Digging in my secret garden: disinhibitory effects of the "hidden observer" on reported sexual fantasies.
Article Type: Report
Subject: Sexual fantasies (Research)
Set (Psychology) (Influence)
Sexual behavior surveys (Methods)
Authors: Burris, Christopher T.
Mathes, Stefanie
Pub Date: 12/22/2011
Publication: Name: The Canadian Journal of Human Sexuality Publisher: SIECCAN, The Sex Information and Education Council of Canada Audience: Academic Format: Magazine/Journal Subject: Psychology and mental health Copyright: COPYRIGHT 2011 SIECCAN, The Sex Information and Education Council of Canada ISSN: 1188-4517
Issue: Date: Winter, 2011 Source Volume: 20 Source Issue: 4
Topic: Event Code: 310 Science & research Canadian Subject Form: Sexual behaviour surveys
Geographic: Geographic Scope: Canada Geographic Code: 1CANA Canada
Accession Number: 286557618
Full Text: Abstract: Self-report biases can affect the results of survey-based sex research, and techniques that minimize their effects are often cumbersome. In this study, 91 undergraduates were asked to submit a narrative of their favourite sexual fantasy in response either to standard instructions or on behalf of their "hidden observer" (HO), the ostensible part of them that knows their deepest secrets and will reveal them if questioned directly (Altemeyer, 1996, Hilgard, 1973). Compared to standard instructions, HO instructions yielded fewer denials, refusals, and/or apologies for the length or content of disclosed fantasies, which were in turn more erotically detailed and explicit. The findings suggest that the HO technique's potential as a user-friendly means of minimizing self-censoring in reports of sexual experiences, attitudes, and behaviours warrants further exploration.


Sexual fantasies are generally assumed to be a normal part of healthy adult sexual functioning, whereas their absence has been associated with sexual dissatisfaction and even sexual dysfunction (Cado & Leitenberg, 1990; Davidson & Hoffman, 1986; Leitenberg & Henning, 1995). Indeed, a paucity of sexual fantasies is currently regarded as one diagnostic indicator of hypoactive sexual desire disorder (American Psychiatric Association, 2000).

Although any two individuals may differ with respect to the frequency, explicitness, or thematic content of the sexual fantasies that they report, one must seriously consider the extent to which such differences reflect, for example, substantive differences in their experience or differences attributable to their respective comfort with such self-disclosure. The issue is far from trivial. As Leitenberg and Henning (1995) crisply noted: "Because sexual fantasies are covert, the only way to measure them is through what a person reports he or she is thinking". Third-party corroboration is impossible, and physiological measures cannot address fantasies' specificity: "There simply is no choice but to rely on self-report, with all its inherent methodological limitations regarding accuracy" (p. 470).

Understandably, then, researchers have been motivated to identify reliable means of minimizing the impact of non-substantive influences on sexual disclosures in research settings. For example, Mitchell et al. (2007) suggested that:

Elsewhere, Kahr (2008) has made a compelling case for sensitivity and acumen in the dialogic process of uncovering fantasy material in clinical interview settings.

Nevertheless, impression management concerns have been associated with lower self-reports of a variety of sexual fantasies and behaviours (based on rating responses to closed-ended measures) even under anonymous survey conditions (Meston, Heiman, Trapnell, & Paulhus, 1998). Moreover, a bogus pipeline procedure (a technique devised by social psychologists studying prejudice in which respondents are led to believe that untruthful statements can be detected by an apparatus applied to them--see Roese & Jamieson, 1993) has been shown to eliminate gender differences in self-reported sexual permissiveness that were evident in anonymous as well as public assessment conditions, suggesting that self-report biases can affect response patterns even when anonymity is assured (Alexander & Fisher, 2003).

Notwithstanding the bogus pipeline's apparent effectiveness in bypassing at least some of the self-report biases associated with sexual disclosures, the procedure is only practicable in an individual lab setting and is likely to be seen as ethically problematic (Roese & Jamieson, 1993). Needed is a technique that can decrease self-report biases more effectively than simple anonymity while being suitable for remote (i.e., non-laboratory) data collection. One such possibility is an adaptation of the "Hidden Observer" (HO) technique introduced by Hilgard (1973, 1986) as a means of assessing whether "some part" of a deeply hypnotized individual might be aware of events and experiences consciously denied by the individual. For example, although a hypnotized individual with an arm immersed in ice water may have denied experiencing discomfort, his/ her hidden observer often admitted experiencing at least some pain when questioned directly.

Altemeyer (1996) reasoned that the apparently evocative power of a purported hidden observer could be harnessed outside a hypnosis context as a means of encouraging more candid responses in anonymous surveys. As Altemeyer put it:

Indeed, paralleling Hilgard's results, Altemeyer demonstrated that right-wing authoritarian (RWA) individuals were much more likely to admit private religious doubts--a threatening confession for people who generally endorse a fundamentalist religious worldview--in their survey responses following an HO prompt versus standard survey instructions.

Although we are unaware of any other empirical demonstrations of the effects of this adapted HO technique beyond Altemeyer (1996), we conducted the present study to determine whether the technique might have a similarly disinhibitory effect on sexual fantasy disclosures. Participants were asked to write about their "absolute favourite sexual fantasy" prompted with either standard on-line instructions or HO-enhanced on-line instructions. Participants' text responses were subsequently content-coded so that we could examine the impact of the HO manipulation on the details they were willing to provide, as well as the language that they used when doing so (cf. Follingstad & Kimbrell, 1986; Moreault & Follingstad, 1978; Pelletier & Herold, 1988).

Paralleling the apparently disinhibitory effect of a posited hidden observer as revealed in RWAs' greater willingness to confess private religious doubts in an anonymous survey context (Altemeyer, 1996), we expected HO participants--relative to standard instructions participants--to show less inhibition in response to a direct request for a description of their absolute favourite sexual fantasy. Thus, in terms of the coded text variables, we hypothesized that HO participants would: (1) express less hesitance with respect to disclosing their favourite sexual fantasy; (2) provide more erotic details in their proffered fantasy narratives; and (3) use more explicit language in their proffered narratives. In order to show the specificity of the hypothesized disinhibitory effect of the HO manipulation, we also developed "neutral" analogues to the "erotic details" and "explicit language" categories--that is, details not directly connected to the erotic narrative, and the use of euphemistic language when referring to sexually intimate acts or body parts, respectively. We did not expect HO participants to average higher on either of these measures compared to standard instructions participants.


The sample consisted of 91 undergraduate student volunteers (49 women, 38 men, 4 not indicated) enrolled in one of several psychology courses offered at a large university located in southwestern Ontario, Canada. Each participant completed two sessions online, in a location of his/her choosing, in exchange for credit points in the applicable psychology course. The study was straightforwardly described as exploring "how and why variations exist in the sexual fantasies that people report." All participants were assured of their anonymity as part of the informed consent procedure (cf. Alexander & Fisher, 2003).

Participant characteristics

Participants provided demographic information in the first on-line session. Mean age was 19.27 years (SD = 1.65 years; range 17-25). Forty-six percent were Caucasian and 33% were East Asian, with the remaining 21% representing Middle Eastern, Hispanic, South Asian, and African/Caribbean backgrounds; 36% were born outside of Canada. No attempt was made to solicit or screen participants on the basis of sexual orientation. Separate, preliminary analyses of variance (ANOVAs) were performed on the coded outcome variables (described below) that included gender and the gender x manipulation interaction. These preliminary findings revealed no significant effects involving gender, so the latter variable was excluded from the reported analyses. (Given the novel nature of the coding scheme, we were unaware of any previous research that would provide a compelling basis for either expecting or not expecting gender effects.)

Pretest confirmation of groups" equivalence on a sexuality-related variable

In order to assess the effectiveness of the random assignment procedure employed as part of the experimental sessions (described below), it seemed prudent to first compare the pre-experimental mean scores of the standard instructions group and HO group participants on a conceptually relevant variable. Thus, in a pretest session, participants completed the 5-item Sexual Opinion Survey-Short Form, a brief measure of erotophobia/erotophilia reflecting broad-based negative or positive attitudes toward sex-related activities (see Semph, 1979, as cited in Rye, Meaney, & Fisher, 2011; and Fisher, Byrne, White, & Kelley, 1988). A 7-point Likert-type response format was utilized (1 = strongly agree to 7 = strongly disagree) to assess responses to statements such as "Masturbation can be an exciting experience". Items were reverse-coded as appropriate so that higher scores indicated greater erotophobia. Cronbach's alpha was a rather modest .60. Two participants failed to complete this measure.

Second (experimental) session

No sooner than forty-eight hours after completion of the first session, participants were scheduled for a second on-line (experimental) session. After indicating their consent to participate electronically, the program randomly assigned participants to one of two instructional sets whereby sexual fantasy text accounts were solicited. Specifically, the standard instructions to participants (n = 47) read as follows:

In contrast, the instructions to HO participants (n = 44) read (of. Altemeyer, 1996):

Participants in both the standard instructions and HO groups subsequently typed their responses in a text box and submitted them.


As part of the post-session debriefing protocol, participants read a feedback letter that explained in detail the conceptual background, goals, hypotheses, and methodological details of the study, including the HO manipulation. Subsequently, they were asked to e-sign a post-debriefing consent form indicating their basic understanding of the study; all did. Participants were also given the option of requesting that their data be withdrawn from further consideration; none did. Thus, due diligence was exercised with respect to minimizing any adverse effects of the experimental procedure.


Preliminary analyses of erotophobia/ erotophilia scores

With a possible range of 5-35, the sample means were near the theoretical midpoint on the erotophobia/ erotophilia measure: 17.78 (SD = 4.13) for men and 21.42 (SD = 5.76) for women. Therefore, notwithstanding the significant gender difference, t (85) = 4.04, p < .001, the sample appeared neither extremely erotophilic nor extremely erotophobic overall. Moreover, mean erotophobia/erotophilia scores did not differ by experimental condition (Ms = 19.44 and 19.95 for the standard instructions and hidden observer [HO] conditions, respectively), t < 1. It would appear, therefore, that our random assignment did in fact equate the two conditions on at least one pre-existing variable that could arguably have influenced participants' responsiveness to the request for disclosure of their absolute favourite sexual fantasy.

Coding text responses to the fantasy prompts

Blind to condition and based on a number of thematic categories, the second author content coded participants' text responses. All words/phrases representing distinct units of meaning relevant to a given category were tallied such that each participant received a tally score for each category described below. (All text examples have been corrected for grammar and spelling.)

Hesitation index

As a gauge of participants' reluctance or sheepishness with respect to disclosing the details of their absolute favourite sexual fantasy, we computed a hesitation index by tallying all distinct occurrences of apologies, denials, and refusals. For example, participants sometimes apologized for the writing style, brevity, or thematic conventionality of their proffered fantasies, as evidenced in the following quotes: "I am sorry, I do not know much about it;" "Sorry about the disorganized manner of my descriptions;" "This was quite difficult to write;"; and "Not much else I can say." Examples of denial- or refusal-related quotes include: "I don't have one specific fantasy;" "I don't have super explicit fantasies;" "I do not have any sexual fantasies at this point in my life;" "My fantasies do not tend to get very detailed;" "I'm too naive to understand when my friends make sexual jokes;" and "I'm wondering what kind of person conducts a study like this?"

Erotic details index

As a gauge of the degree to which participants were willing to construct an explicit, detailed sexual narrative in their text responses, we computed an index of erotic details by tallying all distinct occurrences of: passionate and/or overtly sexual actions (e.g., "I move on to his lip and we both moan;" "we would switch positions"); affectionate, romantic actions such as cuddling and "pillow talk" (e.g., "we position ourselves in spooning position;" "afterwards we cuddle and kiss extensively"); and references to the specific temporal and spatial context of both types of actions (e.g., "starts in the daytime;" "we are at her place, in her bed").

Explicit language index

As an additional index of explicitness, we tallied all distinct occurrences of non-euphemistic sexual language. This included both technical (e.g., "vagina/ penis," "thrusting") and vulgar or slang terms (e.g., "eating her out," "shooting his load," "blow job").

Peripheral details and euphemistic language

Finally, we computed two additional indices that were intended to serve as neutral comparisons for the erotica and explicit language indices. We tallied all distinct details referring to non-romantic/nonsexual activities that were peripheral to the actual erotic narrative (e.g., "It's a hot summer day and we've spent the day hiking," "I decided to drink until my senses were all woozy," "We looked out to the beautiful view"). Separately, we tallied all distinct occurrences of euphemistic language, whereby reference was made to sexual activities without the use of technical, vulgar, or slang terms (e.g., "hands wandering down;" "he services me;" "become intimate;" "everything gets done there").

Coding reliability

A second coder, also blind to condition, was first trained on a randomly selected initial subset of fantasy texts (n = 10), and then coded a second randomly selected subset (n = 10). Because the lowest inter-rater correlation based on any coded index for the latter subset was .98, we concluded that the coding system was reliable.

Hypothesis testing

Means on the five coded outcome variables as a function of instructional set appear in Table 1. As expected, the "Hidden Observer" (HO) instructions-relative to the standard instructions (SI)--virtually eliminated refusals, denials, and apologies pertaining to the disclosure of sexual fantasies, as measured by the hesitation index. HO participants also provided more detailed narratives of their fantasized sexual/ erotic activities, with more frequent graphic forms of expression (i.e., explicit language), in their responses relative to standard instructions participants. Importantly, both of the latter effects remained statistically significant (p < .05) even when the analysis was limited to those who scored "0" on the hesitation index (and who therefore, by definition, unapologetically provided a sexual fantasy narrative). In contrast, scores on the peripheral details and euphemistic language indices did not differ by condition, indicating the specificity of the HO effect.

Moreover, testifying to the validity of our coding categories, greater self-reported erotophobic tendencies predicted lower scores on the erotic details and explicit language indices overall, r (87) = -.21, p < .05, and r (87) = -.33, p = .002, respectively, whereas erotophobia/erotophilia did not predict scores on the peripheral details or euphemistic language indices (rs = .05 and .04, respectively). The correlation between erotophobia/erotophilia and scores on the hesitation index was not significant, r (87) = -.04, perhaps because erotophobic individuals might be more likely to deny having and/or refuse to disclose an absolute favourite sexual fantasy, but less likely to apologize for the brevity or conventionality of their fantasies if they chose to disclose them (cf. Fisher et al., 1988). None of the erotophobia/erotophilia correlations differed significantly by condition, and the effects of the HO manipulation on hesitation, erotic details, and explicit language use reported above remained significant (p < .05) when controlling for erotophobia/erotophilia.


As hypothesized, participants who were prompted to disclose their absolute favourite sexual fantasy on behalf of their "hidden observer" (HO) were more cooperative and/or less apologetic relative to participants prompted by standard disclosure instructions. HO participants' proffered fantasies were also more explicit in terms of both content and diction. Clearly, the HO manipulation had an effect. How should this effect be interpreted?

It could perhaps be argued that HO participants' greater willingness to provide erotically explicit narratives was the result of confabulation or fabrication and that participants' responses to the more basic disclosure instructions were in fact "truer" or more "honest" overall. Although this interpretation cannot be ruled out definitively based on the present data, its parsimony can be seriously challenged. We note that all participants, regardless of condition, were specifically instructed to write about their "absolute favourite sexual fantasy," and that the on-line instructions introducing a presumed hidden observer that has direct access to participants' "deepest intimate information" framed the HO as an internal, not an external, audience. Moreover, HO participants were not in a hypnotic state induced by a physically present researcher (cf. Hilgard, 1973; 1986) but simply engaged in an anonymous on-line survey. Thus, any additional experimental demand specifically induced by the HO instructions arguably involved participants' accountability to themselves, not to the researchers. Consequently, neither confabulation nor fabrication strikes us as a particularly compelling interpretation of HO effects.

Instead, we suggest that the effects of the HO prompt in our study parallel Altemeyer's (1996) evoked admissions of private religious doubts among his right-wing authoritarian (RWA) respondents. Given the strong link between RWA and religious fundamentalism, such admissions were likely seen as socially undesirable, even personally threatening, by those respondents, so it seems rather unlikely that HO instructions would prompt them to make a "false confession" concerning religious doubts in an anonymous survey to appease a researcher. Like religious doubts, sex-related personal information--such as fantasy content--is often regarded as deeply private and not readily disclosed in routine circumstances (e.g., Kahr, 2008; Mitchell et al., 2007). Thus, we would suggest that the more parsimonious interpretation of our findings is that--paralleling previously observed effects of "truth-telling" prompts in non-sexual (e.g., Roese & Jamieson, 1993) as well as sexual domains (Alexander & Fisher, 2003)--the HO prompt disinhibited participants, so that they were more willing to disclose the details of their favourite sexual fantasy as instructed.

Study limitations and concluding observations

Of course, as with any single study, sample limitations must be acknowledged. In the present case, participants were age-traditional Canadian undergraduates who volunteered for a sexual fantasy study. Notwithstanding the limited age range, more than half of the participants were non-Caucasian and over one-third were born outside of Canada, so the sample was rather diverse in the latter respects. Moreover, although it is certainly possible that self-selection into a study concerning sexual fantasy constrained our sample's variability with respect to unmeasured variables such as sex guilt (cf. Leitenberg & Henning, 1995), our sample appeared neither excessively erotophilic nor erotophobic, at least based on the modestly reliable brief measure used here. Moreover, we demonstrated empirically that erotophobic/erotophilia could not account for the differences observed in text responses to the fantasy prompt as a function of the HO manipulation.

Overall, then, the present study's results point suggestively toward the potential of HO instructions as a simple technique for reducing "false negatives" on sex-related self-report variables. Beyond a purely pragmatic standpoint, the question of exactly why the HO technique works remains intriguing: Is it the case that the threatening nature of sensitive self-disclosures is reduced because it is not "the self" but "the HO" that is disclosing (cf. Altemeyer, 1996)? Alternatively, does the HO function more like an internal bogus pipeline that holds the self accountable for the truthfulness of one's claims (cf. Alexander & Fisher, 2003; Roese & Jamieson, 1993)? Pinpointing the mechanism behind the HO technique's ostensible disinhibitory properties remains a worthwhile goal for future research. In addition, a more mundane, but no less important, goal of future research is to determine the robustness of the effect described herein--whether, for example, HO instructions would decrease the likelihood of underreporting (especially socially sensitive) sexual behaviours whose frequency could be independently verified, for example. If this proves to be the case, then the HO technique should interest any sex researcher whose investigations are to any extent reliant upon self-report.

Indeed, to the extent that the disinhibitory effect of HO instructions can be shown to be robust across a breadth of sexual self-report domains, the implications for not only conducting future research, but also for interpreting past research, will be considerable. For example, sex guilt/erotophobia has been shown to be inversely correlated with self-reports of sexual fantasies involving force or coercion directed toward the self (Strassberg & Lockerd, 1998). In light of the results of the present research, we must ask whether this inverse relationship would emerge if respondents were prompted with HO instructions--and if not, how might this challenge our collective understanding of the mechanisms underlying sex guilt/erotophobia? Likewise, induced relationship threats have been shown to affect the thematic content of sexual fantasies under standard disclosure conditions (Birnbaum, Svitelman, Bar-Shalom, & Porat, 2008). Would disinhibitory HO instructions intensify such effects? Because we did not focus on thematic fantasy content in the present research, we ourselves cannot answer such questions. Nevertheless, our results suggests that the "Hidden Observer" technique may offer considerable promise as a tool for digging deep in the "secret garden" (Friday, 1973) of sexual fantasy.

Acknowledgements: Our thanks to Armand Munteanu and Kristina Schrage for assistance with data collection and content coding and to John Rempel for helpful comments. Portions of this paper were presented at the 2011 World Congress for Sexual Health, Glasgow, UK.


Alexander, M.G., & Fisher, T.D. (2003). Truth and consequences: Using the bogus pipeline to examine sex differences in self-reported sexuality. Journal of Sex Research, 40, 27-35. doi: 10.1080/00224490309552164

Altemeyer, B. (1996). The authoritarian specter. Cambridge, MA: Harvard University Press.

American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental Disorders (4th Ed., text rev.). Washington, DC: Author.

Birnbaum, G.E., Svitelman, N., Bar-Shalom, A., & Porat, O. (2008). The thin line between reality and imagination: Attachment orientations and the effects of relationship threats on sexual fantasies. Personality and Social Psychology Bulletin, 34, 1185-1199. doi: 10.1177/0146167208319692

Cado, S., & Leitenberg. H. (1990). Guilt reactions to sexual fantasies during intercourse. Archives of Sexual Behavior, 19, 49-63. doi: 10.1007/BF01541825

Davidson, J.K., & Hoffman, L.E. (1986). Sexual fantasies and sexual satisfaction: An empirical analysis of erotic thought. Journal of Sex Research, 22, 184-205. doi: 10.1080/00224498609551299

Fisher, W.A., Byrne, D., White, L.A., & Kelley, K. (1988) Erotophobia-erotophilia as a dimension of personality. Journal of Sex Research, 25, 123-151. doi: 10.1080/00224498809551448

Follingstad, D.R., & Kimbrell, C.D. (1986). Sex fantasies revisited: An expansion and further clarification of variables affecting sex fantasy production. Archives of Sexual Behavior, 15, 475-486. doi: 10.1007/ BF01542312

Friday, N. (1973). My secret garden: Women's sexual fantasies. New York: Simon and Schuster.

Hilgard, E.R. (1973). The domain of hypnosis, with some comments on alternative paradigms. American Psychologist, 28, 972-982. doi: 10.1037/h0035452

Hilgard, E.R. (1986). Divided consciousness: Multiple controls in human thought and action. New York, NY: Wiley.

Kahr, B. (2008). Who's been sleeping in your head?: The secret world of sexual fantasies. Toronto, ON: Viking Canada.

Leitenberg, H., & Henning, K. (1995). Sexual fantasy. Psychological Bulletin, 117, 469-496. doi: 10.1037/0033-2909.117.3.469

Meston, C.M., Heiman, J.R., Trapnell, P.D., & Paulhus, D.L. (1998). Socially desirable responding and sexuality self-reports. Journal of Sex Research, 35, 148-157. doi: 10.1080/00224499809551928

Mitchell, K., Wellings, K., Elam, G., Erens, B., Fenton, K., & Johnson, A. (2007). How can we facilitate reliable reporting in surveys of sexual behaviour? Evidence from qualitative research. Culture, Health, and Sexuality, 9, 519-531. doi: 10.1080/13691050701432561

Moreault, D., & Follingstad, D.R. (1978). Sexual fantasies of females as a function of sex guilt and experimental response cues. Journal of Consulting and Clinical Psychology, 46, 1385-1393. doi: 10.1037/0022-006X.46.6.1385

Pelletier, L.A., & Herold, E.S. (1988). The relationship of age, sex guilt, and sexual experience with female sexual fantasies. Journal of Sex Research, 24, 250256. doi: 10.1080/00224498809551420

Roese, N., & Jamieson, D. (1993). Twenty years of bogus pipeline research: A critical review and meta-analysis. Psychological Bulletin, 114, 363-375. doi: 10.1037/0033-2909.114.2.363

Rye, B.J., Meaney, G.J., & Fisher, W.A. (2011). Sexual Opinion Survey. In T.D. Fisher, C.M. Davis, W.L. Yarber, & S.L. Davis (Eds.), Handbook of sexuality-related measures, 3rd Ed. (pp. 231-236). New York, NY: Routledge.

Semph, M.E. (1979). Emotional orientation towards sexuality: Its relation to expecting and perceiving contraceptive side effects. Unpublished honours thesis. University of Western Ontario, Department of Psychology, London, Ontario, Canada.

Strassberg, D.S., & Lockerd, L.K. (1998). Force in women's sexual fantasies. Archives of Sexual Behavior, 27, 403-414. doi: 10.1023/A:1018740210472

Christopher T. Burris (1) and Stefanie Mathes (2)

(1) Department of Psychology, St. Jerome's University, Waterloo, ON

(2) Department of Psychology, McGill University, Montreal, PQ

Correspondence concerning this article should be addressed to Christopher T. Burris, Department of Psychology, St. Jerome's University, Waterloo, ON N2L 3G3. E-mail: cburris@uwaterloo.ca
commitment to the aims of the research;
   belief in the legitimacy of the survey;
   assurances of confidentiality; a professional
   approach on the part of the interviewer; and
   perceptions of the therapeutic benefit of
   disclosure can facilitate ease and accuracy
   of disclosure of sensitive [sex-related]
   information" (p. 526, brackets inserted).

Hilgard considered the Hidden Observer a
   subconscious process that knows the real
   truth. I saw it as a way to allow people,
   hypnotized or not, to admit things they could
   not admit otherwise: they do not spill the
   beans; the Hidden Observer does (p. 137).

Please bring to mind your absolute favourite
   sexual fantasy. Once you have it in mind,
   please describe it in as much detail as
   possible, using whatever words seem
   appropriate to you. Give details regarding the
   setting, the activities and sequence of events,
   your role and the role of any other actors,
   plus any other details that come to mind,
   regardless of how random they may seem.

Imagine that there exists a part deep
   within you that you could call the "hidden
   observer". This part of you knows every
   detail of all behaviour and thought processes
   that go on within you and is aware of
   the deepest intimate information about
   yourself. However, it cannot be asked
   to talk about this intimate information
   unless it is allowed by you to do so.

   Now, please ask your "hidden observer"
   to bring to mind your absolute favourite
   sexual fantasy. Once your "hidden observer"
   has this fantasy in mind, please ask your
   "hidden observer" to describe it in as
   much detail as possible, using whatever
   words seem appropriate. Ask your "hidden
   observer" to give details regarding the
   setting, the activities and sequence of
   events, your role and the role of any other
   actors, plus any details that come to mind,
   regardless of how random they may seem.

Table 1 Hesitation to disclose and content-related features of
self-reported sexual fantasies as a function of instructional

                     Standard       Hidden
                   Instructions    Observer    t      p
                       (SI)          (HO)
                     (n = 47)      (n = 44)

Hesitation            .21             .02     2.45   .02
                     (.51)           (.15)

Erotic Details      10.11           19.80     2.39   .02
                    (9.50)         (26.01)

Explicit             1.74            3.73     2.13   .04
Language            (2.55)          (5.82)

Peripheral           2.49            1.02      .85    ns
Details            (11.62)          (1.83)

Euphemistic          1.02            1.07      .17    ns
Language            (1.29)          (1.34)

Note. Standard deviations appear in parentheses. Significance
levels are based on two-tailed tests, equal variances not
assumed; significance levels are unchanged if equal variances are
Gale Copyright: Copyright 2011 Gale, Cengage Learning. All rights reserved.