An Analysis of Empirical Estimates of Sexual Aggression Victimization and Perpetration

Article excerpt

Estimates of prevalence for several categories of sexual coercion, including rape and attempted rape, were statistically aggregated across 120 studies, involving over 100,000 subjects. According to the data, almost 13% of women and over 3% of men have been raped, and almost 5% of men claim to have perpetrated rape. In contrast, about 25% of women and men claim to have been sexually coerced and to have perpetrated sexual coercion. In general, the mediating variables examined-population type, decade, date of publication, and type of operationalization-were not consistently related to rates of victimization or perpetration. Nevertheless, the extensive variation among study estimates strongly suggests the possibility of systematic sources of variation that have yet to be identified. Further analyses are called for to disentangle such sources.

Sexual violence is an incendiary topic in society. Various groups lay claim to their preferred estimates of this problem in order to advance their agendas, while individual victims often give mute testimony through the lenses of summary objective data. Estimates of the incidence of sexual violence provide at least one barometer of social and cultural health. They also serve as a key stimulus for rhetorical and political exigency, which in turn affect public policy and societal response to such problems. Finally, publicized estimates of the risk of crime potentially play a significant role in individuals' formulation of their sense of personal security. Public paranoia about the personal risk of crime victimization can translate into lives of fear, seclusion, and excessive and even aggressive defensiveness. A physically, psychologically, and emotionally traumatizing problem experienced involuntarily by 1% of all Americans is often viewed as a major concern, but such a problem experienced by 5%-25% of the public could easily be viewed as a national crisis. Estimates of sexual violence and coercion victimization run this gamut. Thus, the rancor over the legitimate estimation of sexual violence is no small matter. Such estimates may well signal the priority the problem occupies in the pecking order of various cultural, social, political, and individual agendas (see, for example, Gilbert, 1991; Klein, 1994; Laura, X., 1994; Muehlenhard, 1994; Muehlenhard, Powch, Phelps & Giusti, 1992; Muehlenhard, Sympson, Phelps, & Highby, 1994; Orton, 1994; Roiphe, 1993).

The traditional scientific basis for deciding among various estimates of virtually anything is to examine critically which estimates were based on more methodologically sound procedures. Studies with larger samples, random samples, more valid operationalizations, and higher reliability ordinarily would be viewed as the preferred estimates. However, even when methods arguably evidence all of these features, they are open to further criticism for being valid only in circumscribed or systematically biased ways. For example, despite what is clearly the largest and probably most representative data base, the Bureau of Justice Statistics (BJS) National Crime Victimization Survey (NCVS), is often criticized for employing overly narrow or constrictive questions about sexual victimization (see Koss, 1992,1996; Koss, Gidycz, & Wisniewski, 1987), although the redesigned survey seems to have satisfied some of the previous criticisms (Bachman, 1994; Bachman & Taylor, 1994; BJS, 1994; cf. Koss, Goodman, Browne, Fitzgerald, Keitz, & Russell, 1994).

Others have attempted to develop alternative approaches that avoid some of the limitations of the NCVS by assessing forms of sexual victimization in terms of highly objective reports of behaviors that meet legal standards of offense. By far the most common scholarly method currently employed is Koss's Sexual Experiences Survey (SES), which has undergone extensive development and application (e.g., Koss & Gidycz, 1985; Koss, & Oros, 1982; Muehlenhard & Linton, 1987). …