Introduction

The concern over rates of sexual coercion and date rape at universities and colleges has led many of these higher institutions to set up awareness programs focused on sexual communication and behavior. Campus programs are designed to educate students about sexual violence, and as in legal cases, they typically stress the importance of asking and giving sexual consent. Although sexual consent is of central importance in cases of sexual assault, there has been only limited research on this topic. Without a solid understanding of how sexual consent is negotiated and communicated, the effectiveness of awareness programs is questionable. The purpose of this study was the development of a survey measure to enhance our understanding of the attitudes and behaviors of university students regarding sexual consent and to test some theoretical predictions based on gender and relationship history, to provide initial validation of the new measure.

Given the dearth of research on sexual consent itself, research in the related areas of sexual communication (Hickman and Muehlenhard 1999; O’Sullivan and Byers 1992; Quina et al. 2000; Sawyer et al. 1993), safer sex negotiations (Allen et al. 2002; Lewis et al. 1997; Molitor et al. 1999) and sexual coercion, especially date rape (Adams-Curtis and Forbes 2004; Byers and O’Sullivan 1996; Grauerholz and Koralewski 1991) helps to inform the study of sexual consent. Sexual consent is important for several reasons. Consent is the defining criterion for both legal and research definitions of what constitutes sexual assault, rape, and abuse. Most importantly, the interpretation of sexual consent in legal cases affects judgments about the guilt or innocence of the accused. Also, research definitions of sexual consent influence the construction of measurement instruments and ultimately the prevalence estimates for sexual violence which researchers find in the population (Muehlenhard et al. 1992).

Defining Sexual Consent

Muehlenhard (1996) and Lim and Roloff (1999) have analyzed the key components of sexual consent. Firstly, sexual consent requires knowledge. An individual must have a clear understanding of what she or he is consenting to, before consent can be considered legitimate. This requires knowledge about what the other person is expecting in terms of sexual behavior. Secondly, sexual consent is meaningless unless given freely, which means being free of coercion or undue influence.

One of the key questions in this area of research is whether consent is a cognitive or behavioral act (Muehlenhard 1996). If consent occurs by means of a cognitive decision, then a partner cannot ‘know’ with any certainty that consent has been given. In this case, an individual’s consent is inferred from his or her nonverbal behaviors (Lim and Roloff 1999). Nonverbal behaviors can be difficult to read accurately and men have a tendency to interpret nonverbal cues from women more sexually than women intend them (Abbey 1982, 1991; Koukounas and Letch 2001). If consent occurs through verbal articulation, then consent should be clearly understood; however, most sexual encounters involve few explicit verbal statements (Greer and Buss 1994; O’Sullivan and Byers 1992; Sawyer et al. 1993). What tends to occur is a tacit agreement to engage in sexual behavior.

Sexual Script Theory

Sexual scripts are cognitive frameworks learned through socialization that delineate how people are expected to behave within sexual interactions (Byers 1996). According to Simon and Gagnon (1973), scripting occurs on multiple levels: cultural, interpersonal and intrapsychic. The cultural level of scripting contains sexual scripts commonly understood among members of a society, community or group. For example, in Western society, there is a common understanding that sexual consent is required for sexual activity to take place; however, how this actually occurs is in the domain of interpersonal and intrapsychic levels of scripting. Other cultural scripts, such as that of women concealing their genuine interest in sexual activity for fear of being perceived as promiscuous and that of men interpreting women’s refusals as only token gestures (Muehlenhard and Hollabaugh 1988; Sprecher et al. 1994) can seriously hamper how sexual consent is negotiated. Interpersonal scripts represent a compromise position between the incoming cultural messages and the individual’s private understandings. Scripts are modified to meet the expectations of others during social interactions. In other words, the context influences the form that scripts take when people are interacting. Given that the context is important for the development of scripts, factors such as the length of a relationship and past sexual history may be influential when negotiating sexual consent. Intrapsychic scripts represent the individual’s private cognitive world. This level of scripting occurs when individuals take into account cultural notions and social interaction to derive scripts unique to themselves. Although described separately for clarity, these three levels of scripting are dynamically integrated. This involves a continuous process of interpreting and reinterpreting messages and actions from self and others (Gagnon 1990).

Differences in gender socialization in North America suggest that the issue of sexual consent may be more important to women than to men. If women are socialized to be the limit-setters of relationships, then part of their traditional sexual script is the giving of consent to engage in sexual activity. It is expected that women, more than men, should decide whether sexual activity will proceed. Men, on the other hand, have been socialized to seek sexual involvement at every opportunity, which suggests that men are more likely to ask for consent as part of their sexual script. However, there is also a competing cultural script for men suggesting they should ignore women’s resistance to their initiations because it is merely a token gesture (Osman 2003; Sprecher et al. 1994).

At its core, sexual script theory emphasizes that sexual attitudes and behaviors are derived from our interaction with the culture that socializes us. The research cited above, describing the particular intricacies of sexual consent, comes from the United States and Canada. Therefore the conceptualization of sexual consent in this research is situated in a North American cultural context and is not believed to differ significantly between Canada and the United States.

Research on Sexual Consent

Research into the process by which individuals negotiate sexual consent is relatively new. Only three studies have focused specifically on sexual consent (Beres et al. 2004; Hall 1998; Hickman and Muehlenhard 1999).

Hall (1998) examined the giving of sexual consent in day-to-day sexual interactions involving both coital and non-coital sexual behaviors. In his sample of 310 sexually active heterosexual college students, Hall found that sexual behaviors usually occurred without overt consent being given. Permission giving was reported most frequently when the behavior in question was vaginal or anal intercourse. The rate of permission giving for other sexual behaviors (e.g., kissing, touching) was markedly lower. In addition, consent giving was most often nonverbal in nature, except for intercourse, for which consent was verbal about half of the time. Hall’s research suggests that there is a sliding scale of priority when it comes to making certain consent has been given. Sexual behaviors that are deemed more intimate are more likely to lead individuals to overtly negotiate consent.

Hickman and Muehlenhard (1999) investigated how university students inferred and conveyed consent to engage in penile-vaginal intercourse. Using scenarios, 378 heterosexual participants were presented with situations in which sex was initiated verbally or nonverbally, along with 34 possible responses to the initiation. Participants indicated how likely they would be to use each of the responses, as well as how frequently they used each response in actual situations. The findings indicated that there is a wide range of behavioral approaches used to signal sexual consent to a partner. Factor analysis revealed that sexual consent behaviors could be categorized as direct or indirect and verbal or nonverbal. Direct consent signals were defined as straightforward and unambiguous tactics (e.g., stating, “I want to have sex with you”) whereas indirect consent signals were more veiled and ambiguous tactics (e.g., “she/he touches and kisses you”).

Beres et al. (2004) analyzed verbal and nonverbal sexual consent behaviors in same-sex relationships. Participants completed online questionnaires assessing the extent to which they had used 26 verbal and nonverbal sexual consent behaviors to initiate or respond to sexual activity. For both men who had sex with men (MSM) and women who had sex with women (WSW), the findings indicated that nonverbal behaviors are used more often than verbal behaviors both to ask for and to give sexual consent.

Although these studies have provided important insights regarding the sexual consent process, especially with respect to the categorization of consent signals, there are still significant gaps in our knowledge. Hickman and Muehlenhard (1999) focused only on consent for penile-vaginal intercourse; however, consent signals can be used for a range of behaviors. Hall’s (1998) research suggested that there may be differences in consent based on the type of sexual behavior being negotiated. By focusing only on intercourse, the process of consent cannot be fully understood. Researchers also have not examined whether consent is situation specific. It is possible that inferring and conveying consent on a first date is very different than in a long-term relationship. For example, a smile in response to being asked “do you have a condom?” could indicate consent giving in an established relationship, but might indicate nervous apprehension on a first date. The context of the relationship and the behavior requiring consent both need to be taken into consideration when assessing sexual consent signals. Finally, we do not know the meaning that young women and men attach to sexual consent. Do they perceive it as an important aspect of relationships?

The Present Study

The goal of the present study was to develop a new scale to measure sexual consent attitudes and behaviors. The study was designed to be more comprehensive than previous ones in that it also incorporated relationship and contextual variables that influence the negotiation of consent. Additionally, we considered attitudes toward universities having formal requirements for obtaining sexual consent, such as the Antioch University Policy of requiring consent at each stage of a sexual encounter. These findings were presented in an earlier publication (Humphreys and Herold 2003).

Principle components factor analysis was completed on the new sexual consent measure. The goal of this analysis was to investigate the latent structure of the factors for the items. To assess the construct validity of the new sexual consent measure, a number of theoretical predictions were developed. Following from previous research on sexual consent (Beres et al. 2004; Hall 1998; Hickman and Muehlenhard 1999), it was hypothesized that there would be gender differences in attitudes and behaviors about sexual consent, with women placing more importance on the explicit use of sexual consent than men. The gender difference on consent can act as a preliminary estimate of the validity of the new scales.

Secondly, research on sexual precedence has suggested that having previously had sex with a partner is perceived as an obligation to continue sexual activity on subsequent occasions (Monson et al. 2000; Shotland and Goodstein 1992). This obligation, based on experience, should reduce the need for explicit consent negotiations between couples. Therefore, sexual and relationship experience variables should be related to sexual consent beliefs and practices in predictable ways. If we can predict scores on the sexual consent measure based on gender (as mentioned above) and sexual and relationship experience variables, then we will have further strengthened the construct validity of the newly developed scales. Multiple regression was used to assess the contributions of gender, intercourse experience, number of intercourse partners, and relationship status to the variance in sexual consent attitudes and behaviors. It was hypothesized that greater sexual and relationship experience will result in less importance being placed on the explicit use or negotiation of sexual consent.

Finally, we explored the ability of the sexual consent attitude measures to predict the sexual consent behavior measures and vice versa. This analysis was exploratory given that there is currently no theoretical or empirical support for a causal direction.

Method

Focus group interviews were used to gain an initial understanding of the key themes in order to generate items for the quantitative survey. Two female groups (N = 7, 5) and one male group (N = 6) consisting of heterosexual students unknown to each other, participated in the focus group interviews with a same-sex facilitator. Participants ranged in age from 19 to 43 (M = 23.7, Mode = 21). Use of focus groups prior to developing the survey instruments improved the phrasing and relevance of the items as well as ensuring adequate coverage of diverse aspects of the topic. The main questions asked of focus group participants included: (1) their own understanding of sexual consent, (2) whether sexual consent was an issue being discussed among their friends or with their partners, (3) the signals they used to ask for or give sexual consent to a partner as well as how often this occurs in a typical sexual encounter, (4) whether they thought that women and men differed in their sexual consent perceptions or behaviors, and (5) how the length of a relationship affects consent negotiations.

The descriptive themes from the qualitative data were subsequently used in the development of the survey. Themes were translated into one or more Likert-type questions which represented core aspects of that theme. For example, when asked how often verbal consent is obtained during a typical encounter, focus group participants frequently talked about the length or stage of a relationship as being a determining factor. This theme of relationship context lead to the development of survey items such as “Obtaining sexual consent is more necessary in a new relationship than in a committed relationship” and “The necessity of asking for sexual consent decreases as the length of an intimate relationship increases”. In addition, discussion regarding the general issue of context led to the exploration of how consent may vary depending on which sexual behaviors were the focus. Accordingly, survey items were developed to assess how consent was negotiated for different sexual behaviors. For example, survey items such as “sexual intercourse is the only sexual activity that requires explicit verbal consent” and “consent should be asked before any kind of sexual behavior, including necking and petting” were developed to explore this theme.

Survey Participants

A stratified random sample of 1,200 students from the undergraduate population at a Canadian university was mailed a questionnaire regarding attitudes and behaviors toward sexual consent. At this university, about two-thirds of the students are female. We purposely over-sampled males by sending questionnaires to an equal number of males and females to ensure sufficient numbers of males. Five hundred and fourteen usable questionnaires were returned, giving an overall response rate of 43%. The response rate for women was notably higher (330 of 600 or 55%), whereas the rate for men was 184 of 600 (31%). The gender distribution of the sample was 64.2% female and 35.8% male. Mean age was 20.8 years of age (SD = 1.58). Participants were selected for inclusion in the analysis if they were full-time, 18 to 27 years old, heterosexual, and not married.

The majority of students (93%) had experienced consensual forms of non-coital sexual activity such as petting or sexual touching with a member of the other sex (93% female; 92% male). The majority (75%) had also experienced consensual sexual intercourse (73% female; 77% male). The reported number of intercourse partners ranged between 0 and 35 with a mean of 2.8 and a median of 2.0 (SD = 4.2) (for females, range 0–22, M = 2.7, SD = 3.7; for males, range 0–35, M = 3.3, SD = 5.2).

Measure

Sexual Experience

Sexual experience was measured by asking: (1) “Have you ever willingly engaged in mild forms of sexual activity such as petting or sexual touching with a member of the opposite sex?”, (2) “Have you ever willingly engaged in sexual intercourse (that is, penile–vaginal intercourse)?”, and (3) “If yes, with how many partners have you had penile–vaginal sexual intercourse?” The response categories for the first two questions were “yes” and “no”. Participants were asked to provide a numerical response to question three. Question two above was used to assess sexual intercourse experience in subsequent analyses.

Sexual Consent Scales

Based on focus group responses, two sexual consent scales were constructed: the Sexual Consent Attitudes Scale and the Sexual Consent Behaviors Scale. The Sexual Consent Attitudes Scale was a 23-item scale measuring beliefs and attitudes toward sexual consent. Example items include: “When initiating sexual activity, it is okay to assume consent and proceed sexually until the partner indicates no”, and “Sexual consent should always be obtained BEFORE the start of any sexual activity”. Response options ranged from 1 (Strongly disagree) to 7 (Strongly agree). The Sexual Consent Behaviors Scale was a 12-item scale measuring students’ sexual consent behaviors. Example items include: “I have discussed sexual consent issues with a friend” and “During a sexual encounter, I typically only ask for consent once”. Response options ranged from 1 (Strongly disagree) to 7 (Strongly agree).

Preferred Method of Obtaining Consent

A single forced-choice item was used to measure preference for one of two ways of obtaining sexual consent. The question instructed students to “Check which of these two statements you agree with more.” The two response choices were: “In making sexual advances, it is okay to continue until a partner indicates otherwise (i.e., assume ‘yes’ until you hear a ‘no’),” and “BEFORE making sexual advances, one should always ask for and obtain a verbal ‘yes’ to engage in any sexual activities (i.e., assume ‘no’ until you get a ‘yes’).”

Procedure

Twelve hundred questionnaires were delivered by campus mail services to a stratified random sample of students’ on-campus mail boxes. A reminder card was sent to the survey participants two weeks following initial mailing to thank those who had responded and to encourage others to do so.

Principle Components Analysis

Principal components factor analyses (PCA) with varimax rotation were conducted separately for the Sexual Consent Attitudes Scale and the Sexual Consent Behaviors Scale to reduce the number of items to a smaller set of more meaningful dimensions. In order to determine the ideal number of components to be extracted, two guidelines were used: only components having eigenvalues (proportion of variance explained) of greater than one were considered to be significant (Hair et al. 1987) and the scree plot was examined for the point at which the curve began to straighten out. The conceptual fit of the variables loading onto the component was also considered. A label was assigned to the resulting subscale to reflect, to the greatest extent possible, what the items loading onto it represented (Hair et al. 1987). The reliability of each subscale was assessed with Cronbach’s alpha coefficient.

PCA of the Sexual Consent Attitudes Scale

An examination of the scree plot for the PCA of the 23 items measuring sexual consent attitudes revealed a two component solution (see Table 1 for factor loadings). The model converged in three iterations and the Kaiser-Meyer-Olkin measure (.88) suggested that the distribution of values was adequate to conduct a PCA. Four items were removed from the solution because they did not load on any component at .3 or higher, cross-loaded on more than one component at .35 or higher, or did not fit conceptually. The two resulting subscales for the Sexual Consent Attitudes Scale were labeled (1) Asking for Consent First is Important; and (2) Commitment Reduces Asking for Consent. The communalities indicated that between 14 and 70% of the variance in each item was accounted for by the resulting subscales. Subscale scores were computed by summing and dividing by the number of items. Some items were reverse scored, as noted in Table 1. Higher scores represent stronger agreement with the factor.

Table 1 Item loadings for the sexual consent attitudes scale.
Full size table

The model accounted for 37% of the variance. The Asking for Consent First is Important subscale (α = .81) accounted for 27.6% of the variance and the Commitment Reduces Asking for Consent subscale (α = .78) accounted for 9.5% of the variance. Internal consistency for the entire Sexual Consent Attitudes Scale was .85 (women α = .83 and men α = .84).

PCA of the Sexual Consent Behaviors Scale

An examination of the scree plot for the PCA of the 12 items measuring sexual consent behaviors revealed a two component solution (see Table 2 for item loadings). The model converged in three iterations and the Kaiser-Meyer-Olkin measure (.66) suggested that the distribution of values was adequate to conduct a PCA. Three items were removed from the solution because they did not load on any component at .3 or higher, cross-loaded on more than one component at .35 or higher, or did not fit conceptually. The two resulting subscales for the Sexual Consent Behaviors Scale were labeled (1) Consent Discussions/Awareness; and (2) Consent is Negotiated Once. The communalities indicated that between 25 and 69% of the variance in each item was accounted for by the resulting subscales. Subscale scores were computed by summing and dividing by the number of items. Some items were reverse scored, as noted in Table 2. Higher scores represent greater agreement that the factor correspond to their personal behavior.

Table 2 Item loadings for the sexual consent behaviors scale.
Full size table

The model accounted for 46% of the variance. The Consent Discussions/Awareness subscale (α = .70) accounted for 28% of the variance and the Consent is Negotiated Once subscale (α = .56) accounted for 18% of the variance. Internal consistency for the entire Sexual Consent Behavior Scale was 0.66 (women α = .69 and men α = .60).

In summary, the two principal components factor analyses resulted in four subscales representing significant dimensions of sexual consent. Three of the four variables had moderate to high reliability. The variable with a lower alpha coefficient measured an important aspect of sexual consent and therefore was included in subsequent analyses. The principal components analysis was performed on women and men separately and produced an almost identical structure to the combined structure presented here. Therefore the structure seems to reflect the responses of both women and men equally.

Results

Gender Analyses

A between subjects multivariate analysis of variance was performed on the 4 sexual consent factors as dependent measures. The independent variable was participant’s gender (male and female). Results of homogeneity of variance-covariance matrices were satisfactory.

A significant multivariate effect was found for gender, Wilks’ ë = .86, and F(4, 502) = 19.81, p < .001. The results reflected a modest association between the combined dependent variables and gender scores, \( {\mathop \eta \limits^\prime }^{2} \) = .14. At the univariate level, three of the four sexual consent factors showed significant differences on the basis of gender. Women (M = 5.07, SD = .72) agreed more than men (M = 4.49, SD = .80) that asking for sexual consent prior to the beginning of sexual activity was important, F(1, 507) = 70.66, p < .001, \( {\mathop \eta \limits^\prime }^{2} \) = .12. Men (M = 4.60, SD  = 1.16) agreed more than women (M = 4.33, SD  = 1.24) that committed or longer relationships reduced the need for consent negotiations, F(1, 507) = 5.64, p = 0.015, \( {\mathop \eta \limits^\prime }^{2} \) = .01. Men (M = 4.54, SD = .99) also agreed more than women (M = 4.09, SD = 1.07) that their own sexual consent behaviors represent a single event that happens “in the moment”, F(1, 507) = 21.43, p < .001, \( {\mathop \eta \limits^\prime }^{2} \) z= 0.04. Women (M = 4.38, SD = 1.38) and men (M = 4.22, SD  =  1.34) did not differ on their general awareness or discussions of consent with friends or partners. These findings support the hypothesis that women place more importance on sexual consent than men.

Preferred Method of Consent Negotiation

More students preferred to ask for consent first (60%) than to assume it (39%) before engaging in sexual activity. Women (65%) were more likely than men (53%) to prefer the method of obtaining consent which involved asking prior to engaging in any sexual activity, whereas men (47%) were more likely than women (35%) to prefer assuming consent and continuing with sexual activity until the partner indicates otherwise, χ 2 (1, N = 507) = 6.62, p < .01. This finding also provides support for the hypothesis that women place more importance on sexual consent than men.

Predicting Sexual Consent Attitudes and Behaviors

A standard multiple regression analysis was performed for each of the sexual consent attitude subscales and the sexual consent behavior subscales. Table 3 displays the correlations between the variables. The predictor variables in the regressions on sexual consent attitudes were gender, sexual intercourse experience, current relationship status, number of sexual intercourse partners, and the two sexual consent behavior subscales (see Table 4). The predictor variables in the regressions on sexual consent behaviors included gender, sexual intercourse experience, relationship status, number of partners and the two sexual consent attitudes subscales (see Table 5). The tolerances for each of the predictors were near 1.00, suggesting that multicollinearity was not a concern in these analyses.

Table 3 Correlations of predictor variables and sexual consent.
Full size table
Table 4 Predictors of sexual consent attitudes.
Full size table
Table 5 Predictors of sexual consent behaviors.
Full size table

In the regression models for sexual consent attitudes (see Table 4), the most consistently significant predictors were sexual intercourse experience and the two sexual consent behavior subscales. They were the only variables significant across both sexual consent attitude subscales. For Asking for Consent First is Important, gender, sexual intercourse experience, current relationship status, number of sexual partners, consent discussions/awareness, and consent is negotiated once were statistically significant unique predictors, indicating that being female, not having experienced sexual intercourse, being in a committed relationship, having fewer partners, having more general discussions about consent, and having more of a process orientation to consent were related to a stronger belief in establishing consent prior to sexual activity beginning. For Commitment Reduces Asking for Consent, having experienced sexual intercourse, having fewer consent discussions, and agreeing that one’s own consent behavior is a one-time event were related to the belief that having a committed relationship reduces the need for sexual consent negotiations.

For the Behavioral Consent Scale (see Table 5), the only significant predictor of the subscale, Consent Discussions/Awareness was the importance of asking for consent first. In other words, the more individuals stressed the establishment of consent prior to sexual activity, the greater their awareness and discussions about consent. For Consent is Negotiated Once, the significant predictors were both attitudinal subscales, gender and sexual experience. In other words, placing less emphasis on consent prior to sexual activity, believing that commitment reduces the need for sexual consent, being male and having experienced sexual intercourse predicted a behavioral approach to sexual consent that was more of a one-time, in the moment, event and less of an ongoing process.

Discussion

There has been a lack of research on the fundamental attitudes, awareness, and behavioral approaches that young adults take with respect to sexual consent. This study is the first to develop scales focused on more general attitudes and behaviors regarding sexual consent, including (a) the necessity of explicit sexual consent, (b) the influence of relationship and contextual variables, (c) how much sexual consent discussion is occurring, and (d) what behavioral approaches students take with respect to consent.

Two subscales were identified in the attitudinal scale: (1) Asking for Consent First is Important assessed the need and desire to establish consent before any sexual activity is initiated, and (2) Commitment Reduces Asking for Consent assessed whether the commitment or length of a relationship influences the need to ask for consent. The two subscales of the behavioral scale were: (1) Consent Discussion/Awareness which assessed current awareness and discussions of sexual consent with peers, friends, or partners and (2) Consent is Negotiated Once which assessed whether students’ own consent behavior, in their most recent encounter, represented a one-time event or more of a process. The construction of the scales and the delineation of the four subscales provide significant conceptual and methodological advancements in our understanding of the complexities of sexual consent. The subscales themselves demonstrated modest to high internal consistency. Also, the significant relationship between the attitudinal and behavioral subscales provides evidence of predictive validity.

Sexual Consent and Gender

Women preferred a more cautious approach to sexual consent by stressing the need to obtain it more explicitly. Women, more than men, stressed the importance of consent on three of the four subscales, including: the importance of asking for consent first, that more committed relationships do not lessen the need for consent, and that consent is preferred as an ongoing process rather than a one-time event. This difference, albeit subtle, suggests a more thorough and deliberate approach to consent negotiations for women than men. The one behavior not related to gender was consent discussions/awareness. It appears that men and women do not differ in the amount of awareness and discussion they have with friends and partners regarding sexual consent issues. It is possible that women and men are exposed in equal measure to sexual coercion/assault prevention campaigns in the university environment. However, given that the subscale on consent discussion and awareness was only slightly above the midpoint, campus campaigns still have some work to do. Future research should examine gender differences in consent awareness levels among non-university/college populations.

When given the choice of methods for negotiating sexual consent, more students preferred to ask verbally rather than assume consent until a “no” was signaled. Women were more likely than men to prefer that a verbal “yes” be asked for prior to any sexual advances. This preference for verbally establishing consent is interesting given research findings that young adults are using more nonverbal signals and cues during sexual encounters than verbal ones (Beres et al. 2004, Hall 1998; Hickman and Muehlenhard 1999). The specific wording in the current study was one of preference, not actual behavior, and therefore probably represents an ideal instead of a reality. In addition, the question left open the matter of exactly who should be doing the verbal asking. The gender difference found on this item suggests that men are slightly more comfortable with nonverbal consent while women clearly prefer a consent process that is more overt. This finding seems to fit well with traditional sexual scripts for women and men. According to the traditional sexual script, the male role is to initiate sexual encounters, whether verbally or nonverbally, while the role for women is to control the level of sexual intimacy. O’Sullivan and Byers (1992) have found that men initiate sexual activity more frequently than do women, while research examining scripts used on a first date indicated that these traditional stereotypes regarding gendered roles continue to persist (Byers 1996; Rose and Frieze 1989). Based on the traditional sexual script, it makes intuitive sense that women would prefer more verbal asking to occur, largely on the part of men, because this allows women greater agency in their traditional role as limit setters or ‘gatekeepers’. Although this study did find a number of statistically significant gender differences caution should be exercised in interpreting their practical significance. There is a considerable amount of overlap in the distributions of women and men on these sexual consent attitudes and behaviors. For example, both women and men scored on the same side of the agree/disagree scale for each of the newly developed factors. In other words, they did not fundamentally disagree. One gender simply felt more strongly than the other on these sexual consent issues. The implication of this for consent negotiations between men and women is for the most part positive. At least on these particular factors, there does not seem to be a great divide or misunderstanding on how consent is negotiated.

Predicting Sexual Consent

The predictor variables of gender, sexual intercourse experience, relationship status, and number of sexual intercourse partners produced mixed results across the four attitude and behavior subscales. Sexual intercourse experience was related to three of the four subscales and in all cases those without intercourse experience indicated a greater need for sexual consent. This finding is supported by past research suggesting that sexual experience is linked in predictable ways to the perception of sexual intent and interest (Fisher and Walters 2003; Kelly and Bazzini 2003). In terms of experience variables, whether or not someone has experienced sexual intercourse seems to be a better predictor of consent attitudes and behaviors than current relationship status or the number of sexual intercourse partners.

Our exploration of the relationship between sexual consent attitudes and behaviors provided strong support for a reciprocal process in which consent attitudes influence behaviors and consent behaviors also influence attitudes. As an example, believing that consent needs to be asked before beginning sexual activity predicted awareness and discussions of sexual consent with friends and partners. However, heightened awareness and discussions also predicted the belief that consent needs to be established before sex starts. Social psychology has a long history of research delineating when attitudes predict behaviors and when behaviors will predict attitudes (Ajzen and Fishbein 1977; Cialdini 1988; Festinger and Maccoby 1964; Zimbardo 1970). Based on the finding of the present study, it seems reasonable to suggest a reciprocal relationship exists in which both are a cause and an effect of each other. Students who stress the importance of sexual consent are more likely to have conversations about it and associate with other individuals who see it as important. This can also lead to behavioral approaches to consent that reflect this concern. Likewise, reflecting on ones consent behaviors may lead to particular attitudinal stances that justify and reinforce current behavior.

Although the attitudes toward sexual consent were predictive of consent behaviors and vice versa, the relationships were modest. Weak attitude–behavior consistency has been an issue of long standing interest in human sexuality and social psychology research (Baumeister and Tice 2001). The difficulty when trying to obtain high attitude–behavior consistency is that there are “plenty of immediate situational factors and pressures [that] can intervene between an attitude and a behavior.” (Baumeister and Tice 2001, p.136). In sexual contexts at college or university, one of the primary culprits is alcohol. Sexual consent behaviors do not occur in a vacuum and therefore the strictest or most cautious standards may not be followed when the complexity of the immediate sexual situation presents itself.

Given that our study did find a relationship between sexual consent attitudes and behaviors, there are implications for sexual consent and/or sexual violence programming. College or university campaigns presenting strong attitudinal messages about sexual consent may have a positive influence on behavior. However, these programs may be strengthened if in additional to dealing with attitudes, the students are provided with concrete examples of successful behavioral strategies for negotiating sexual consent.

A number of limitations should be briefly mentioned. First, university based samples are not representative of the larger population and it is likely that the level of awareness surrounding date rape issues on college and university campuses has differentially influenced the perception of these students. Second, this research was also restricted to a heterosexual sample. Gay, lesbian, bisexual and transgender individuals may have different attitudes and approaches with respect to sexual consent and future researchers may want to assess GLBT individuals with modified versions of the scales developed here. A final limitation is the moderate reliability of the consent is negotiated once subscale delineated by the PCA. Further research is necessary to refine and strengthen this variable.

Determining the reliability and validity of a newly developed instrument is a continuous process and future research should include further efforts to improve the sexual consent scales presented here. These scales could also be related to other predictor variables. For example, measures of sexual assertiveness may be positively correlated with establishing consent prior to sexual activity and negatively related to anticipating difficulties with verbal consent.

Future research should also expand on the conceptualization of sexual consent. While the current attitude subscales focused on “shoulds” and contextual influences, this is not the only way to conceive of consent. For example, concerns about ensuring consent may also be understood in terms of self-doubts about one’s own sexual knowledge or skill, the desire to please one’s partner or the desire to be reassured as a sexual being. These cognitive–emotional representations of sexual consent should be incorporated in future research.

In conclusion, this study has provided valuable insights into how young adults think about and behave with respect to sexual consent. Given the centrality of the issue of sexual consent in situations of sexual coercion, it is vital that research continue on this topic. We believe that the attitudinal and behavioral sexual consent scales developed in this study are important contributions to future research on this topic.