Use of a Bogus Pipeline to Detect Men’s Underreporting of Sexually Aggressive Behavior

Abstract

Researchers rely on men’s self-reports of sexually aggressive behavior to identify the prevalence and correlates of men’s perpetration (i.e., their use of verbal or physical coercion to obtain sex from a nonconsenting partner). However, results from research based on men’s self-reported sexual perpetration are suspect because men may intentionally conceal their sexually aggressive behavior. The current study used a Bogus Pipeline (BPL) methodology to determine whether young, community men (N = 93) intentionally underreport their use of sexually aggressive strategies on two self-report measures of perpetration. Compared with men in a Standard Testing (ST) condition, men in an experimental BPL condition—who believed that the honesty of their responses was being monitored—had 6.5 times greater odds of endorsing the use of illegal sexual assault strategies on a widely used measure. Indeed, over a third of men in the BPL condition admitted to using such strategies. However, there was no significant difference in men’s reports of verbally coercive strategies in the BPL versus the ST condition on the same measure. There were also no significant differences in reporting on the other, less commonly used measure of sexual aggression as a function of condition. Based on these preliminary findings, men’s use of sexual assault strategies may be far more common than is suggested by most self-report studies. Self-reports of verbal coercion, however, may be more trustworthy.

Keywords

sexual assault adult victims offenders

Using behaviorally specific self-report measures, approximately 15% of women in the United States report experiencing an act perpetrated by a man that is consistent with legal definitions of rape at some point in their lifetime; paradoxically, only approximately 5% of men self-report perpetrating an act against a woman that is consistent with legal definitions of rape (see Kolivas & Gross, 2007; Spitzberg, 1999, for reviews). Some researchers have noted that perpetrators often commit multiple acts of sexual aggression, and thus, they have explained this paradox by saying, “A relatively small proportion of men are responsible for a large number of rapes” (Lisak & Miller, 2002, p. 80). This explanation is somewhat comforting, as it suggests that, if the small group of repeat rapists can be incarcerated or rehabilitated, women will be safe from rape. However, there is reason to suspect that sexually aggressive behavior actually may be more common than would be suggested by this conclusion.

For example, in a study of college students, Edwards, Bradshaw, and Hinsz (2014) found that 31.7% of men reported that they would act on intentions to force a woman to engage in sexual intercourse “if nobody would ever know and there wouldn’t be any consequences” (p. 190). Although this study asked about a hypothetical behavior, the fact that nearly a third of men reported their willingness to engage in sexual assault suggests that rape may be part of a larger cultural problem, rather than the result of a few “bad apples.”

In addition, many women report having been raped multiple times (Classen, Palesh, & Aggarwal, 2005) either by the same perpetrator or by multiple perpetrators. For example, Daigle, Fisher, and Cullen (2008) found that there were high rates of repeat sexual victimization among women during a 2-year period of college, with 47% of those who experienced sexual victimization experiencing more than one act of victimization and 21% experiencing three or more victimizations, leading the authors to conclude that “a small proportion of college women experience a large proportion of . . . sexual victimizations” (p. 1296). Thus, although many perpetrators engage in multiple acts of sexual assault, this is unlikely to fully account for the discrepancy between reported rates of perpetration and victimization given that many victims also experience multiple acts of sexual assault.

Given these findings that challenge the idea that a small number of men perpetrate the majority of rapes, it seems important to consider another—potentially less comforting—explanation for this paradoxical discrepancy in men’s reports of aggression and women’s reports of victimization: The discrepancy may result from men’s underreporting of their sexually aggressive behavior on self-report measures (Cook, 2002; Kolivas & Gross, 2007; Strang, Peterson, Hill, & Heiman, 2013).

The Sexual Experiences Survey (SES; Koss, Gidycz, & Wisniewski, 1987; Koss & Oros, 1982; Sexual Experiences Survey–Long Form Perpetration [SES-LFP]; Koss et al., 2007) and the Sexual Strategies Scale (SSS; Strang et al., 2013) are examples of self-report measures of sexually aggressive behavior. Both measures employ behaviorally specific language (i.e., describe sexually aggressive behavior using operational definitions rather than asking explicitly about “sexual coercion,” “sexual assault,” or “rape”), and both measures assess two types of strategies that could be used to obtain nonconsensual sexual acts: (a) strategies that are not illegal but that involve verbal coercion, including pressure and manipulation, and (b) strategies that are consistent with illegal sexual assault, including taking advantage of an individual’s intoxication/incapacitation, threatening physical harm, or using physical force. Self-report measures such as the SES and the SSS are the primary methods that researchers use to identify nonincarcerated men who have perpetrated sexual aggression and thus to examine attitudinal, personality, and social risk factors associated with sexual aggression perpetration. In particular, most research on men’s perpetration has relied on some version of the SES.

Self-report research on sexual aggression using measures such as the SES and the SSS could be threatened by both intentional and unintentional misreporting. Recognizing a sexual experience as coercive or nonconsensual depends on subjective perceptions and interpretations of others’ behavior and intent, as well as on personal schema or “scripts” for what coercive versus consensual sex look like (Carroll & Clark, 2006). It is possible for one individual to perceive a sexual experience as aggressive or nonconsensual, while another individual involved in the same sexual encounter perceives the experience as consensual. In addition, it is likely that some perpetrators of sexual aggression might have difficulty recognizing, acknowledging, or recalling their own aggressive behavior, resulting in unintentional misreporting on measures of sexual aggression. The problem of unintentional misreporting has been addressed in a few prior studies (e.g., Buday & Peterson, 2015; Ross & Allgeier, 1996; Strang & Peterson, 2016).

Less commonly addressed is the problem of intentional misreporting of sexual aggression on self-report measures. Sexual perpetration research seems intuitively threatened by intentional underreporting or concealment because sexual aggression measures assess particularly sensitive behaviors, which are socially undesirable and sometimes illegal (Cook, 2002; Heilbrun & Loftus, 1986; Ouimette, Shaw, Drozd, & Leader, 2000; Strang et al., 2013). Socially desirable responding refers to the tendency toward strategic self-reporting “for the purpose of looking good” (Meston, Heiman, Trapnell, & Paulhus, 1998, p. 148). Intuitively, one would suspect that social desirability affects sexual aggression reporting; however, some studies demonstrate a relationship between measures of socially desirable responding and self-reports of sexual aggression, and other studies find no evidence of a relationship (Cook, 2002; Porter, Critelli, & Tang, 1992; Strang & Peterson, 2012; Strang et al., 2013; Walker, Rowe, & Quinsey, 1993). However, even a lack of relationship between social desirability measures and self-reports of sexual aggression does not provide assurance that men are not intentionally underreporting sexual aggression. Measures of social desirability include items such as “I never swear” and “I sometimes tell lies if I have to” (Paulhus, 1991). It is feasible that some men might be willing to admit to swearing or lying but not be willing to admit to sexual aggression. Thus, although there is reason to suspect that men may intentionally underreport their sexual aggression, it is difficult to prove that they are doing so.

In the 1970s, social psychologists developed the Bogus Pipeline (BPL; Jones & Sigall, 1971) procedure as a laboratory tool to increase honesty in self-reporting. In this procedure, participants are led to believe they are being monitored by a device, resembling a lie detector, which can determine their truthfulness in responding. A BPL condition creates a demand for honest, self-aware reporting that surpasses the demand for socially desirable responding. Across a variety of BPL studies, findings demonstrate that the procedure reliably reduces socially desirable responding and increases honesty, especially for issues of fact, like the presence or absence of specific behaviors (Roese & Jamieson, 1993; Tourangeau, Smith, & Rasinski, 1997).

Several studies have demonstrated that a BPL condition affects self-reporting of sexual behavior (Alexander & Fisher, 2003; Fisher, 2013; Tourangeau et al., 1997). For example, Fisher and colleagues (Alexander & Fisher, 2003; Fisher, 2013) found that women report significantly less sexual behavior than men on traditionally administered questionnaires. Men and women report similar levels of sexual activity, however, when researchers use a BPL condition. That is, women’s self-reports of sexual activity increase significantly when there is an increased demand for honesty; presumably this occurs because women are influenced by sexual double standards that socially reward men for sexual behavior while punishing women for the same behavior (Alexander & Fisher, 2003; Fisher, 2013). With respect to coercive sexual behaviors, researchers have used the BPL procedure to investigate cognitive distortions held by men who have sexually abused children; the manipulation resulted in increased endorsement of offense-supporting beliefs for men in the BPL condition compared with standard conditions (Ganon, Keown, & Polaschek, 2007).

In the present study, men were randomly assigned to a BPL or Standard Testing (ST) condition and completed two different self-report sexual perpetration measures. We hypothesized that substantially more men in the BPL than the ST condition would endorse having used (a) verbal coercion and (b) sexual assault strategies on the self-report measures. Results in the predicted direction would suggest that intentional concealment may threaten the accuracy of men’s reports of sexual aggression perpetration on traditionally administered surveys.

Method

Participants

Unmarried men, aged 18 to 30 inclusive, who self-identified as sexually attracted to women and who reported some sexual experience (i.e., vaginal, oral, and/or anal intercourse) with a woman were eligible for participation. These inclusion criteria were selected because most sexual aggression occurs within a dating context, and young adults are in the age group at the highest risk of sexual aggression victimization and perpetration (Koss, Dinero, Seibel, & Cox, 1988; Teten, Hall, & Capaldi, 2009). Participants were recruited from the greater St. Louis community. Flyers, Internet ads, and a newspaper ad invited men to participate in a study of “sexual experiences” and offered US$20 for participation.

We sought to recruit approximately 100 men into this study. Consistent with a prior self-report study of men’s perpetration using the SES-LFP (Strang & Peterson, 2013), we anticipated that, in the ST condition, about 28% of men would endorse one or more item on our measures of sexual aggression (including both verbal coercion and sexual assault tactics). We anticipated that, in the BPL condition, men’s reported rates of perpetration would be fairly consistent with women’s reported rates of victimization from a recent study using the revised SES, with about 59% endorsing one or more items (Senn et al., 2014). Given these expected proportions and α = .05, 100 participants would allow us to detect effects with 86% power for the hypothesis-testing analyses (Faul, Erdfelder, Lang, & Buchner, 2007).

A total of 249 men made initial contact with the researcher. Of these, 12 individuals did not meet inclusion requirements, 109 men did not follow up to schedule an appointment, and 30 men no-showed or canceled. Of the 98 men who completed the study, two declined permission to use their data after learning of the deception in this study, one indicated he was above 30 years old, one failed to provide his age, and one was removed due to experimenter error in randomization. Thus, our final sample consisted of 93 men; 35 men were randomized into the ST condition, and 58 men were randomized into the BPL condition using a computer randomization program. Because our groups were—by mere chance—quite uneven in size, we reran our power analysis based on the uneven group sizes; with these group sizes and our expected proportions of endorsement, we still anticipated approximately 80% power to detect effects at α = .05. Because our analyses are slightly underpowered, we report and discuss all effect sizes below. Men in the ST and BPL conditions evidenced no statistically significant differences in demographics, and effect sizes were small to near zero. Demographics are presented in Table 1.

Table 1.

Demographic Information, N = 93.

	M	SD
Age	25.2	3.0
Years of education	14.1	2.4
	n	%
Sexual orientation
Heterosexual	88	94.6
Bisexual	4	4.3
Undecided	1	1.1
Race
White	47	50.5
Black	39	41.9
Biracial/multiracial	7	7.5
Ethnicity
Hispanic	6	6.5
Non-Hispanic	87	93.5
Highest degree earned
Less than high school	6	6.5
High school diploma/GED	49	52.7
Associate’s degree	13	14.0
Bachelor’s degree	20	21.5
Master’s degree	4	4.3
Doctorate degree	1	1.1
Current university/college
Student	26	28.0
Relationship status
Monogamous relationship	28	30.1
Nonmonogamous/dating	44	47.3
Not dating	21	22.6

Note. GED = General Educational Development high school equivalency credential.

Measures

Demographic questionnaire

Participants completed a demographic questionnaire, which collected relevant personal information such as age, race and ethnicity, relationship status, and years of education.

SES-LFP

The SES-LFP (Koss et al., 2007) includes a total of 20 questions assessing sexual teasing and harassment as well as attempted and completed nonconsensual sexual acts ranging from kissing to intercourse. For the analyses presented here, men completed only the SES-LFP items that inquire about completed acts of oral sex, vaginal penetration by a penis or object, and anal penetration by a penis or object. Men were asked about any perpetration since age 14. The published SES-LFP uses gender-neutral language; given the focus of this study, the pronouns were changed to ask only about perpetration by a man against a woman. For each sexual behavior, the SES-LFP assesses for 13 verbal coercion and sexual assault strategies. Notably, the sexual assault strategies on the SES-LFP are designed to closely correspond to most legal definitions of sexual assault.

SSS

The SSS (Strang et al., 2013) asks participants, “In the past, which if any of the following strategies have you used to convince a woman to have sex (oral, anal, or vaginal intercourse) after she initially said ‘no’?” Participants may check any of 22 sexually coercive and aggressive strategies they have used to obtain sex. This scale is a revision and extension of the Postrefusal Persistence Scale developed by Struckman-Johnson, Struckman-Johnson, and Anderson (2003). The strategies examined for the current study included (a) verbal coercion (e.g., “Telling her lies [e.g., saying ‘I love you’ when you don’t]”), (b) use of intoxication (e.g., “Taking advantage of the fact that she is drunk/high”), and (c) threats or force (e.g., “Using physical restraint”). The SSS was designed to assess a wide range of coercive and aggressive strategies; thus, the sexual assault strategies measured by the SSS do not necessarily correspond closely to legal definitions.

Balanced Inventory of Desirable Responding (BIDR)

The BIDR (Paulhus, 1991) measures social desirability with two subscales—Self-Deception (i.e., the tendency to deny psychologically threatening thoughts or feelings) and Impression Management (i.e., the tendency to overreport socially desirable behaviors and underreport socially undesirable behaviors). In prior studies, the BIDR has demonstrated good internal consistency, good test–retest reliability, and good concurrent validity based on its correlations to other measures of social desirability (Paulhus, 1991). In the current study, Cronbach’s alphas for the BIDR were .75 for the Self-Deception subscale and .78 for the Impression Management subscale.

Brief Sexual History Questionnaire (BSHQ)

This series of questions was created for the purposes of the current study. Participants were asked two key questions: (a) “Do you think you may have ever verbally coerced a woman into oral, vaginal, or anal sex?” and (b) “Do you think you may have ever raped or sexually assaulted a woman?” These questions were embedded in a list of several other questions about sexual history (e.g., age at first intercourse, frequency of masturbation, etc.). The questions were included to assess whether participants who endorsed sexually aggressive behavior labeled their behavior as coercive or assaultive.

Filler measures

Two filler measures—one assessing consensual but risky sexual behavior and the other assessing sexual responsivity—were included with the goal of disguising the purpose of the study. The Sexual Risk Survey (SRS; Turchik & Garske, 2009) was designed to measure risky but consensual sexual behaviors in college students, and the Sexual Inhibition/Sexual Excitation–Short Form (Carpenter, Janssen, Graham, Vorst, & Wicherts, 2010) measures individuals’ tendency toward sexual excitation and arousal inhibition.

Posttest Questionnaire

A Posttest Questionnaire, taken from Fisher (2013), inquired about participants’ beliefs about the legitimacy of the BPL equipment. Men were asked to answer the following four questions with a 5-point Likert-type scale with lower scores indicating lower likelihood, lower influence, or lower pressure: (a) “How likely do you think it is that the equipment could be used to assess your anxiety level?”; (b) “How likely do you think it is that the equipment could be used to assess your honesty level?”; (c) “How much influence did the equipment have on your responses to the questions you answered?”; and (d) “How much pressure did you feel from the equipment to answer the questions honestly?”

Procedure

The methods of this study were approved by the Institutional Review Board at the University of Missouri–St. Louis.

Interested prospective participants contacted the researcher through a laboratory email account to receive more information about the study. Men were told that they would be answering questions about sexual experiences and that, during a portion of their visit to the laboratory, they would “be hooked up to a device designed to measure physiological responses.” If interested and eligible, men scheduled an appointment for participation.

Procedures for this study were modeled after those used by Fisher (2013). When men arrived at their individually scheduled appointment, they were greeted by a male researcher. After arrival, participants were randomly assigned to either the BPL (n = 59) or ST (n = 36) conditions. The researcher presented and reviewed an informed consent statement.

BPL condition

In the BPL condition, the researcher positioned the participant at a computer and informed the participant that he would be attached to a physiological measurement device that is “similar to a polygraph or lie detector test.” The participant was told that the machine was being attached to encourage honest responding and that his responses to individual questions may be compared with his physiological output to determine the accuracy of his answers. The researcher then attached functioning electrodes to the participant’s wrist and ankle and a heart monitor to his fingers, and the participant was temporarily able to see his physiological reactivity (heart rate and skin conductance) on the computer screen. The researcher then moved the screen out of participant’s view into the adjacent room. The researcher told the participant to look at the computer screen on the desk in front of him and provide an inaccurate response to the first question on the screen (“Are you in Seattle?”) and a truthful response to the second question on the screen (“Are you in the St. Louis metropolitan area?”) to “calibrate” the machine. The participant was then left alone in the room to complete the questionnaires administered on the computer. The participant’s physiological responses were not actually recorded during the questionnaire completion.

All participants first completed the demographic questionnaire. The order of the SES-LFP, SSS, BIDR, and two filler measures was randomized. Participants completed the BSHQ last. Then, the participants were unhooked from the BPL machine. Finally, participants completed a manipulation check—the Posttest Questionnaire.

Men in the BPL condition were then debriefed. Men were asked if they had any prior knowledge about the nature of the study or any suspicion about the purpose of the device. No participant reported knowledge or suspicion. They were informed about the deception used in the study, as well as the function of this deception. After learning this information, participants had the right to refuse the use of their data. One participant in the BPL condition elected to withdraw his data, leaving 58 men in the BPL condition.

ST condition

In the ST condition, after giving informed consent, the researcher positioned the participant at a computer and informed the participant that he would be attached briefly to a physiological measurement device designed to “determine your level of anxiety prior to starting the questionnaire.” Men in the ST condition were then briefly attached to the same machine used in the BPL condition. This procedure provided a control for the researcher contact involved in hooking up the BPL machine in the experimental condition and is consistent with the procedure used by Fisher (2013). The researcher then moved the physiological measurement screen out of the participant’s view and informed the participant that the machine would take a reading for 1 min. After 1 min, the researcher detached the participant and instructed him to begin the questionnaires on the computer. The researcher then left the participant alone in the room. Men in the ST condition then completed a questionnaire battery identical to the one completed by men in the BPL condition. Participants were debriefed at the conclusion of their participation. Following debriefing, one participant in the ST condition elected to withdraw his data, leaving 35 men in the ST condition.

Results

Data Cleaning

Two measures in the battery—the SES-LFP and the BIDR—had missing values. The SES-LFP is a behavioral sampling measure of coercive and aggressive behavior and does not produce a meaningful total summed or averaged score; therefore, the 15 missing values on the SES-LFP were treated as nonendorsement.¹ No participant was missing more than three values on the measure. Men in the BPL condition produced 12 missing data points (13.8% of men in BPL were missing at least one data point), and men in the ST condition produced three missing data points (8.6% of men in the ST condition were missing at least one data point). Results of Fisher’s exact test suggested that condition (BPL vs. ST) had a small and nonsignificant effect, p > .25, odds ratio (OR) = 0.59, on participants’ likelihood of skipping a sexual aggression item.

Missing values on the BIDR were replaced using mean imputation. There were a total of 14 imputed values, with 10 participants missing one data point out of 40 and two participants missing two data points out of 40. Nine men from the BPL condition and three men from the ST condition had one or more missing values. Results of Fisher’s exact test suggested that condition (BPL vs. ST) had a small and nonsignificant effect, p > .25, OR = 0.51, on participants’ failure to answer BIDR items.

Manipulation Check

Two manipulation checks—a measure of socially desirable responding (BIDR) and questions specifically about the impact of the physiological equipment (Posttest Questionnaire)—were incorporated into the protocol to confirm that the BPL procedure had an impact on men’s reporting in the expected direction. BIDR Self-Deception subscale scores ranged from 0 to 17 (M = 6.44, SD = 3.75), and BIDR Impression Management subscale scores ranged from 0 to 15 (M = 4.64, SD = 3.50) out of a maximum possible score of 20 on each subscale, with higher scores indicating more socially desirable responding. Data were normally distributed, as evaluated by skewness and kurtosis, and required no transformation. There was a moderately strong and significant difference in self-deception scores for men in the BPL condition (M = 5.81, SD = 3.46) versus the ST condition (M = 7.49, SD = 4.02), t = −2.12, p = .04, d = 0.45. There was also a moderately strong and significant difference in Impression Management scores for men in the BPL condition (M = 4.06, SD = 3.20) versus the ST condition (M = 5.60, SD = 3.84), t = −2.08, p = .04, d = 0.44.

Men in both conditions completed the four-item Posttest Questionnaire, in which they reported their beliefs about the legitimacy of the physiological equipment and its effects on their reporting. Results from a MANOVA indicated a moderate and statistically significant difference in reporting between the BPL condition and the ST condition across the four items in the expected direction, Wilks’s λ = 0.78, F(4, 88) = 6.10, p < .001, $η_{p}^{2}$ = 0.22. See Table 2 for descriptive statistics by condition and follow-up univariate tests of significance for each item. Compared with men in the ST condition, men in the BPL condition were significantly more likely to report that the machine “influenced” their responding, F(1, 91) = 17.10, p < .001, $η_{p}^{2}$ = 0.16, and that they felt more “pressure” to respond honestly, F(1, 91) = 13.79, p < .001, $η_{p}^{2}$ = 0.13. Thus, results from both the BIDR and the Posttest Questionnaire provided support for the idea that our BPL manipulation was successful.

Table 2.

Descriptive Statistics and ANOVA Results for Posttest Manipulation Check Items by Condition, N = 93.

Posttest Items	Condition		F	$η_{p}^{2}$
	ST	BPL
	M (SD)	M (SD)
1. How likely do you think it is that the equipment could be used to assess your anxiety level?	3.77 (1.09)	3.59 (1.03)	0.68	0.007
2. How likely do you think it is that the equipment could be used to assess your honesty level?	3.09 (1.27)	3.52 (1.03)	3.21	0.034
3. How much influence did the equipment have on your responses to the questions you answered?	1.37 (0.88)	2.40 (1.30)	17.1***	0.158
4. How much pressure did you feel from the equipment to answer the questions honestly?	1.43 (0.95)	2.43 (1.42)	13.8***	0.132

Note. Scores range from 1 to 5 with lower scores indicating less agreement. ST = Standard Testing; BPL = Bogus Pipeline.

***

p < .001.

Furthermore, for men in the BPL condition, we created a composite variable by averaging the three posttest items measuring (a) the extent to which they felt the equipment could assess their honesty, (b) the extent to which it influenced their responses, and (c) the extent to which they felt pressure to answer honestly (α = .72; M = 2.78; SD = 1.01). For men in the BPL condition, higher scores on this composite measure were significantly associated with the men’s likelihood of endorsing any sexual coercion or aggression item on the SES-LFP and/or the SSS, p = .004, OR = 4.06. This provides support for the conceptual basis of the BPL manipulation; that is, as participants more strongly perceive that their honesty is being monitored, the pressure for honest responding overrides the pressure for socially desirable responding.

Descriptive Statistics

Table 3 presents percentages of endorsement of sexual aggression (including oral, vaginal, and anal sex obtained using verbal coercion and sexual assault tactics) as a function of condition. Endorsement of items on the two measures of sexual aggression (SES-LFP and SSS) is presented both separately and combined (i.e., a man is considered to have endorsed sexual aggression if he endorsed an item on either or both scales) for three distinct tactics: verbal coercion, use of drugs and/or alcohol, and force. Given the low reporting rate for use of force (only six men in total reported use of force on the SES-LFP, and only nine men in total reported use of force on the SSS), as well as the lack of legal distinction between the use of drugs and/or alcohol and force to overcome nonconsent, tactics of using drugs and/or alcohol and force were combined into a composite “sexual assault” variable for the purpose of hypothesis testing.

Table 3.

Percentage of Men Endorsing Sexual Aggression by Condition and Measure, N = 93.

Measure and Tactic	ST Condition (n = 35)	BPL Condition (n = 58)	Fisher’s Exact Test	Odds Ratio
Measure and Tactic	%	%	p	Odds Ratio
SES-LFP and SSS combined
VC tactics	60.0	67.2	.509	1.37
DA tactics	25.7	51.7	.017	3.10
F tactics	8.6	17.2	.357	2.22
SA tactics (DA + F)	28.6	53.4	.030	2.87
SES-LFP
VC tactics	22.9	41.4	.077	2.38
DA tactics	8.6	36.2	.003	6.05
F tactics	5.7	6.9	1.00	1.22
SA tactics (DA + F)	8.6	37.9	.002	6.52
SSS
VC tactics	60.0	58.6	1.00	0.94
DA tactics	25.7	44.8	.079	2.35
F tactics	2.9	13.8	.146	5.44
SA tactics (DA + F)	28.6	46.6	.125	2.18

Note. Items on the SES-LFP and the SSS queried use of these tactics to obtain completed oral, vaginal, or anal sex. ST = Standard Testing; BPL = Bogus Pipeline; SES-LFP = Sexual Experiences Survey–Long Form Perpetration (Koss et al., 2007); SSS = The Sexual Strategies Scale (Strang et al., 2013); VC = verbal coercion; DA = drugs or alcohol; F = force; SA = sexual assault. Bolded text highlights statistically significant differences in reporting as a function of condition.

Overall, men were significantly more likely to report using verbal coercion strategies on the SSS (59.1%) as compared with the SES-LFP (34.4%), p < .001, OR = 6.36. Similarly, men were significantly more likely to report using sexual assault strategies on the SSS (39.8%) compared with the SES-LFP (26.9%), p < .001, OR = 17.06. This pattern of reporting behavior was demonstrated by men in both conditions: Men in the BPL condition were significantly more likely to report using verbal coercion strategies on the SSS (58.6%) compared with the SES-LFP (41.4%), p = .008, OR = 4.81, as well as more likely to report using sexual assault strategies on the SSS (46.6%) as compared with the SES-LFP (37.9%), p < .001, OR = 13.50. Men in the ST condition were significantly more likely to report using verbal coercion strategies on the SSS (60.0%) compared with the SES-LFP (22.9%), p = .009 (OR could not be calculated due to a cell with a 0 value; however, Risk Difference = .48, 95% confidence interval [CI] = [0.33, 0.71]). Men in the ST condition were also significantly more likely to report using sexual assault strategies on the SSS (28.6%) compared with the SES-LFP (8.6%), p = .018, Risk Difference = .22, 95% CI = [0.11, 0.42].

Hypothesis Testing

It was hypothesized that men in the BPL condition would be more likely to acknowledge using sexually aggressive strategies, compared with men in the ST condition. Fisher’s exact tests were used to compare endorsement in the two conditions across sexual tactics (verbal coercion and sexual assault) and across measures (SES-LFP and SSS). Four separate Fisher’s exact tests were performed. The independent variable for each of the Fisher’s exact tests was the condition—BPL or ST. The dependent variables were the dichotomous endorsement or nonendorsement of (a) any verbal coercion item on the SES-LFP, (b) any sexual assault item on the SES-LFP, (c) any verbal coercion item on the SSS, and (d) any sexual assault item on the SSS (see Table 3).

On the SES-LFP, counter to predictions, men in the BPL condition were only slightly and nonsignificantly more likely to report verbal coercion (41.4%) compared with men in the ST condition (22.9%), p = .08, OR = 2.38. Consistent with predictions, men in the BPL condition were substantially more likely to report use of sexual assault tactics (37.9%) than men in the ST condition (8.6%) on the SES-LFP, p = .002, OR = 6.52, such that, on the SES-LFP, the odds of men in the BPL condition reporting sexual assault tactics were 6.52 times greater than the odds of men in the ST condition reporting sexual assault tactics.

On the SSS, contrary to expectations, men in the BPL condition were not significantly more likely to report verbal coercion (58.6%) compared with men in the ST condition (60.0%), and the effect size was near zero, p = 1.00, OR = 0.94. Also contrary to expectations, men in the BPL condition were only slightly and nonsignificantly more likely to report use of sexual assault tactics (46.6%) than men in the ST condition (28.6%) on the SSS, p = .13, OR = 2.18.

Items on the SES-LFP and SSS inquire about sexual aggression tactics in behaviorally specific language that do not contain the terms “coercion,” “rape,” or “sexual assault.” Toward the end of the survey, for exploratory purposes, participants were asked two explicitly worded questions about their sexual aggression history: In response to the question, “Do you think you may have ever verbally coerced a woman into oral, vaginal, or anal sex?” 69.0% of men in the BPL condition and 57.1% of men in the ST condition answered “yes.” This difference in reporting was small and nonsignificant, p > .25, OR = 1.67. Men were also asked, “Do you think you may have ever raped or sexually assaulted a woman?” Three men in the BPL condition answered “yes,” whereas zero men in the ST condition positively endorsed this question. Given the minimal endorsement of this item, no statistical tests were performed.

Exploratory Analyses

Although the questionnaire items related to consensual sexual experiences (e.g., from the BSHQ and the SRS) were originally included to help disguise the purpose of the study, some kinds of consensual sexual behavior have been shown to be associated with sexual aggression: namely, early sexual debut and an impersonal approach to sex, which is reflected in a larger number of partners and more frequent engagement in sex with unknown partners (e.g., Malamuth, Linz, Heavey, Barnes, & Acker, 1995). Thus, to examine whether the underreporting of sexual assault on the SES-LFP in the ST condition might affect the association between sexual aggression and its correlates, we examined the point biserial correlation between endorsement of sexual aggression on the SES-LFP and three relevant consensual sex variables—age at first penile–vaginal intercourse (from the BSHQ), number of lifetime sexual partners (from the SRS), and number of times the man has had sex with someone he did not know well or had just met (from the SRS)—separately for men in the ST and the BPL conditions. We then compared the strength of the correlations for the two groups. Only one correlation differed significantly for the two groups: There was a stronger correlation between age at first intercourse for men in the BPL condition than for men in the ST condition, such that in the BPL condition, but not the ST condition, younger age at first intercourse was associated with greater likelihood of endorsing sexual assault tactics on the SES-LFP. Although there were not statistically significant differences between the correlation coefficients for the other variables, the pattern was such that, the correlations were stronger in the BPL than in the ST condition (see Table 4). This provides some preliminary evidence that sexually aggressive men who are willing to admit to sexual aggression under ST conditions may differ systematically from sexually aggressive men who are unwilling to admit to sexual aggression under ST conditions.

Table 4.

Comparison of Correlations Between Reports of Sexual Assault Tactics on the SES-LFP and Consensual Sexual Behaviors as a Function of Condition.

	ST	BPL	Fisher’s z	p (Two-Tailed)
	r_pb (n)	r_pb (n)	Fisher’s z	p (Two-Tailed)
Age at first intercourse	.15 (34)	−.30^a (58)	2.05	.04
Lifetime sexual partners	−.02 (35)	.33^a (58)	−1.63	.10
Number of sexual encounters with someone you did not know	−.08 (35)	.28^a (57)	−1.65	.10

Note. SES-LFP = Sexual Experiences Survey–Long Form Perpetration; ST = Standard Testing; BPL = Bogus Pipeline; r_pb = point biserial correlation coefficient.

r_pb is significant at p < .05.

Notably, none of these consensual sex variables differed significantly as a function of condition, and effect sizes were near zero, suggesting that self-reports of these consensual sexual behaviors were not influenced by intentional underreporting.

Discussion

Manipulation Checks

Men’s responses to the Posttest Questionnaire supported the success of the BPL manipulation. As expected, men in the BPL condition were moderately more likely to report that the machine “influenced” their responding and caused them to feel more “pressure” to respond honestly than men in the ST condition. Consistent with other studies using the BPL (e.g., Fisher, 2013; Ganon et al., 2007), significant and moderately large differences were also detected between men in the BPL and ST conditions on responses to a measure of socially desirable responding. These findings provide evidence that the BPL manipulation was successful in reducing social desirability and encouraging honest responding.

Primary Findings

Findings from this study revealed that men in the BPL and ST conditions reported equal—and surprisingly high—rates of verbally coercive sexual strategies. Based on this, men appear somewhat comfortable disclosing the use of verbal coercion, suggesting that they may view this behavior as qualitatively distinct from the use of more “severe” sexual assault strategies. The consistency of verbal coercion reports across conditions is, perhaps, not entirely surprising; verbal coercion strategies are not illegal and are somewhat normalized within the context of consensual heterosexual dynamics (Muehlenhard, & Peterson, 2004). Indeed, cultural expectations of male persuasion and pressure in the face of female resistance are often included in traditional sex and seduction scripts (e.g., Littleton, Axsom, & Yoder, 2006; Wiederman, 2005). The fact that 64.5% of the sample, overall, endorsed the use of verbal coercion on at least one measure demonstrates the typicality of these tactics. Furthermore, 64.5% of men answered “yes” to the face-valid question, “Do you think you may have ever verbally coerced a woman into oral, vaginal, or anal sex?” This suggests that most men are fully aware of and willing to label their own verbally coercive behavior. The current results are also consistent with some older findings showing that, unlike with sexual assault and rape, the victimization-perpetration reporting discrepancy for verbal coercion is almost nonexistent; in other words, men report sexually coercing women at nearly the same rate that women report being sexually coerced by men (Spitzberg, 1999).

In contrast to the lack of significant group difference for reports of verbal coercion, a reporting discrepancy for sexual assault tactics across conditions was clearly evident on one measure in this study—the SES-LFP. The SES-LFP is specifically designed to measure illegal sexual behaviors; on that measure, the odds of men admitting to illegal sexual assault were 6.5 times greater in the BPL than in the ST condition. This finding is important to investigate further, given the widespread use of the SES-LFP in sexual aggression perpetration research.

Over a third of men (37.9%) in the BPL condition admitted to perpetrating illegal sexual assault on the SES-LFP. This rate is dramatically higher than in most prior studies using similar samples and measurement tools. For example, Abbey, Jacques-Tiura, and LeBreton (2011) found that 7.2% of men endorsed nonconsensual oral, vaginal, or anal sex through force or intoxication. In Strang et al. (2013), 1.7% of young men reported use of threats or force, and 10.1% reported use of intoxication to obtain oral, anal, or vaginal sex. Notably, reported rates of sexual assault in both the BPL and ST conditions of the current study were higher than reported rates in many other studies. Completing the measures in a lab with a researcher next door may have served to increase honest responding even in the ST condition. It is also worth noting that the percentage of men who endorsed engaging in sexual assault in the BPL condition in the current study was remarkably similar to the percentage of men in the Edwards et al. (2014) study who reported that they would be willing to force a woman to engage in intercourse if they knew that there would be no consequences.

Importantly, there was no significant difference across conditions between rates of reported use of sexual assault tactics on the SSS, and the effect size was fairly small. The SSS was designed to assess a range of sexually aggressive behaviors and to minimize socially desirable responding. The fact that rates of reporting on this measure were less discrepant across conditions as compared with the SES-LFP suggests that it may be successful in eliciting more honest responding. However, the SSS also does not correspond to legal definitions of sexual assault; thus, the assault tactics on the SSS may reflect lower levels of aggression as compared with the sexual assault tactics on the SES-LFP. More research is needed on the SSS, as it is a much less commonly used measure as compared with the SES-LFP.

The current findings suggest that the use of illegal sexual assault tactics may be far more prevalent than is suggested by results from standard self-report procedures. Furthermore, this study provides preliminary evidence that the use of a manipulation designed to promote honest responding can produce increased reports of illegal sexual assault perpetration. There are at least two possible pathways to this result. Most likely, as intended, men in the BPL condition were pressured by the manipulation to endorse items they would have intentionally failed to endorse if they were assigned to the ST condition. However, it also may be the case that men’s differences in reporting across conditions sometimes resulted from more subtle differences in decision-making processes. For example, men in the BPL may have erred on the side of overreporting, whereas men in the ST condition may have erred on the side of underreporting, for experiences they perceived as “gray” or ambiguous. In the future, qualitative follow-up interviews could be used to help elucidate the decision-making process.

Interestingly, although 53.4% of men in the BPL condition endorsed using sexual assault tactics on one or both scales, only three of these men answered affirmatively to the question, “Do you think you may have ever raped or sexually assaulted a woman?” This finding suggests that, even under pressure to answer honestly, men may fail to accurately identify or label their use of sexual assault. This unwillingness and/or inability to recognize that their behavior qualifies as sexual assault may represent a key intervention target.

Unquestionably, underreporting on standard measures of sexual aggression history can affect studies that are designed to assess the prevalence of sexual aggression perpetration. What is less clear is whether underreporting also affects the conclusions drawn from research designed to assess the correlates of sexual aggression. In other words, do sexually aggressive men who are willing to admit to sexual aggression under ST conditions differ systematically from sexually aggressive men who are unwilling to admit to sexual aggression under ST conditions? We used exploratory analyses to examine the association between previously identified (Malamuth et al., 1995; Peterson, Janssen, & Heiman, 2010) correlates of sexual aggression—earlier age at first intercourse, larger number of sexual partners, and more times engaging in sex with an unknown partner—for men in the ST and BPL conditions separately to provide some initial evaluation as to whether underreporting in the ST condition affected the findings regarding correlates of sexual aggression. All three previously identified correlates were significantly associated with sexual assault reports on the SES-LFP in the BPL condition, but none were significant in the ST condition. One correlation—with age at first intercourse—was significantly stronger in the BPL than in the ST condition. This provides some preliminary evidence that underreporting may affect researchers’ conclusions about the correlates of sexual aggression. However, the results also raise additional questions: For example, given that these variables have previously demonstrated a significant relationship with men’s sexual aggression in studies that have used a standard self-report procedure, why did we find a relationship only in the BPL and not in the ST condition? More research is needed to examine how underreporting might influence findings related to correlates of sexual aggression.

Reporting Differences Across Scales

The differences in reporting on the SES-LFP as compared with the SSS are substantial and worth noting. Men were substantially more likely to report both verbal coercion and sexual assault tactics on the SSS than on the SES-LFP. This was true in both the ST and BPL conditions; this suggests that the differences in reporting on the two measures likely were not primarily related to intentional concealment of behaviors on the SES-LFP but rather to unintentional misreporting on one or both measures (i.e., unintentional overreporting on the SSS and/or unintentional underreporting on the SES-LFP).

The different reporting across the measures is likely attributable to language differences between the scales (see Buday & Peterson, 2015; Strang et al., 2013, for further discussion). For example, the SES-LFP uses the phrase “without her consent” when asking about sexual strategies, whereas the SSS asks about strategies used “after she initially said ‘no.’” It is possible that some men who endorsed verbal coercion on the SSS read the more legalistic terminology (i.e., “consent”) on the SES-LFP and dissociated their behavior from these more “serious” or threatening items. Also, as noted above, the SES-LFP sexual assault items were specifically designed to correspond to legal definitions of rape and sexual assault (Koss et al., 2007), whereas this was not necessarily the goal of the SSS sexual assault items. Thus, there are some real and meaningful differences between the two measures. As a result, some discrepancies in reporting across the measures may have reflected legitimate differences in the behaviors being measured, and some discrepancies may have reflected inconsistent responding on the part of the participant. Future research should further investigate how men interpret and respond to items on these two scales, as well as the degree to which researchers and participants possess shared understandings of concepts such as consent, coercion, and aggression.

Limitations

This study involved a relatively small sample size, so results must be considered preliminary. Replication with a larger sample is clearly needed. In particular, such replication could allow for separate analyses of reported use of drug and/alcohol sexual assault tactics versus physical force sexual assault tactics, which may be more appropriate than a collapsed analysis of all sexual assault perpetration given the dramatically different rates of endorsement of these two strategies. Indeed, in this study, reporting differences in the use of sexual assault tactics across conditions on the SES-LFP were driven largely by differences in reporting drug and/or alcohol tactics rather than differences in reporting use of force (see Table 1), potentially suggesting that reports of physical force tactics were less influenced by unintentional concealment than reports of intoxication tactics on the SES-LFP. Yet, on the SSS, the effect size for the cross-condition comparison of reported physical force tactics was quite large, suggesting a potential problem with intentional concealment of force tactics on the SSS. Unfortunately, because the total number of men in our study reporting use of physical force tactics was so low, it is difficult to draw any meaningful conclusions from these results. However, this clearly represents an interesting direction for future research. Nevertheless, despite the small sample size in this study, the effect sizes of the statistically significant analyses are quite large, and the effect sizes of the nonsignificant analyses are generally small, suggesting that the pattern of results is not unduly influenced by a lack of statistical power.

Furthermore, although this study includes a diverse group of community men, which is a strength of the study, participants self-selected into the study. It is possible that participants represent a subset of the community that is particularly comfortable discussing “sexual experiences” or particularly in need of monetary compensation.

Conclusion

Despite limitations, this study provides preliminary evidence that men do intentionally underreport their use of illegal sexual assault tactics under standard self-report conditions, particularly sexual assault by intoxication or incapacitation. Based on findings from men in our BPL condition, men’s use of illegal sexual assault strategies may be far more common than is suggested by most self-report studies. Thus, men’s underreporting of sexual aggression may help to explain the paradoxical victim–perpetrator reporting discrepancy. These findings are in contrast to the assumption that all sexual assault is perpetrated by a few repeat offenders. Results suggest that broad social change is likely needed to address the widespread problem of men’s sexual aggression.

Footnotes

Acknowledgements

The authors are grateful to William Dement, Joe Davis, and Patrick Graham for their help with data collection and to Drs. Kristen Carbone-Lopez, Ann Steffen, and Brian Vandenberg for their input on the project.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was conducted at the University of Missouri–St. Louis as part of Emily Strang’s dissertation under the supervision of Zoë Peterson, with funding from the Jayne E. Stake Feminist Research Award.

Notes

Author Biographies

Emily Strang, PhD, is a psychologist with the Southern Arizona VA Healthcare System in Tucson, Arizona. She was trained in clinical psychology and earned her PhD from the University of Missouri–St. Louis in 2015. She currently works primarily with veterans experiencing posttraumatic stress disorder (PTSD) secondary to combat and military sexual trauma. Her research focuses primarily on sexual assault perpetration and measurement validity.

Zoë D. Peterson, PhD, is an associate professor of psychological sciences and director of the Sexual Assault Research and Education Program at the University of Missouri–St. Louis. She is a research fellow of the Kinsey Institute for Research in Sex, Gender, and Reproduction at Indiana University. She also serves as an associate editor of the Journal of Sex Research. Her research focuses primarily on individuals’ experiences with and conceptualizations of sexual coercion, sexual aggression, and unwanted sex. Her research has been funded by the Social Science Research Council’s Sexuality Research Fellowship Program and by the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

References

Abbey

Jacques-Tiura

A. J.

LeBreton

(2011). Risk factors for sexual aggression in young men: An expansion of the confluence model. Aggressive Behavior, 37, 450-464. doi:10.1002/ab.20399

Alexander

M. G.

Fisher

T. D.

(2003). Truth and consequences: Using the Bogus Pipeline to examine sex differences in self-reported sexuality. The Journal of Sex Research, 40, 27-35. doi:10.1080/00224490309552164

Buday

Peterson

Z. D.

(2015). Men’s and women’s interpretation and endorsement of items measuring self-reported sexual aggression. Journal of Sex Research, 52, 1042-1053. doi:10.1080/00224499.2014.967373

Carpenter

D. L.

Janssen

Graham

C. A.

Vorst

Wicherts

(2010). The Sexual Inhibition/Sexual Excitation Scales–Short form. In Fisher

T. D.

Davis

C. M.

Yarber

W. L.

Davis

S. L.

(Eds.), Handbook of sexuality-related measures (3rd ed., pp. 236-240). New York, NY: Routledge.

Carroll

M. H.

Clark

M. D.

(2006). Men’s acquaintance rape scripts: A comparison between a regional university and a military academy. Sex Roles, 55, 469-480. doi:10.1007/s11199-006-9102-3

Classen

C. C.

Palesh

O. G.

Aggarwal

(2005). Sexual revictimization: A review of the empirical literature. Trauma, Violence, & Abuse, 6, 103-129. doi:10.1177/1524838005275087

Cook

S. L.

(2002). Self-reports of sexual, physical, and nonphysical abuse perpetration. Violence Against Women, 8, 541-565. doi:10.1177/107780102400388425

Daigle

L. E.

Fisher

B. S.

Cullen

F. T.

(2008). The violent and sexual victimization of college women: Is repeat victimization a problem? Journal of Interpersonal Violence, 23, 1296-1313. doi: 10.1177/0886260508314293

Edwards

S. R.

Bradshaw

K. A.

Hinsz

V. B.

(2014). Denying rape but endorsing forceful intercourse: Exploring differences among responders. Violence and Gender, 1, 188-193. doi:10.1089/vio.2014.0022

10.

Faul

Erdfelder

Lang

A. G.

Buchner

(2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175-191.

11.

Fisher

T. D.

(2013). Gender roles and pressure to be truthful: The Bogus Pipeline modifies gender differences in sexual but not non-sexual behavior. Sex Roles, 68, 401-414. doi:10.1007/s11199-013-0266-3

12.

Ganon

T. A.

Keown

Polaschek

D. L. L.

(2007). Increasing honest responding on cognitive distortions in child molesters: The Bogus Pipeline revisited. Sex Abuse, 19, 5-22. doi:10.1007/s11194-006-9033-0

13.

Heilbrun

A. B.

Loftus

M. P.

(1986). The role of sadism and peer pressure in the sexual aggression of male college students. Journal of Sex Research, 22, 320-332. doi:10.1080/00224498609551312

14.

Jones

E. E.

Sigall

(1971). The Bogus Pipeline: A new paradigm for measuring affect and attitude. Psychological Bulletin, 76, 349-364. doi:10.1037/h0031617

15.

Kolivas

E. D.

Gross

A. M.

(2007). Assessing sexual aggression: Addressing the gap between rape victimization and perpetration prevalence rates. Aggression and Violent Behavior, 12, 315-328. doi:10.1016/j.avb.2006.10.002

16.

Koss

M. P.

Abbey

Campbell

Cook

Norris

Testa

. . . White

(2007). Revising the SES: A collaborative process to improve assessment of sexual aggression and victimization. Psychology of Women Quarterly, 31, 357-370. doi:10.1111/j.1471-6402.2007.00385.x

17.

Koss

M. P.

Dinero

T. E.

Seibel

C. A.

Cox

S. L.

(1988). Stranger and acquaintance rape: Are there differences in the victim’s experience? Psychology of Women Quarterly, 21, 1-24. doi:10.1111/j.1471-6402.1988.tb00924.x

18.

Koss

M. P.

Gidycz

C. A.

Wisniewski

(1987). The scope of rape: Incidence and prevalence of sexual aggression and victimization in a national sample of higher education students. Journal of Consulting and Clinical Psychology, 55, 162-170. doi:10.1037/0022-006X.55.2.162

19.

Koss

M. P.

Oros

C. J.

(1982). Sexual Experiences Survey: A research instrument investigating sexual aggression and victimization. Journal of Consulting and Clinical Psychology, 50, 455-457. doi:10.1037/0022-006X.50.3.455

20.

Lisak

Miller

P. M.

(2002). Repeat rape and multiple offending among undetected rapists. Violence and Victims, 17, 73-84. doi:10.1891/vivi.17.1.73.33638

21.

Littleton

H. L.

Axsom

Yoder

(2006). Priming of consensual and nonconsensual sexual scripts: An experimental test of the role of scripts in rape attributions. Sex Roles, 54, 557-563. doi:10.1007/s11199-006-9017-z

22.

Malamuth

N. M.

Linz

Heavey

C. L.

Barnes

Acker

(1995). Using the confluence model of sexual aggression to predict men’s conflict with women: A 10-year follow-up study. Journal of Personality and Social Psychology, 69, 353-369. doi:10.1037/0022-3514.69.2.353

23.

Meston

C. M.

Heiman

J. R.

Trapnell

P. D.

Paulhus

D. L.

(1998). Socially desirable responding and sexuality self-reports. The Journal of Sex Research, 35, 148-157. doi:10.1080/00224499809551928

24.

Muehlenhard

C. L.

Peterson

Z. D.

(2004). Conceptualizing sexual violence: Socially acceptable coercion and other controversies. In Miller

A. G.

(Ed.), The social psychology of good and evil (pp. 240-268). New York, NY: Guilford.

25.

Ouimette

P. C.

Shaw

Drozd

J. F.

Leader

(2000). Consistency of reports of rape behaviors among nonincarcerated men. Psychology of Men & Masculinity, 1, 133-139. doi:10.1037//1524-9220.1.2.133

26.

Paulhus

D. L.

(1991). Measurement and control of response bias. In Robinson

J. P.

Shaver

P. R.

Wrightsman

L. S.

(Eds.), Measures of personality and social psychology attitudes (Vol. 1, pp. 17-59). San Diego, CA: Academic Press.

27.

Peterson

Z. D.

Janssen

Heiman

J. R.

(2010). The association between sexual aggression and HIV risk behavior in heterosexual men. Journal of Interpersonal Violence, 25, 538-556. doi:10.1177/0886260509334414

28.

Porter

J. F.

Critelli

J. W.

Tang

C. S. K.

(1992). Sexual and aggressive motives in sexually aggressive college males. Archives of Sexual Behavior, 21, 457-468. doi:10.1007/BF01542270

29.

Roese

N. J.

Jamieson

D. W.

(1993). Twenty years of Bogus Pipeline research: A critical review and meta-analysis. Psychological Bulletin, 114, 363-375. Retrieved from http://www.apa.org/pubs/journals/bul/index.aspx

30.

Ross

R. R.

Allgeier

E. R.

(1996). Behind the pencil/paper measurement of sexual coercion: Interview-based clarification of men’s interpretations of sexual experiences survey items. Journal of Applied Social Psychology, 26, 1587-1616. doi:10.1111/j.1559-1816.1996.tb00087.x

31.

Senn

C. Y.

Eliasziw

Barata

P. C.

Thurston

W. E.

Newby-Clark

I. R.

Radtke

H. L.

, . . . SARE Study Team. (2014). Sexual violence in the lives of first-year university women in Canada: No improvements in the 21st century. BMC Women’s Health, 14, Article 135. Retrieved from http://bmcwomenshealth.biomedcentral.com/articles/10.1186/s12905-014-0135-4

32.

Spitzberg

B. H.

(1999). An analysis of empirical estimates of sexual aggression victimization and perpetration. Violence and Victims, 14, 241-260. doi:10.1177/088626099014001001

33.

Strang

Peterson

Z. D.

(2012). The relationships among perceived peer acceptance of sexual aggression, punishment certainty, and sexually aggressive behavior. Journal of Interpersonal Violence, 28, 3369-3385. doi:10.1177/0886260513502126

34.

Strang

Peterson

Z. D.

(2013). The relationships among perceived peer acceptance of sexual aggression, punishment certainty, and sexually aggressive behavior. Journal of Interpersonal Violence, 28, 3369-3385. doi:10.1177/0886260513502126

35.

Strang

Peterson

Z. D.

(2016). Unintentional misreporting on self-report measures of sexually aggressive behavior: An interview study. Manuscript submitted for publication.

36.

Strang

Peterson

Z. D.

Hill

Heiman

J. R.

(2013). Discrepant responding across self-report measures of men’s coercive and aggressive sexual strategies. Journal of Sex Research, 50, 458-469. doi:10.1080/00224499.2011.646393

37.

Struckman-Johnson

Anderson

P. B.

(2003). Tactics of sexual coercion: When men and women won’t take no for an answer. Journal of Sex Research, 40, 76-86. doi:10.1080/00224490309552168

38.

Teten

A. L.

Hall

G. C. N.

Capaldi

D. M.

(2009). Use of coercive sexual tactics across 10 years in at-risk young men: Developmental patterns and co-occurring problematic dating behaviors. Archives of Sexual Behavior, 38, 574-582. doi:10.1007/s10508-007-9309-6

39.

Tourangeau

Smith

T. W.

Rasinski

K. A.

(1997). Motivation to report sensitive behaviors on surveys: Evidence from a Bogus Pipeline experiment. Journal of Applied Social Psychology, 27, 209-222. doi:10.1111/j.1559-1816.1997.tb00629.x

40.

Turchik

J. A.

Garske

J. P.

(2009). Measurement of sexual risk taking among college students. Archives of Sexual Behavior, 38, 936-948. doi:10.1007/s10508-008-9388-z

41.

Walker

W. D.

Rowe

R. C.

Quinsey

V. L.

(1993). Authoritarianism and sexual aggression. Journal of Personality and Social Psychology, 65, 1036-1045. doi:10.1037/0022-3514.65.5.1036

42.

Wiederman

M. W.

(2005). The gendered nature of sexual scripts. The Family Journal, 13, 496-502. doi:10.1177/1066480705278729