Comparing Perception of Sexual Intent and the Use of Sexual Coercion in Men and Women College Students

ABSTRACT

Despite research showing that many college men and women have experienced misperception of their friendliness as sexual intent, such research has focused on this type of misperception only as a correlate of men’s sexual aggression. In fact, regardless of methodology used many researchers seem to suggest women do not misperceive men’s sexual intent, and in some instances may actually under-perceive it. We used a hypothetical scenario to determine whether men (n = 324) and women (n = 689) college students perceived similar sexual intent from a character who is not the same gender as they are, as depicted in a story about a man and woman on a “date.” Our results revealed that men and women in our sample reported similar levels of perceived sexual intent on the part of the character with a different gender as described in the scenario, even after that character clearly indicated to the partner that they “think they do not want to have sex.” In addition, the perceived level of the character’s sexual intent as solicited in response to this scenario design was related to sexual coercion intentions among both men and women (though it appears more strongly related among men), and these relationships remained even after controlling for other known correlates of sexual coercion (e.g., rape myth acceptance, level of sexual arousal). Implications for the study of misperception and its origins are discussed.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ The scenario used in this study is based partly on a longer version developed by Norris et al. (2009). The authors would like to thank Dr. Norris and her colleagues for allowing us to review their materials to aid us in constructing our scenario.

² We chose to estimate Ordinary Least Squares regression models only for the mean verbal coercion outcome variable, as it appeared normally distributed. We used binary Logistic regression for the illegal coercion outcome because that data, especially among women was not normally distributed (see below). The coefficient comparison test (Clogg et al., 1995), however, is not suited to use with coefficients from non-linear regression models (e.g., Logistic regression). As mentioned, OLS models assume that the dependent variable was normally distributed. Several authors have argued that data can be considered normally distributed if skewness is between ‐2 to +2 and kurtosis was between −7 to +7 (see, Byrne, 2010; George & Mallery, 2010; Hair et al., 2010). Among the combined sample of men and women, our mean verbal coercion scale did in fact appeared to be normally distributed (skewness = 0.90, kurtosis = −0.25), such that use of OLS appeared appropriate. More specifically, among men the verbal scale also appeared to be normally distributed (skewness = 0.44, kurtosis = −1.02), as it did among the sample of women (skewness = 1.13, kurtosis = 0.49). On the other hand, our mean illegal coercion variable was non-normally distributed (skewness = 2.51, kurtosis = 6.35) because many respondents reported a 0% intention to use either one or the other of those “illegal” tactics. Among men specifically, the mean illegal coercion variable did actually appear to be normally distributed (skewness = 1.93, kurtosis = 3.12); however, among women this variable is non-normally distributed (skewness = 2.88, kurtosis = 9.01). Since illegal coercion, especially among women is not normally distributed, we report results from Logistic regressions for the illegal coercion outcome variable. In addition, we re-estimated all of the OLS models (i.e., for verbal coercion) presented in Tables 3 and 4 using Logistic regressions, and in each case the relationships of perceived willingness, gender, rape myth acceptance and arousal level to the odds of reporting coercion intentions were similar to those we generated using OLS methods examining sexual coercion likelihood scores.

³ We also compared this same perceived willingness variable between men and women in the arousal condition only, and also in the control group only and again there were no gender differences in the perceptions of the other person among these two separate sub-samples either. Specifically, among those in the control group men (56.9%) and women (55.3%, t(518) = −.564, n.s.) reported similar perceptions of willingness. Among those in the arousal group, men (57.6%) and women (54.8%, t(486)= −.901, n.s.) reported similar perceptions of willingness as well.

⁴ In addition to the results for the entire sample (i.e., cases from arousal and control conditions together), we also examined these correlations in the arousal and control groups separately, and found them to be similar between men and women within each of these groups. For instance, among men in the control group the correlation between the perceived willingness and likelihood of coaxing the other person to remove their clothes to obtain sex was r = .43 (p < .01) and among women in the control group the correlation was r = .23 (p < .01). Within the arousal condition group, this same correlation was r = .35 (p < .01) for men and r = .30 (p < .01) for women. As in the entire sample, perceived willingness was not significantly correlated with intentions to get the other person drunk among women in either sub-sample, but they were for men in the arousal and control groups separately. Other results available from the first author upon request.

⁵ An anonymous reviewer suggested we examine the relationship of perceived willingness to each coercion measure, without including the controls for arousal and rape myths, and then estimating the model again with those controls included to determine if misperception had a relationship over and above other known correlates. We did estimate these kinds of models and the relationship between perceived willingness and each verbal and illegal coercion intention outcome was always significant regardless of what other variables were included in the models.

⁶ In this case the perceived willingness measure was mean-centered prior to multiplying it by the gender variable to help control for any multicollinearity in the models once the interaction term is added (e.g., statistical overlap between the interaction term and original variables, gender and willingness). This process appears to have been successful in that none of the relevant coefficients (gender, perceived willingness, nor the interaction term) yielded Variance Inflation Factors (VIF) in the OLS regression for verbal coercion that were greater than 1.5 (According to SPSS Help, VIF greater than 2.0 are considered problematic).

Additional information

Funding

This work was not supported by any funding sources.