Introduction

Men consistently report a more pronounced interest in short-term mating than do women (Buss & Schmitt, 1993; Kinsey, Pomeroy, & Martin, 1948; Schmitt, 2005; Schmitt et al., 2012). In a meta-analysis of studies on human sexuality performed from 1993 to 2007, Peterson and Hyde (2010) found a sizable gender difference concerning the interest in short-term mating (average effect size of Cohan’s d = .45). The effect size approaches that found in frequency of masturbation (d = .53) and use of pornography (d = .63), with men engaging more frequently in both. A meta-analysis conducted by Oliver and Hyde (1993) found even larger gender differences. Peterson and Hyde (2010) later demonstrated that the effect, consistent as it is, was mediated by the year of publication and by gender equity in nations and ethnic groups. For interest in short-term mating, the gender difference was smaller in newer publications and nations with stronger tendencies towards gender equality. We hypothesize that other mediating factors may have been overlooked altogether, and the factor of familiarity with potential mates plays a decisive role. We first review some recent findings and then report two experiments to test the hypothesis.

Evidence of Gender Differences

Most findings with regard to mating preferences are based on questionnaire surveys; however, they may be particularly problematic in this domain, as subjects may be in a conflict between privacy and compliance concerns in the sensitive area of sexuality (see, e.g., Alexander & Fisher, 2003). Even if subjects want to tell the truth, they are prone to the cognitive bias of the “hot-to-cold” empathy gap. That is, they are inaccurate in anticipating their own actions in an altered emotional state, such as when being aroused or attracted to someone (Ariely & Loewenstein, 2006).

To avoid these problems, Clark and Hatfield (1989) conducted a field experiment with high external validity. In their study, four male and five female psychology students approached a total of 96 opposite-sex subjects at Florida State University. After a standard introduction, “I have been noticing you around campus. I find you to be very attractive,” the experimenter’s confederates asked one of three questions randomly: “Would you go out with me tonight?”, “Would you come over to my apartment tonight?” or “Would you go to bed with me tonight?” Men and women were equally likely to consent to a date, but men were significantly more likely to agree to the apartment or casual sex offers. No female participant agreed to having casual sex. The study was conducted in 1978 right before the appearance of the first diagnosed case of AIDS in 1981. To tackle the issue of renewed awareness of sexually transmitted diseases, Clark and Hatfield conducted a second study in the same setting in 1982 (for more information on the history of this experiment, see also Clark & Hatfield, 2003). It produced almost identical results. However, since AIDS was not well known in 1982, Clark (1990) replicated the experiment again in 1990. Once again, the same pattern emerged. Men agreed to casual sex two-thirds of the time whereas women never agreed to casual sex.

In all three experiments, Clark and Hatfield did not find any effect of the requestors’ attractiveness, probably because they had not varied it systematically. Guéguen (2011) conducted a study that systematically varied the attractiveness of the requestor, with two men and women being very attractive and two men and women being averagely attractive. The study was carried out in the pedestrian zone of a medium size town in the west of France, thus recruiting not just students. Confederates told subjects “I find you very likeable and attractive” and then asked one of two questions: “Would you come to my apartment to have a drink?” or “Would you go to bed with me?” Guéguen’s results largely matched those of Clark and Hatfield (1989) and Clark (1990), with significantly more men than women agreeing to having sex. However, contrary to prior findings, Guéguen found that attractiveness played an important role. The only condition were this was not significant was for women consenting to casual sex. This might be due to the fact that out of all 60 women who participated in the study only one agreed to casual sex.

In 2009, two Danish researchers replicated the original Clark and Hatfield (1989) study and added measures of subjects’ relationship status, age, rating of confederate attractiveness, and subject awareness, asking if the subject was in any form aware that it was not an authentic request (Hald & Høgh-Olesen, 2010). We prefer to call this “suspiciousness” and later introduced a measure of it (see Exp. 2). They claimed that these factors would explain additional variance the former experiments could not account for. It is noteworthy that Hald and Høgh-Olesen had tried to replicate the experiment before, but their results were difficult to interpret due to the large number of rejections based on relationship status of the subjects and the small overall sample size (Hald & Høgh-Olesen, 2009). To test their hypothesis in the new study, 10 male and 11 female confederates approached a total of 389 subjects (173 men and 216 women) at either a university campus or a centrally located pedestrian or park area in one of the four largest Danish cities. As in all prior studies, Hald and Høgh-Olesen found a large gender difference in the acceptance of the sexual invitation, with women uniformly rejecting the advances. Contrary to Guéguen (2011), the confederates’ attractiveness was a good predictor for consent of female subjects in the date condition, but not of males. Finally, Hald and Høgh-Olesen also found that relationship status significantly contributed to the decision of the subject, with singles being six times more likely to accept any form of advance compared to subjects who reported being in a relationship.

Combining all studies, 508 women and 464 men have been approached (Table 1). The results of these studies strongly suggest a large gender difference in the acceptance of sexual invitations (Cramér’s V = .40). However, there is also some criticism that needs to be considered.

Table 1 Cumulative results of previous studies by sex and request

Most criticism aims at the explanation for the low number of female participants accepting the sexual offer. Interestingly, all experimenters but one (Hatfield) who found such low numbers were male, although male and female authors were equally critical about the close-to-zero findings. This is important because it shows that almost all female authors are very skeptical about the close-to-zero findings. The tally is based on the data reported by Conley (2011), Schützwohl, Fuchs, McKibbin, and Shackelford (2009), and Voracek, Fisher, Hofhansl, Rekkas, and Ritthammer (2006). Conley, Schützwohl et al., and Voracek et al. asked participants how they hypothetically would react to the Clark and Hatfield scenario, varying systematically the attractiveness of the requestor. They found that women were significantly more likely to accept sexual advances from a very attractive male requestor, a finding not present in field studies (Guéguen, 2011; Hald & Høgh-Olesen, 2010).

One question that cannot be ignored is that of the general validity of the Clark and Hatfield study and its successors. Conley, Moors, Matsick, Ziegler, and Valentine (2011) pointed out that what is measured in this scenario could be young people’s risk taking behavior instead of their interest in short-term mating. Maybe women who were approached in the study would have liked to have casual sex but perceived the scenario as dangerous. Indeed, young women have a high prevalence of becoming victims of sexual violence. Black et al. (2011) surveyed a sample of 16,507 U.S. adults (9,086 women and 7,421 men) with the representative random digit dial (RDD) method and found that 45 % of women reported having been victims to some form of sexual violence. Eighteen percent reported attempted or completed rape. Among the women who reported rape, 40 % implicated an acquaintance, such as a neighbor, a family friend, a first date, or someone briefly known. Women within the typical age bracket of college students, ages 20 to 24, experienced the largest per capita rate of nonfatal intimate partner violence (Catalano, 2007). Also, sexual double standards are still very much alive, which socially restrain female sexuality and endanger sexually liberal women to be called sluts or worse (Albury & Crawford, 2012).

One possibility to assess the impact of risk perception is to present participants with an alternative scenario that offers casual sex in a safe setting. Clark (1990) tried to do so by instructing four male and four female confederates to call their single friends and tell them that a good childhood friend would visit them who just came out of a relationship, was good looking, and a fantastic lover. Furthermore, the two would “really be made for each other.” They then asked “Would you be willing to go with her/him?” or”Would you be willing to go to bed with her/him?” While more than 90 % of the men and women agreed to go on a date, only 5 % of the women agreed to have sex, compared to 50 % of the men. This design clearly reduced the perceived physical danger for the female participants. However, the psychological danger was even higher, since a friend of the participant would know about the encounter and possibly share it with others, thus leading to a possible reputational risk.

The Current Study

We wanted to find out how the original findings would stand up to a more naturalistic setting, such as a cocktail bar, and a more safe setting, namely a laboratory. To establish a baseline, we first replicated the original Clark and Hatfield study on a German university campus, leaving out the “apartment” request because we felt it would not explain further variance in our design. In a second step, the same confederates approached different subjects in nightclubs and bars and repeated their request. In a further experiment, we told participants an elaborate cover story that made them believe they could have a date or sex without physical or psychological danger.

We predicted for Experiment 1 that the replication would yield similar results as former studies, with men consenting more often to a sexual invitation than women. We further predicted that in the party setting, the consent rate for date and sex for both men and women would be significantly higher than the consent rate in the campus setting. We also assumed that relationship status, attractiveness, and sexual skills would significantly predict consent for both sexes beyond location and type of request. While agreeableness and conscientiousness are somewhat negatively associated with sexual risk taking (Hoyle, Fejfar, & Miller, 2000), we did expect that the effect of personality traits would be too weak to have any predictive value. However, since the effect of personality was ignored in the past, we wanted to include it as an explorative factor.

We predicted for Experiment 2 that, given a safe setting, men and women would no longer differ in their consent rates for dating and sex. We also assumed that women and men would have a significantly higher consent rate for dating than for sex.

Experiment 1

Method

Participants

In total, 281 subjects (119 female, 162 male) were approached in a medium sized town in the south-west of Germany (about 200,000 inhabitants) on a university campus or in student clubs. All subjects were approached between June and July 2013. On campus, subjects were approached on sunny weekdays between 10 am and 6 pm and in the night clubs between 10 pm and 4 am.

Confederates were 14 psychology students (8 female, 6 male) from an experimental psychology course at the University of Mainz. The confederates were neatly dressed in the same outfit across conditions: casual attire, with women not wearing strong make-up, and avoiding big cleavage or short skirts (so the percentage of skin they showed was similar to that of the male confederates). Their average age was 22.00 years (SD = 2.38) for women and 21.83 years (SD = 0.68) for men. Confederates obtained partial course credit for their participation.

A total of 281 subjects were approached. We only retained subjects who self-identified as heterosexual. Thus, 5 subjects were excluded from data analysis, who self-identified as gay or lesbian. Another 9 subjects were excluded from the analysis because they claimed suspiciousness either at the outset or in the questionnaire.

Measures

We used a slightly modified methodology from the Clark and Hatfield (1989) study, leaving out the apartment condition. Instead, we introduced the new variable location, accounting for former criticism that the campus might be an unusual location for a date or sex request. This made it a 2 × 2 × 2 (Sex × Approach × Location) between-group design. We extended former studies by including measures of subjects’ age, and relationship status. Further, sexual history was measured by the number of former partners. Ratings of confederate attractiveness (“How attractive did you find the person that approached you?”) and sexual ability (“How sexually skilled would you estimate the person that approached you?”) were made on scales from 1 to 10, as were suspiciousness (“How believable did you find what just happened to you?”) and mood (“How good was your mood when you were approached?”). To assess personality, the short BFI-10 was used (Rammstedt & John, 2007). The BFI-10 consists of 10 questions, two for each personality factor (openness, conscientiousness, extraversion, agreeableness, and neuroticism).

Procedure

As part of the course introduction, confederates learned about and reflected on the experiment. In a special two hour session, confederates were carefully instructed and trained for their performance. In the campus condition, confederates approached subjects at strongly frequented areas during the day. In the party condition, confederates approached subjects at the bar, the dance floor, and the smoking area at night. They were instructed to approach unknown members of the opposite sex who were without obvious company. As in previous studies, they were further instructed to approach only people they found attractive. However, they were asked to be liberal about their judgment.

After the confederate chose a subject, s/he had to rate the attractiveness of the subject on a scale from 1 to 10 and state his or her own self-confidence on the same scale. The confederate then was randomly assigned one of the two conditions. The confederate approached the subject using a standardized introduction: “Hi, normally I don’t do anything like this, but I find you totally attractive.” The confederate then asked one of the two questions: “Would you like to go out for a coffee with me sometime?” or “Would you like to have sex with me?” If the subject tried to start a conversation, the confederate was instructed to insist on an answer to the initial question. Each confederate approached 20 subjects, five in each condition.

Confederates usually went out in groups (2–5 people). After a subject was approached, the group left the area and went to a different site to avoid being noticed prior to the encounter by potential subjects. One confederate usually was able to approach roughly 3 subjects per hour.

Following the request, the confederate left the scene and another member of the research team approached the subject and debriefed him or her discretely. Subjects were informed about the experiment and that they were chosen because the confederate found them attractive. Subsequently, they were asked to fill in a questionnaire asking about their age, relationship status, sexual orientation, sexual history, rating of the confederate’s attractiveness and sexual ability, suspiciousness about the experimental context, their current mood, personality, and, in the party condition, about how inebriated or drunk they felt (on a scale from 1 to 10). Subjects’ drunkenness was also rated by the confederate after their interaction. After subjects filled in the questionnaire, or declined to do so, they were thanked for their participation and offered an email address to learn more about the experiment. Upon request, they later received an email with information and results about the experiment.

Results

A total of 267 subjects were included in the analysis. Of these subjects, 45 % left out at least one question. We used multiple imputation (5 iterations) to avoid a substantial loss of precision and power in subsequent analysis (Sterne et al., 2009).

As predicted, using a Chi square-test, significantly more men than women consented to a sexual invitation, χ2(1) = 18.18, p < .001, Φ = .38. Interestingly, the same was true for the date condition, χ2(1) = 10.37, p = .001, Φ = .27. Further, significantly more men agreed to either offer at the party as compared to the campus condition, χ2(1) = 26.24, p < .001, Φ = .41. Women did not produce a location effect, χ2(1) = 0.38, p = .534, Φ = .05 (Table 2).

Table 2 Consent rate by relationship status

A binary logistic regression was conducted, including sex, location, type of request, relationship status, age, sexual experience, attractiveness, self-confidence, and sexual skills of the confederate, and the Big Five personality traits as predictors. A backwards likelihood-ratio test with the criterion p < .05 was used for predictor inclusion into the regression. The same was repeated for women and men separately, leaving out the predictor sex (Table 3).

Table 3 Results of binary logistic regression

Attractiveness and consent correlated significantly, r(265) = .29, p < .001. The same was true for the confederate’s perceived sexual skills and consent, r(265) = .38, p < .001. Both predictors intercorrelated with r(265) = .53, p < .001.

Female subjects (M = 4.88, SD = 2.98) were rated by our confederates to be as attractive as male subjects (M = 4.58, SD = 2.70) on a scale from 1 to 10, F(1, 265) < 1, η 2p  = .03, using a one-way ANOVA. However, female confederates were rated to be significantly more attractive (M = 6.84, SD = 1.89) than the male confederates (M = 5.13, SD = 2.09), F(1, 265) = 48.18, p < .001, η 2p  = .15. As expected from the above correlations, female confederates were also rated as significantly more sexually skilled (M = 6.76, SD = 1.46) than male confederates (M = 5.59, SD = 1.55), F(1, 265) = 28.53, p < .001, η 2p  = .12. Interestingly, while male confederates clearly had a lower success rate than females, they were significantly more self-confident (M = 7.22, SD = 2.45) before approaching the subjects than were women (M = 6.60, SD = 2.25), F(1, 265) = 4.53, p = .034, η 2p  = 0.01.

We also found that for the party condition, self-perceived level of alcohol intoxication correlated significantly with consent rate, r(127) = .20, p = .023. Confederates’ ratings of the subjects after their interaction also correlated significantly with self-perceived level of alcohol intoxication, r(127) = .43, p < .001. Confederates’ attractiveness ratings correlated significantly with subjects’ drunkenness, r(127) = .21, p = .016. However, the time of the approach did not correlate with the attractiveness scores of confederates or subjects, r(127) = −.01, and r(127) = .05, respectively.

Discussion

Whereas former studies (Guéguen, 2011; Hald & Høgh-Olesen, 2010) found that the confederate’s attractiveness was a good predictor of subject’s consent, we found a low correlation between attractiveness and consent. The confederate’s perceived sexual skills, as proposed by Conley (2010), were a better predictor of consent. Both predictors intercorrelated significantly, and when the predictor “sexual skills” was excluded from the binary logistic regression, attractiveness took its place. This leads us to conclude that judged sexual skills and attractiveness both should be seen as part of a common higher-order construct.

Gladue and Delaney (1990) found that, controlling for alcohol consumption, bar patrons of the opposite sex were perceived to be more attractive as the night progressed. However, we did not find such a trend in the party condition. As the night progressed, neither the confederates nor the subjects received increased attractiveness scores. However, with increasing drunkenness, subjects rated the confederates to be more attractive. We can ascertain that this effect was not merely due to a change in self-perception. Our subjects’ self-reported blood alcohol levels (which may be prone to misjudgment; see Grant, LaBrie, Hummer, & Lac, 2012) and confederates rated how inebriated the subjects had seemed during their interaction. Subjective and outside evaluations matched, indicating that here self-perceived level of alcohol intoxication can be used as a crude estimate for drunkenness.

Especially for women, there might have been concerns to consent to casual sex for safety reasons. To counter these concerns and measure their possible effects on the outcome of Experiment 1, we conducted a second experiment. In Experiment 2, we made up an elaborate cover story and invited 60 male and female subjects into our laboratory. We presented 10 pictures of persons of the opposite sex to these subjects and led them to believe that the latter had either consented to date them or to have sex with them. The subjects then could choose from the pictures whom they would like to meet. We made subjects believe that sex/a date would occur later in a safe environment. In this setting, subjects were not exposed to the danger of gaining a reputation since the experiment was anonymous. Also, possible physical risk should be perceived as minimal because subjects believed the people in the pictures were pre-selected.

Experiment 2

Method

Participants

Subjects were 30 men and 30 women who answered an online announcement. All subjects self-identified as heterosexual and were randomly assigned to either the sex or the date condition. The majority of subjects were students at the University of Mainz.

Measures

The questionnaire asked about age, sex, and relationship status, with the answer categories “single,” “in a relationship,” “complicated/unclear,” and “other.” Personality was assessed using the BFI-10. After the experiment, the subjects were asked how suspicious they felt about the experiment on a scale from 1 (very suspicious) to 4 (not suspicious at all).

The pictures of potential male and female partners were matched on the basis of three primary categories: facial expression, photo context, and skin. Facial expression meant whether the person in the picture was smiling or not. Photo context indicated where the picture was taken and what was in the background, and made sure there were no other cues like alcohol or animals present. In the last category, we checked that all participants showed a similar amount of skin and women had no big cleavage. All people in the pictures were rated to be in the range of average attractiveness by 87 independent raters.

Procedure

Subjects were recruited by announcing online that the psychology department was looking for participants in a dating study. The ad mentioned a cooperation with a large online dating site developing and validating a novel match-making process. Subjects would have the chance to meet people and learn more about their dating behavior. When participants contacted the given email address, they received a request for a picture as well as for an answer to the questions “Do you like horror movies?”, “Do you like the taste of beer?”, “What is your sexual orientation?”, and “Have you participated in any study on courtship behavior before?” They also were informed that they would be invited to the laboratory and after a screening would have the chance to meet new people.

A week after they had sent in their picture, subjects were invited to the laboratory. There, we handed them a written description of the cover story. It said that the subject would be presented with 10 pictures of people who had seen their picture beforehand and were interested in either having a date or having sex with them, depending on the condition. If the subject agreed to meet with any one of them, the research team would film the first half of an hour of their encounter and then leave them alone to continue their date or to have sex. The meeting would occur at a location provided by the research team. Subjects also signed a consent form that explained that not all information about the experiment could be disclosed beforehand and that a complete debriefing would follow after the experiment. For the sex condition, it was stressed that the people in the pictures particularly wanted to have sex with them. It might be important to know that in Germany, arranging for two people to have sex is not illegal, as there are countries where this might be the case.

Subjects were seated in front of a laptop and had the opportunity to ask questions. The experimenter also explained to them that they could withdraw their consent to the experiment at any time. They then were presented with 10 pictures of people of the opposite sex in a random order. For each picture, they were asked how attractive they found the person and if they wanted to have a date or sex with them, respectively. During the testing, subjects were left alone in the room, with a researcher outside the door so subjects could ask questions any time. After finishing the testing, subjects completed a questionnaire. At that point subjects still believed they would meet the people they had chosen from the pictures.

After the experiment, subjects were fully debriefed about the purpose of the experiment. They were also asked to not talk about the purpose of the experiment with other for the next couple of month till the data collection was finished to avoid “deception leakage”. All subjects were offered to meet other participants at a come-together later and to receive additional information upon request.

Results

One male subject who was in a relationship and claimed suspiciousness (saying that he knew he would not actually meet any women in the experiment) was subsequently excluded from further analysis. The rest of the subjects were single and did not express strong suspicions (M = 2.98, SD = 0.63). The numbers show that, on average, people thought our cover story was “mostly believable.” However, 15 % of the subjects said they found the cover story “somewhat unbelievable.” When asked of what in particular they were suspicious, they expressed rather minor concerns and did not have any doubts about meeting the people in the pictures.

A new group of 67 female and 20 male psychology students rated the attractiveness of portrait pictures (female pictures: M = 5.21, SD = .90; male pictures: M = 4.59, SD = 0.99), as did the subjects (female pictures: M = 5.56, SD = 0.88; male pictures: M = 4.73, SD = 1.34), on a scale from 1 to 10. A dependent-samples t test for the new group and an independent-sample t-test for the original subjects revealed that these differences were significant: attractiveness ratings differed between female and male pictures. A significant effect was found for gender of the pictures for the new group of students, t(86) = −7.88, p < .001, d = .68, as well as for the original subjects, t(57) = 2.92, p = .005, d = .74.

Of all male subjects, 100 % agreed to have a date or sex with at least one woman. This rate did not differ from the female consent rate (97 %). Men chose on average to meet 3.67 (SD = 1.45) women for a date and 3.57 (SD = 1.16) women to have sex with. Women chose on average to meet 2.80 (SD = 1.21) men for a date and 2.73 (SD = 1.87) men to have sex with.

A two-way ANOVA indicated a significant main effect of gender on the number of selected partners, F(1, 55) = 5.07, η 2p  = .08. There was no main effect of condition on the number of selected partners, F(1, 55) < 1, η 2p  < .01, and no interaction effect between gender and condition on number of partners selected, F(1, 55) < 1, η 2p  < .01 (see Fig. 1).

Fig. 1
figure 1

Results of experiment 2. Men chose to meet significantly more women compared to women choosing men, independent of the condition. Error bars represent ±1 SEM

General Discussion

As proposed in our first hypothesis, we found that both on campus and at parties, men agreed more willingly to casual sex than did women. For female subjects, the close-to-zero acceptance rate persisted. Only one out of 119 women agreed to casual sex.

Contrary to the second hypothesis, only men consented more willingly to any type of request in the party condition. Women practically had the same rate of consent for the campus and the party conditions. It is noteworthy that men agreed less often to casual sex than to a date, even when stratified for relationship status. This differs from previous studies (Clark & Hatfield, 1989), where men were more likely to agree, the more explicit the offer was. We also found that alcohol intoxication correlated significantly with rated attractiveness and consent rate. This is well in line with previous findings that alcohol leads to a loosening of sexual inhibition (George & Stoner, 2000). However, this leads us to question whether the more naturalistic setting or the alcohol consumption was responsible for the higher consent rate in men.

In line with the third hypothesis, relationship status and sexual skills were good predictors for male and female consent rate. We did not find that attractiveness was a good predictor. However, it seems participants used the same criteria to assess attractiveness and sexual skills, as both correlated significant with each other. It makes sense to assume that subjects use the simple heuristic that the better looking someone is, the more pleasure they will get from them in bed. Also, conscientiousness was a good predictor for male consent rate. This is surprising as a meta-analysis by Hoyle et al. (2000) found that conscientiousness was associated with low promiscuity and less sexually risky behavior. We suspect that this might be due to the fact that we used the short BFI-10. While reliable and valid, it holds only two questions per trait, which cannot cover all facets of the Big Five personality traits (Rammstedt & John, 2007).

Male confederates were significantly more confident before approaching female subjects but significantly less successful than female confederates. This might reflect the Dunning–Kruger effect, a cognitive bias in which unskilled individuals overestimate their abilities (Kruger & Dunning, 1999). Alternatively, it could be that in our society men are taught to be self-confident in romantic interactions whereas women are more often taught to be coy.

In our second experiment, the first hypothesis was affirmed. Women and men had the same consent rate for both conditions. However, male subjects wanted to meet significantly more potential partners than did female subjects (Cohen’s d = .61). We could not confirm our second hypothesis of Experiment 2. Female subjects chose as many partners to meet for a date as to have sex with.

Results of both experiments can be explained with both evolutionary and social theories. Sexual strategies theory (Buss & Schmitt, 1993) predicts that women are less likely to engage in casual sex than men. This was found in the first experiment. However, women do engage in casual sex as a means to evaluate or attract potential long-term partners or to mate with males who have highly desirable traits, genes, positions, and/or resources (Hald & Høgh-Olesen, 2009). This would explain the findings of the second experiment. When given the chance to choose from a range of males, women did not hesitate to select the most attractive. Thus, the high consent rate to casual sex might be explained in evolutionary terms as a high consent rate to evaluate a potential partner through short-term mating. Alternatively, the result can be explained by the sexy son hypothesis, stating that females engage in short-term mating with attractive males as a mean to get physically attractive sons, who in turn will produce more grandchildren (e.g., Dawkins, 1976).

Social role theory (Eagly, Wood, & Diekman, 2000) predicts that women are less likely to engage in casual sex than men because it would be seen as inappropriate behavior and might carry negative consequences. Even in a situation where it is socially acceptable for women to be sexually less restrained, such as a party, society dictates clear boundaries, such as not to be easy to get. This is in line with the first experiment. However, women will engage more willingly in casual sex if the environment encourages them to and they can expect a positive outcome of the encounter (Conley, Ziegler, & Moors, 2013). This explains the high consent rate for casual sex in the second experiment. Also, women were aware that they would have a short time before they would have sex with the potential partner and could resign from the experiment at any time, thus staying in control of the situation at all times.

One challenge both experiments encountered is the sampling bias. It is possible that we encountered subjects with different traits at a campus and at a club, albeit a student club. Also, it is very likely that only subjects who were interested in meeting potential partners applied for Experiment 2. This might lead to a limitation of the generalizability to the general population. However, the setting used in Experiment 1 has limited ecological validity as offers to date or have sex are usually made in a safe and friendly social environment. We think Experiment 2 stands out in that is has achieved a higher degree of ecological validity than Experiment 1 and most other existing experiments concerning this research question.

We also found that women were rated as significantly more attractive than men by the subjects in both studies. This was not due to poor sampling but represents the actual population. Men are generally perceived as less physically attractive than women by both sexes (Grammer, 1994; Symons, 1979). In Experiment 1, the difference in attractiveness (expressed in Cohen’s d) was d = .85 while the difference in acceptance rate was d = .65. Thus, the difference in attractiveness was even larger than the difference in acceptance rate. However, in Experiment 2, the difference in attractiveness rating persisted whereas the difference in acceptance rate vanished. This leads us to conclude that attractiveness was not solely responsible for the large gender difference in the acceptance rate in Experiment 1.

Our findings qualify the common belief that men more readily consent to casual sex than do women. We found a strong context effect. Whereas the gender differences are large when subjects are approached on the street or at a party, these gender differences are significantly smaller in a non-threatening environment. Women no longer had a close-to-zero rate for the acceptance of casual sex. However, it is noteworthy that a significant gender difference in the number of potential partners that were accepted for any given activity persisted.

Further studies should focus on three issues: First, it was shown that the environment had a huge effect on the consent rate for sex of women and men. However, these findings were the first of their kind and need to be validated. Second, former studies have focused solely on Western white students in their early 20 s. It would be interesting to vary age, education, and culture to see if the findings hold up in more diverse samples. Third, since there is no field study to date that varies sexual orientation, it would be very interesting to repeat the former experiments in the lesbian, gay, bisexual, transgender, and queer (LGBTQ) community.