavior that is deemed at the least as inappropriate, and more typically as intentionally causing unwanted suffering (Pontara, 1978). Researchers tend to assume that there is a universal or at least culture-wide standard for interpreting acts, but gender, race, class, overall relationship quality, trauma history, and other issues probably affect what is perceived as "violence." Although the percentage of reports affected by ambiguous incidents may be small, partner violence research should still pay more attention to the problem of distinguishing between aggressive and nonaggressive acts.

Evidence of Overreporting from Records-Check Studies

Records-check studies have primarily found evidence of significant underreporting. Some studies, however, have also found evidence of overreporting. Although errors in official records, or a failure to obtain the records of the relevant jurisdiction, can hamper interpretation of apparent overreporting, estimates of the number of individuals who report official police contact when no record can be found have ranged from 10% to 30% (Huizinga & Elliott, 1984).