The effects of target attractiveness on men's sexual arousal in response to erotic auditory stimuli.

People become sexually aroused through a variety of external stimuli, including auditory stimuli and written accounts. Sexual arousal produced from such stimuli likely requires mental imagery processes that have been investigated in past research. The present study seeks to manipulate the visual imagery used by participants in conjunction with a standard erotic auditory stimulus in order to determine the effects of such stimuli on rates and levels of subjective sexual arousal as predicted from our understanding of underlying mental processes.

Past research has indicated the importance of the visual sensory modality for male sexual arousal (Byrne, 1977; Ellis & Symons, 1999; Gold & Gold, 1991; Przybyla, Byrne, & Kelley, 1983). Men, more than women, focus on the physical aspects and characteristics of sexual fantasy partners and on the sexual behaviors being performed (Ellis & Symons, 1999; Geary, Vigil, & Byrd-Craven, 2004) and may also be more readily prompted to fantasize by exposure to external, visual cues (Gold & Gold, 1991). In the absence of external visual stimuli, the ability to form clear, vivid images still appears to be important in facilitating arousal (Harris, Yulis, & LaCoste, 1980; Przybyla & Byrne, 1984; Przybyla et al., 1983; Smith & Over, 1987, 1988).

Studies examining the phenomena of sexual fantasy and arousal draw heavily from cognitive research on mental imagery. Sexual fantasy, which utilizes mental imagery, acts as an intermediary step between external sexual stimuli and overt sexual responses. Several past studies suggest that mental imagery facilitates the production of sexual arousal from external stimuli (Harris et al., 1980; Janssen, Everaerd, Spiering, & Janssen, 2000; Przybyla & Byrne, 1984; Smith & Over, 1987, 1988; Tokatlidis & Over, 1995), that individual differences exist in the ability to form and utilize mental imagery (Reisberg, Culver, Heuer, & Fischman, 1986), and that this ability relates to greater sexual arousal through fantasy (Harris et al., 1980).

In a study by Smith and Over (1987), men who were able to form more vivid visual images during sexual fantasy reported higher levels of sexual arousal, but the content of fantasy was independent of imagery ability. In subsequent research (Smith & Over, 1988), men's ability to form vivid images again predicted reported levels of sexual arousal during fantasy but not when participants simply viewed erotic material and were instructed not to fantasize. These studies suggest that the formation of vivid mental images is a key factor in an individual's ability to become aroused during sexual fantasy, regardless of whether external erotic stimuli are present.

Janssen and colleagues (2000) proposed an automatic information-processing model in which fantasy-based sexual arousal results from two principal stages. The first stage, appraisal, involves encoding of the sexual stimulus and matching in memory, which provides emotional meaning to the stimulus. The second stage involves response generation, in which the meaning of the stimulus is integrated with motor sensory processes, resulting in the subjective experience of sexual arousal. The authors hypothesized that erotic content presented in the form of pre-attentive priming can be enough to elicit sexual arousal in response to a neutral, overt stimulus. The type of overt stimulus should not matter, as long as the sexual meaning is stored within semantic memory. However, they acknowledged that erotic pictures, which are easily perceived without higher-order cognitive processes, most readily activate meaning within semantic memory. Words or sounds require further processing.

Controlled and attentional processes also affect these central pathway operations. In the second stage of the model by Janssen and colleagues (2000), regulatory cognitive processes have the potential to inhibit or facilitate subjective sexual response. Przybyla and Byrne (1984) also demonstrated that the path from attention to an erotic stimulus to sexual arousal could be disrupted by a distraction task, further suggesting that a link between stimulus and response is not always automatic. Instead, the stimulus elicits cognitions in the form of an internal, working fantasy or mental imagery that, once manipulated and processed, leads to physiological and subjective sexual arousal.

Research into sexual fantasy demonstrates that erotic auditory stimuli facilitate sexual arousal through a two-step cognitive process that involves appraising the stimuli and translating it into visual imagery (Przybyla & Byrne, 1984; Przybyla et al., 1983). According to this theory, the qualities of visual, as opposed to auditory, stimuli may be processed differently in the information-processing relay between stimulus and response. Participants tend to report that visual stimuli are more sexually arousing than are auditory stimuli, suggesting that there may be an additional step in which the auditory stimuli is processed into an internal, visual form (Przybyla & Byrne, 1984). This suggests that sexual arousal to auditory stimuli should occur at a slower rate than arousal produced by visual stimuli or that produced by a combination of the two modalities (Gaither & Plaud, 1997).

Although therapeutic techniques that seek to alter mental imagery, such as covert sensitization, have often been used as a method to decrease paraphilic arousal (Dougher, 1995), a less addressed question is whether sexual response to a highly erotic, auditory stimulus can be altered by pairing it with a prior congruent or incongruent visual stimulus. A recent summary of research on expectancy and expectancy violation by Roese and Sherman (2007) can inform such a hypothesis. These authors explain that in order to make sense of current experiences, inward-bound sensory information must be constantly matched with associated information that exists in semantic memory. Sensory input that is congruent with expectancies held in semantic memory facilitates cognitive processing, whereas incongruencies result in disrupted processing and an increased expenditure of cognitive resources. If an erotic auditory stimulus must first be transformed into internal, visual imagery, then matching it with an expectancy-incongruent visual image (such as an unattractive speaker) should disrupt the production of subjective sexual arousal, whereas an expectancy-congruent visual image (such as an attractive speaker) should facilitate arousal. This study hypothesizes that participants who are exposed to a visually attractive (A) target prior to hearing an erotic auditory stimulus should report higher levels of sexual arousal than should participants who are exposed to a visually unattractive (U) target. Because incongruent pairings require more effortful cognitive processing, it is also expected that participants in the U condition should be slower to reach their peak level of sexual arousal compared with those in the A condition.

Because we assume that visual processes mediate between auditory stimuli and arousal, we predict that participants in a no visual stimuli, or control (C) condition, should show no differences in levels of peak arousal compared with subjects in the A condition. When compared with participants in the U condition, those in the C condition should show higher levels of peak arousal because they will most likely generate imagery that is personally arousing to them. Second, the rate at which C participants reach their peak arousal should be slower compared with those in the A condition because they will not be presented with a facilitating secondary stimulus and thus may take longer to internally generate their visual imagery. Whether the rate of arousal for C participants is faster than that for U participants should be an open research question because there is little past research to inform this specific issue.

Method

Participants

Sixty-six adult male students at a Pacific university were recruited from undergraduate psychology courses and were given course credit for participation. Four participants who reported being either exclusively or mostly sexually attracted to other males were excluded from further analysis; 2 participants were excluded because of technical problems occurring during the experimental session, and 1 was excluded because he provided incomplete data. The remaining 59 individuals were between the ages of 18 and 31 (M = 21.15, SD = 2.94). Most participants (83.1%) reported being unmarried/not formally committed, although nearly one half were in an exclusive romantic relationship.

Stimuli

Participants were randomly assigned to one of three conditions (attractive [A], n = 19; unattractive [U], n = 18; or no picture [C], n = 22). (1) The pictures used were determined through extensive pilot testing of over 40 naturalistic headshot photographs with raters who did not participate in the study itself. Pictures with the highest and lowest mean scores of attractiveness were used to select the attractive and unattractive stimuli. Two photographs were selected for each condition and counterbalanced between participants to ensure that the responses were not in reaction to a specific visual stimulus. Though several ethnicities were included, all pictures chosen were of Caucasian females.

Participants were told they would hear an erotic recording made by a female volunteer, though the recording was actually made by a person unrelated to the research. This recording was selected because it included graphic descriptions of specific sexual activities and contained several overt references to the target's sexual anatomy, all of which have been shown in past research to be foci of male sexual fantasy (Ellis & Symons, 1999; Geary et al., 2004). The audiotape was narrated by a female and depicted the sexual activity as female-initiated and female-centered because previous research has indicated that this type of stimulus garnered the highest ratings of subjective arousal in similar samples (Hall, Binik, & di Tomasso, 1985; Heiman, 1975). During the 5-min erotic recording, participants heard a series of tones at 30-s intervals. This included a tone at the very beginning and one at the very end of the recording. Participants were trained to attend to this tone by hearing it once before the session began. Each time participants heard this sound, they were instructed to self-report their level of sexual arousal at that moment on an 11-point Likert-type scale ranging from 0 (not at all aroused) to 10 (extremely aroused). Separate Likert-type items were provided for each audio cue, and participants were instructed to progress to the next item each time the sound was heard.

Procedure

Participants were individually led into a room with reduced outside stimuli. Participants were seated at a table with a cassette player and headphones, the experimental testing packet, and a covered folder containing stimulus material. All participants provided informed consent before beginning the experimental session, after which they were told to put on the headphones and press Play on the tape recorder.

On the audiotape, a male instructor greeted participants, assured the confidentiality of their responses, and requested that they provide honest answers. All major instructions from this point were given both verbally (on the tape) and in writing. Participants were told to open their experimental testing packet and to self-report their baseline level of sexual arousal on an 11-point Likert-type scale.

Participants were informed that they would be listening to an erotic tape recording made by a female volunteer. Information on the "volunteer" was provided in a covered folder, including the unattractive or attractive photograph and a brief personal description, which was held constant across conditions. Participants in the C condition received a personal description but no photograph. Participants were asked to open the folder and were given 30 s to review the information inside. They were then told to close the folder and keep it closed for the remainder of the study. Participants were told to relax and to concentrate on the recording, after which the auditory stimulus began and participants rated their subjective arousal at 30-s intervals.

Participants were notified on the tape when the experiment had ended. They then reported demographic data, including their age, relationship and marital status, and sexual preference. They were also asked to rate the attractiveness of the picture they had seen on a 7-point Likert-type scale (control participants were not presented with this question), and they were asked to indicate whether an image of the person who made the recording frequently came to mind while they listened (5-point Likert-type scale). Participants then placed their response packets through a slot in a sealed box. Project assistants debriefed the participants by informing them that the female in the photograph had not actually recorded the erotic stimulus and by explaining the true purpose of the experiment. All participants were asked to avoid discussing the details about the experiment with others.

Results

Manipulation Cheeks

Three manipulation checks were conducted; the first assessed the equivalence of the two attractive and the two unattractive pictures in order to justify combining them into one A condition and one U condition. The second assessed the effectiveness of the experimental manipulation, and the last assessed the use of mental imagery by participants across conditions. An initial analysis of variance (ANOVA) indicated significant differences in attractiveness ratings among the four pictures, F(3, 33) = 53.77, p < .001, and a post hoc Tukey's test demonstrated that the two attractive pictures (M = 6.00, SD = .32; M = 5.00, SD = .31) differed significantly (p < .001) from the two unattractive pictures (M = 1.22, SD = .32; M = 1.77, SD = .32), but neither the two attractive nor the two unattractive pictures differed significantly within pairs (p =. 133 and .62, respectively). Thus, the four groups were combined to create the two experimental conditions, and these combined groups were analyzed with another ANOVA. Participants in the A condition rated their pictures as significantly more attractive (M = 5.47, SD = .985) than did participants in the U condition (M = 1.50, SD = 1.07), t(35) = -11.71, p < .001. Regarding the tendency to visualize the person who made the recording, there was no significant difference between participants assigned to the U (M = 3.28, SD = 1.36), A (M = 3.11, SD = 1.33), and C (M = 2.93, SD = 1.38) conditions, F(2, 55) = .32, p = .73, suggesting that participants across conditions were equally likely to have experienced visual imagery.

Baseline Arousal and Immediate Response to Manipulation

Participants showed no differences in baseline arousal, F(2, 56) = .36, p = .71 ; U (M = .78, SD = 1.80), A (M = 1.11, SD = 1.73), and C (M = 1.23, SD = 1.69). A 3 x 2 ANOVA with repeated contrasts (3 conditions x 2 measurement points) conducted on scores from the baseline to those reported at the first tone (after participants had seen the pictures but before they began listening to the erotic recording) revealed significant effects for time, Wilk's [lambda] = .86, F(1, 56) = 8.93,p = .004); group, F(2, 56) = 3.28,p = .045; and a Time x Group interaction, Wilk's [lambda] = .86, F(2, 56) = 4.58,p = .014. The U group showed a reduction in their reported arousal, whereas increased arousal was evident for the A and C groups (see the first two columns of Figure 1). A post hoc Tukey's test showed only the A and U conditions to differ significantly (p = .037).

[FIGURE 1 OMITTED]

Overall Level of Arousal

A 3 x 11 ANOVA with repeated contrasts (3 between-subject conditions x 11 within-subject measurement points; see Table 1) indicated significant differences between conditions in participants' mean scores of arousal across measurement points. The U group showed lower average arousal scores compared with both the A and C groups (p [less than or equal to] .05), whereas the latter two showed no significant differences. The highest arousal score for each participant was then noted and analyzed with an ANOVA. Differences in peak arousal were not significant, F(2, 56) = 1.08, p = .35, between the U (M = 4.56, SD = .63), A (M = 5.84, SD = .61), and C (M = 5.23, SD = .57) groups.

Rate of Arousal Change

Differential rates of arousal change were tested by using multivariate analyses that could handle the anticipated curvilinear slope of arousal scores (Masters & Johnson, 1966) and the possible different rates of change across conditions. Tests of within-subject contrasts were used to interpret the pattern of arousal responses by examining the highest level of contrast (e.g., linear vs. quadratic vs. cubic) at which a significant result was obtained. Because of the large number of measurement points, a p value of < .01 was set for the acceptance of a within-subject contrast pattern. The same 3 x 11 ANOVA reported earlier was used to examine the changes in arousal from the baseline measurement (Time 1) to the end of the auditory stimulus (Time 11; see Table 1). Because Mauchly's test of sphericity was significant, [chi square] (54) = 479.27, p < .001, the Greenhouse-Geisser test (Greenhouse & Geisser, 1959) was used to interpret within-subject effects. Significant differences existed when comparing scores at each point in time (p < .001) and in terms of a Time x Group interaction (p = .002). Further, tests of within-subject contrasts showed significant linear and quadratic effects both for time (p < .001 or each level) and a Time x Group interaction (p < .001 and p = .003, respectively). No other higher order contrasts reached a significance of p < .01. Thus, all contrast tests within the main ANOVA and subsequent, post hoc parallel tests (see below) were interpreted in quadratic terms. In summary, although all groups showed a curvilinear pattern of arousal over time, this pattern differed across experimental conditions (see Figure 1).

In order to interpret the Time x Group interaction effects, three separate 2 x 11 ANOVAs with repeated contrasts were conducted, comparing two groups at a time (see Table 2). In comparisons of the A and U conditions, multivariate tests in the ANOVA revealed a significant effect for time (p < .001) but no significant Time x Group interaction (p < .05). The Greenhouse-Geisser test indicated significant within-subject effects for time when comparing reported arousal at each point (p < .001) and in terms of a Time x Group interaction (p = .003). Further, the tests of within-subject contrasts revealed significant effects at the quadratic level both for time (p < .001) and for a Time x Group interaction (p = .002). Thus, although both A and U participants showed a curvilinear pattern of arousal over time, this pattern was different between the conditions. Specifically, participants in the U condition were slower to show arousal than those in the A condition (see Figure 1).

These same tests were repeated for a comparison of the U and C conditions. Multivariate tests in the repeated-measures ANOVA revealed a significant effect for time (p < .001) but no significant Time x Group interaction (p =. 13). Again, significant within-subject effects were found when comparing reported arousal at each point in time (p < .001) and in terms of a Time x Group interaction (p = .046). Further, the tests of within-subject contrasts again showed significant quadratic effects both for time (p < .001) and for a Time x Group interaction (p = .005). Thus, although both U and C participants showed a curvilinear pattern of arousal over time, those in the U condition were slower to show arousal.

The final comparison was between the A and C conditions. Multivariate tests revealed a significant effect for time (p < .001) but no Time x Group interaction (p = .56). The Greenhouse-Geisser test indicated significant within-subject differences in arousal at each point in time, F(3.29, 128.32) = 28.16, p < .001, and no significant Time x Group interaction (p = .07). Tests of within-subject contrasts showed significant effects at the quadratic level for time (p < .001) but not for the Time x Group interaction (p = .77), indicating quadratic effects for both groups but no difference in the rate of change.

Discussion

The present study assessed men's self-reported arousal to an erotic auditory stimulus after manipulating the visual attractiveness of the speaker. As expected, participants demonstrated a curvilinear pattern of arousal that differed across experimental conditions. Specifically, participants in the U condition were slower to reach peak arousal and showed lower average arousal levels. No significant differences were found between groups in reported peak arousal levels. In addition, significant differences in arousal existed between the A and U conditions immediately after exposure to the visual stimuli but prior to hearing the auditory stimulus. In line with the expectation that C participants would self-select highly arousing imagery, this group showed comparable peak arousal levels to those receiving an attractive picture. However, it was also predicted that the absence of a facilitating visual stimulus would result in the C group participants being slower to arouse, and although there was a slight trend in this direction, it was not statistically significant.

The present results are in line with the contention that incoming erotic stimuli are pattern-matched with information in semantic memory (Janssen et al., 2000; Spiering, Everaerd, & Janssen, 2003). The data also support the model of Przybyla and colleagues (Przybyla & Byrne, 1984; Przybyla et al., 1983), which places internal visual imagery as a mediating step between erotic auditory stimuli and resulting sexual arousal. As seen here, this mediating imagery can be manipulated by providing differing relevant information to hold in semantic memory with subsequent changes in rates of sexual arousal--a notion in agreement with expectancy research demonstrating that processing stimulus-semantic memory incongruencies is less fluid and more effortful (Bartholow, Fabiani, Gratton, & Bettencourt, 2001). It is important to note that all groups imagined the speaker during the auditory stimulus at equivalent levels, suggesting that such stimuli are transformed into visual imagery even in the absence of overt, visual stimuli (Przybyla & Byrne, 1984; Przybyla et al., 1983).

Several interpretations may explain why U participants were slower to arouse, yet achieved levels of peak arousal similar to the other groups. First, these participants may have tried to substitute their own, more arousing imagery. Although the more effortful processing required to generate a novel mental image and effectively substitute it for the one already held in semantic memory was expected to result in a slower rate of sexual arousal, once done, participants could be expected to attain levels of arousal roughly comparable to the other groups. Second, the chosen auditory stimulus may have been so arousing that participants were able to focus strongly on the graphic descriptions provided in the recording alone and focus less on the visual image held in semantic memory. Finally, the unattractive target may have served as a distractor to which participants habituated over time, allowing them to overcome the impediments of this imagery as the experiment progressed. However, neither of the latter two explanations seem sufficient given that U participants reported visualizing the target at levels equivalent to the other groups. It is interesting to note that the expected facilitative effects on arousal levels were evident during the short period after the visual stimulus was provided, but before exposure to any auditory stimulus. Apparently, the mere expectation of hearing an unattractive or attractive female was enough to facilitate or inhibit arousal.

Although differences were not statistically significant, an examination of Figure 1 suggests that A participants reached their peak level of arousal faster than did C participants and held this peak for a longer period, pointing toward a facilitative effect of attractive imagery. At the same time, C participants appeared to reach a higher level of peak arousal but for a shorter period, which might suggest they chose imagery that was highly arousing but required more effortful processing to create and/or maintain.

Given methodological limitations, findings from the present study should be interpreted prudently. In particular, because the dimensionality of self-reported sexual arousal is unclear, researchers taking this approach should be careful when using statistical methods that assume linearity or curvilinearity. The use of an 11-point ordinal scale to measure perceived arousal required respondents to make individual judgments about what each point on the scale meant, and the distance between the points in the scale may have differed between participants. Given the arbitrary nature of the measure, inferences about the pattern of arousal over time must be made with caution.

Although some studies have reported roughly concordant results between men's subjective self-reports and objective linear measures, such as plethysmography, these actually represent two separate components of sexual response, and there can be a fairly high amount of individual variability in this respect (Hall et al., 1985; Lang, Wincze, Zwick, Feldman, & Hughes, 1981; Wincze, Venditti, Barlow, & Mavissakalian, 1981). For instance, Hall and colleagues (1985) reported that there was greater agreement between objective and subjective measures at lower levels of arousal, and that longer latency in reporting peak subjective arousal predicted greater concordance between measures. If those findings are compared with the present results, it may be that the U group provided the most accurate account of their arousal. On the other hand, to the extent that the attractive stimuli produced more physiological arousal for the A group, attending to such physical cues could be sexually stimulating and increase concordance between the measures (Sakheim, Barlow, Beck, & Abrahamson, 1984).

Clearly, only using self-reports is not the "gold standard" and does not permit a completely continuous measure of arousal. However, multiple assessments protect the ratings from being confounded with the participants' memory of arousal, a shortcoming of the use of a one-time, poststimulus measurement (Hall et al., 1985). On a related note, the finding of approximate equivalency in peak arousal between groups, though unexpected, does provide some evidence against interpreting the results in terms of social desirability bias. Had either of these explanations been the primary influence on the results, one would expect participants in the U condition to have reported peak arousal at much lower levels compared with the other groups and not merely slower rates of arousal response.

Particular characteristics of the sample may have biased the results. The effectiveness of the stimuli used here would likely depend on participants' sexual preferences for women with a particular background or for certain types of erotica and/or sexual behavior. Such preferences may be heavily influenced by both cultural and individual experimental factors. Participants' age, high level of education, and the possibility of volunteer bias also complicate issues of generalizability. It is unknown at the moment whether similar results would be obtained from older men, particularly those who are married and may experience negative feelings when fantasizing about someone other than their partners (e.g., Cado & Leitenberg, 1990). However, it is likely that these relatively young individuals held more permissive views about sexuality and erotica than would older males in the general population (Lambe, 2004; Le Gall, Mullet, & Shafighi, 2002; Wilson, 1975). In previous research, men's level of higher education has also been related to the postponement of sexual initiation, more permissive sexual attitudes (Wilson, 1975) and less support for censoring erotic materials (Lambe, 2004). These qualities may have had special impacts upon participants volunteering for the study and their subsequent reactions. Finally, it is not clear to what extent the same pattern of results would arise in a female sample. Considering, however, that several prior studies have found visual imagery to be an especially important aspect of sexual fantasy and arousal for men (Byrne, 1977; Ellis & Symons, 1999; Gold & Gold, 1991; Przybyla et al., 1983; Smith & Over, 1987, 1988), the present study extends this literature.

Future research should replicate this study by using objective, physiological measures of arousal, alone or in conjunction with self-reports. Extending the research to female participants would also be informative in terms of the relative importance of visual imagery between the sexes, though considering that women's self-reports are also often inconsistent with physiological data (Heiman, 1975; Laan, Everaerd, van Bellen, & Hanewald, 1994) it would be especially important to include more objective measures. Finally, the findings of differential rates, but equivalent levels, of arousal could be further examined through alternate manipulations, such as overtly instructing participants to hold a mental image of the picture given to them, actually showing the image throughout the entire auditory stimulus presentation, or including other types of distractors. Comparisons of highly erotic auditory material with more content-neutral stimuli may also be useful.

The research conducted here provides support for Przybyla and colleagues' (Przybyla & Byrne, 1984; Przybyla et al., 1983) model of sexual arousal elicited from erotic auditory stimuli. The present research extends this model by demonstrating that linking an auditory stimulus with expectancy congruent or incongruent visual information stored in semantic memory results in differential rates of arousal response. Specifically, it appears that individuals not only produce their own imagery in absence of any overt visual stimuli but also that this imagery can be manipulated through providing prior visual cues, with subsequent effects on patterns of subjective sexual arousal.

Though the study was intended as a test and an extension of a theoretical model, the results are not without real-world implications. In the clinical field, the possibility for negative affect and/or intrusive thoughts to impair men's normative pattern of sexual response has been widely discussed in prior literature (e.g., Barlow, 1986; Hale & Strassberg, 1990; Koukounas & McCabe, 2001). Although no data were gathered as to whether the A and U groups differentially experienced visual imagery as voluntary or intrusive, further research should continue to explore the role of negative affective processing, particularly resulting from unwanted and unpleasant visual imagery, for inhibiting men's subjective and physiological sexual arousal. The findings point to the possible importance of imagery and mental processing related to physical attractiveness and suggest it might be useful for clinicians treating sexual difficulties to examine these processes in their cases and to educate clients about the potential effects of competing cognitive or emotional states on patterns of arousal. Further, technological advances have increased the frequency with which people engage in sexual encounters without actually seeing the other person involved, such as phone sex and chat-based cybersex. Anecdotally, we have observed that a description of physical characteristics (or a request for such information), either in word or in picture form, is one of the initial communications that occurs in sexual encounters in which no actual visual input is available. Although no formal research has described standard scripts for these types of interactions, the model of Przybyla and colleagues is useful in explaining how and why sexual arousal occurs in such situations.

References

Barlow, D. H. (1986). Causes of sexual dysfunction: The role of anxiety and cognitive interference. Journal of Consulting and Clinical Psychology, 54, 140-148.

Bartholow, B. D., Fabiani, M., Gratton, G., & Bettencourt, B. A. (2001). A psychophysiological examination of cognitive processing of and affective responses to social expectancy violations. Psychological Science, 12, 197-204.

Byrne, D. (1977). The imagery of sex. In J. Money & H. Musaph (Eds.), Handbook of sexology (pp. 327-350). Amsterdam: Excerpta Media.

Cado, C., & Leitenberg, H. (1990). Guilt reactions to sexual fantasies during intercourse. Archives of Sexual Behavior, 19, 49-63.

Dougher, M. J. (1995). Behavioral techniques to alter sexual arousal. In B. K. Schwartz & H. R. Cellini (Eds.), The sex offender: Corrections, treatment and legal practice (pp. 15.1-15.8). Kingston, NJ: Civic Research Institute.

Ellis, B. J., & Symons, D. (1999). Sex differences in sexual fantasy: An evolutionary psychological approach. Journal of Sex Research, 27, 527-555.

Gaither, G. A., & Plaud, J. J. (1997). The effects of secondary stimulus characteristics on men's sexual arousal. Journal of Sex Research, 34, 231-236.

Geary, D. C., Vigil, J., & Byrd-Craven, J. (2004). Evolution of human mate choice. Journal of Sex Research 41, 27-42.

Gold, S., & Gold, R. (1991). Gender differences in first sexual fantasies. Journal of Sex Education & Therapy, 17, 207-216.

Greenhouse, S. W., & Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24, 95-112.

Hale, V. E., & Strassberg, D. S. (1990). The role of anxiety on sexual arousal. Archives of Sexual Behavior, 19, 569-581.

Hall, K. S., Binik, Y., & di Tomasso, E. (1985). Concordance between physiological and subjective measures of sexual arousal. Behaviour Research and Therapy, 23, 297-303.

Harris, R., Yulis, S., & LaCoste, D. (1980). Relationships among sexual arousability, imagery ability, and introversion-extraversion. Journal of Sex Research, 16, 72-86.

Heiman, J. R. (1975). The physiology of erotica: Women's sexual arousal. Psychology Today, 8, 90-94.

Janssen, E., Everaerd, W., Spiering, M., & Janssen, J. (2000). Automatic process and the appraisal of sexual stimuli: Toward an information processing model sexual arousal. Journal of Sex Research, 37, 8-23.

Koukounas, E., & McCabe, M. P. (2001). Sexual and emotional variables influencing sexual response to erotica: A psychophysiological investigation. Archives of Sexual Behavior, 30, 393-408.

Laan, E., Everaerd, W., van Bellen, G., & Hanewald, G. J. F. P. (1994). Women's sexual and emotional responses to male- and female-produced erotica. Archives of Sexual Behavior, 23, 153-169.

Lambe, J. L. (2004). Who wants to censor pornography and hate speech? Mass Communication and Society, 7, 279-299.

Lang, J. D., Wincze, J. P., Zwick, W., Feldman, S., & Hughes, K. (1981). Effects of demand for performance, self-monitoring of arousal, and increased sympathetic nervous system activity on male erectile response. Archives of Sexual Behavior, 10, 443-464.

Le Gall, A., Mullet, E., & Shafighi, S. R. (2002). Age, religious beliefs, and sexual attitudes. Journal of Sex Research, 39, 207-216.

Masters, W. H., & Johnson, V. E. (1966). Human sexual response. Boston: Little, Brown.

Przybyla, D. P. J., & Byrne, D. (1984). The mediating role of cognitive processes in self-reported sexual arousal. Journal of Research in Personality, 18, 54-63.

Przybyla, D. P. J., Byrne, D., & Kelley, K. (1983). The role of imagery in sexual behavior. In A. A. Sheikh (Ed.), Imagery: Current theory, research, and application (pp. 436-467). New York: Wiley.

Reisberg, D., Culver, L. C., Heuer, F., & Fischman, D. (1986). Visual memory: When imagery vividness makes a difference. Journal of Mental Imagery 10, 51-74.

Roese, N. J., & Sherman, J. W. (2007). Expectancy. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (Vol. 2). New York: Guilford Press.

Sakheim, D. K., Barlow, D. H., Beck, G., & Abrahamson, D. J. (1984). The effect of increased awareness of erectile cues on sexual arousal. Behavior Research and Therapy, 22, 151-158.

Smith, D., & Over, R. (1987). Male sexual arousal as a function of the content and the vividness of erotic imagery. Psychophysiology, 24, 334-339.

Smith, D., & Over, R. (1988). Relationship between male sexual arousal and vividness of imagery in different sensory modalities. Australian Journal of Psychology, 40, 303-310.

Spiering, M., Everaerd, W., & Janssen, E. (2003). Priming the sexual system: implicit versus explicit activation. Journal of Sex Research, 40, 134-145.

Tokatlidis, O., & Over, R. (1995). Imagery, fantasy, and female sexual arousal. Australian Journal of Psychology, 47, 81-85.

Wilson, W. C. (1975). The distribution of selected sexual attitudes and behaviors among the adult population of the United States. Journal of Sex Research, 11, 46-64.

Wincze, J. P., Venditti, E., Barlow, D., & Mavissakalian, M. (1980). The effects of a subjective monitoring task on the physiological measure of genital response to erotic stimulation. Archives of Sexual Behavior, 9, 533-545.

(1) Due to differential drop-out, actual sample sizes in each group varied (see Participant section).

Skyler T. Hawk 

Ryan Tolman

Charles W. Mueller

University of Hawaii at Manoa

Correspondence should be addressed to Skyler T. Hawk, UvA Dept. Sociale Psychologie, Roeterstraat 15 1018 WB, Amsterdam, the Netherlands. E-mail: [e-mail]

Table 1. Repeated-Measures ANOVAs for Arousal Comparing All Groups

                                                               Partial
                        Wilke's                                 [[eta]
                        [lambda]        df           F         .sup.2]

Multivariate tests
  Time                    .27         10, 47       13.03 ***     .74
  Time x Group            .58         20,94         1.47
Within-Ss Effects (a)
  Time                             3.07, 172.15    37.78 ***     .40
  Time x Group                     6.15, 172.15     3.54 **      .11
  MSE                                               6.20
Within-Ss Contrastsb
  Time                                1, 56       109.89 ***     .66
  Time x Group                        2, 56         6.43 **      .19
  MSE                                               3.77
Between-Ss Effects                    2, 56         2.64
  MSE                                              38.21

Note. MSE = mean squared error.

(a) Interpreted with Greenhouse-Geisser. (b) Quadratic values.

** p [less than or equal to] .01. *** p [less than or equal to] .001.

Table 2. Post-Hoc Parallel Tests (Repeated-Measures ANOVA) for Arousal

                                 Wilke's
                                 [lambda]        df             F

Attractive Versus Unattractive
  Multivariate tests
    Time                           .26         10, 26         7.51 ***
    Time x Group                   .60         10, 26         1.76
  Within-Ss effects (a)
    Time                                     2.59, 90.65     17.46 ***
    Time x Group                             2.59, 90.65      5.45 **
    MSE                                                       8.19
  Within-Ss contrasts (b)
    Time                                        1, 35        61.62 ***
    Time x Group                                1, 35        11.56 **
    MSE                                                       3.39
  Between-Ss effects
    Attractive                                  1, 35         4.09 *
    Unattractive
    MSE                                                      32.71
Unattractive Versus No Picture
  Multivariate tests
    Time                           .27         10, 29         7.78 ***
    Time x Group                   .63         10, 29         1.68
  Within-Ss effects (a)
    Time                                    2.98, 113.38     36.77 ***
    Time x Group                            2.98, 113.38      2.76 *
    MSE                                                       5.25
  Within-Ss contrasts (b)
    Time                                        1, 38        52.52 ***
    Time x Group                                1, 38         8.85 **
    MSE                                                       3.93
  Between-Ss effects
    Unattractive                                1, 38         4.30 *
    No Picture
    MSE                                                      39.97
Attractive Versus No Picture
  Multivariate tests
    Time                           .22         10, 30        10.90 ***
    Time x Group                   .77         10, 30          .89
  Within-Ss effects (a)
    Time                                    3.29, 128.32     28.16 ***
    Time x Group                            3.29, 128.32      2.33
    MSE                                                       6.22
  Within-Ss contrasts (b)
    Time                                        1, 39       113.62 ***
    Time x Group                                1, 39          .087
    MSE                                                       3.94
  Between-Ss effects
    Attractive                                  1, 39          .032
    No Picture
    MSE                                                      41.44

                                 Partial
                                  [[eta]
                                 .sup.2]          M             SD

Attractive Versus Unattractive
  Multivariate tests
    Time                           .74
    Time x Group
  Within-Ss effects (a)
    Time                           .33
    Time x Group                   .14
    MSE
  Within-Ss contrasts (b)
    Time                           .64
    Time x Group                   .25
    MSE
  Between-Ss effects
    Attractive                     .11          3.62           .40
    Unattractive                                2.48           .41
    MSE
Unattractive Versus No Picture
  Multivariate tests
    Time                           .73
    Time x Group
  Within-Ss effects (a)
    Time                           .49
    Time x Group                   .07
    MSE
  Within-Ss contrasts (b)
    Time                           .58
    Time x Group                   .19
    MSE
  Between-Ss effects
    Unattractive                   .10          2.48           .45
    No Picture                                  3.73           .41
    MSE
Attractive Versus No Picture
  Multivariate tests
    Time                           .78
    Time x Group
  Within-Ss effects (a)
    Time                           .42
    Time x Group
    MSE
  Within-Ss contrasts (b)
    Time                           .74
    Time x Group
    MSE
  Between-Ss effects
    Attractive                     3.62          .45
    No Picture                     3.73          .41
    MSE

Note. MSE = mean squared error.

(a) Interpreted with Greenhouse-Geisser. (b) Quadratic values.

* p [less than or equal to] .05. ** p [less than or equal to] 01.
*** p [less than or equal to] .001.