Response Bias on Self-Report Measures of Sexual Fantasies Among Sexual Offenders

Abstract

The impact of sexual fantasies in future risk and treatment response among sexual offenders has long been known. However, as we develop objective self-report measures of sexual fantasies, response bias is becoming an increasing concern. In examining a sample of institutionalized sex offenders, the present study suggests that offenders’ responses on these measures are prone to response bias, the bias does not negate their associations with other self-report measures of sexual deviance, and relationship of their sexual fantasies does not appear to relate to actual behavioral indications. Clinical and research implications for these findings are discussed.

Keywords

sexual fantasies sex offenders response bias

The assessment of sexual recidivism remains a prominent area of research given the increasing focus on sex offenders by the legal system and the legislatures (Cortoni, 2009). Although a multitude of measures and isolated predictors exist to assess sexual recidivism, analyzing the correlates of actual sexual offenses has been a challenging process. As it has become clear that shortcomings may exist in our current approaches (Campbell, 2007; Hart, Michie, & Cooke, 2007), research has attempted to improve the methods designed to assess risk and the most important individual risk factors (Beech, Craig, & Browne, 2009). Specifically, sexual interest in deviant behavior has been identified as a strong predictor of recidivism (Hanson & Morton-Bourgon, 2005). One way in which psychologists have long assessed sexually deviant behavior is by examining offenders’ sexual fantasies (Bartels & Gannon, 2011: O’Donohue, Letourneau, & Dowling, 1997). Evidence suggests that sexual fantasies can provide evidence for unreported desires that can hinder treatment and lead to re-offense (Hanson & Morton-Bourgon, 2005; Howitt, 2004; Woodworth et al., 2013). However, there has been concern about the manner in which sexual fantasies are assessed and the degree to which reliable and valid indications of these phenomenon can be obtained in clinical practice (Baumgartner, Scalora, & Huss, 2002).

Leitenberg and Henning (1995) defined sexual fantasies as mental images, occurring while a person is in a conscious state of mind, about sexual experiences that lead to sexual arousal. Sexual fantasies can be experienced by offenders and non-offenders and are considered to be a normal sexual experience for all types of persons (Carnes, 2001; Leitenberg & Henning, 1995). However, non-offenders are able to contain their sexual desires and fantasies in their mind and instead concentrate on the real relationships around them. Others make the decision to put into action or force these fantasies on others. The manner and motivation in which these sexual fantasies are experienced are a significant factor for reoffending in a sex-offender population (Howitt, 2004).

The role of sexual fantasies in sexual offending has been much debated. Carnes, in 1983, may have been one of the first to publicize sexual addiction and sexual fantasies (see Delmonico & Griffin, 2002). His publication sparked an outbreak of research on sexual addiction and its consequences to reduce deviant sexual acts (Carnes, 2001). Carnes (2001) posited that a client could be excited about his or her pornographic tendencies in sexual fantasies, or a client could be disgusted and shamed by his or her pornographic sexual fantasies. The manner in which the person allows sexually deviant behaviors, including fantasies, to consume his life, and his feelings toward the level of obsession, can factor into his cycle of sexual compulsivity and the possibility of recidivism (Carnes, 2001). Carnes (2001) further argued that once the client reaches heightened levels of sexual compulsivity, the negative consequences of unconventional sex and sexual addiction destroy the normality of a person’s life. Relationships are then affected by the isolation from separating one’s sexual deviancy from the intimate relationships that are in place. As a person isolates (whether to keep his or her pleasure a secret or to isolate in shame due to his or her preference for this type of sexual encounters), the relationships around him or her deteriorate (Carnes, 2001). Sexual fantasies become part of the isolation and become a factor leading to sexual deviance. One’s mind can be so engaged with sexual fantasies that these fantasies become an illusion of one’s reality and encourage hypersexuality and sexual behavior (Carnes, 2001).

Experiencing repeated sexual fantasies among sexual offenders commonly leads to hypersexuality (Leedes, 1999). Furthermore, these persons experiencing hypersexuality often have less relational sexual fantasies and, instead, tend to objectify these sexual fantasies (Carnes, 2001). As a result, the person who separates sexual fantasies from intimacy may be very comfortable acting out sexual fantasies with those without relations or interest and acts out frequently, becoming more sexually compulsive (Carnes, 2001). In particular, the role of sexual fantasy provides an interesting insight into offender motivation. Sometimes, offenders report the sexual fantasies, especially deviant fantasies, as more arousing than the act itself (Howitt, 2004). However, because the fantasy itself is so arousing, this reality can be the driving factor and motivation to actually perpetrate these fantasies or some type of sexually deviant behavior (Howitt, 2004). The debate is not whether sexual fantasies are important to offending; the debate is about the process by which sexual fantasies may lead a person to sexually offend.

Unfortunately, the assessment of sexual fantasies is difficult, and many complain about the face validity of these measures and that response bias may pose difficulties for treatment (Baumgartner et al., 2002). These difficulties may change the focus of treatment from examining an offenders’ overall psychopathology to specifying areas that affect that person sexually (Huss, 2014). Nonetheless, self-report measures remain a widely used method to assess sexual deviance and specifically sexual fantasies because of their ease of administration (Adherhold, Huss, & Palarea, 2011). By sexual deviance, we use Hanson and Morton-Bourgon’s (2005) definition that includes deviant sexual interests in paraphilias, children, rape, and other sexual preoccupations.

Even further complicating the process is the limited amount of research available examining the systematic collection of sexual fantasies. In fact, there are few commonly recognized self-report measures, the Sex Fantasy Questionnaire (SFQ; O’Donohue et al., 1997) and the Wilson Sex Fantasy Questionnaire (WSFQ) created in 1978 (Wilson, 1988) being the most long-standing. The WSFQ was first created to assess and standardize sexual fantasies (Wilson, 1988). The WSFQ measures sexual fantasies and attempts to distinguish normal from paraphilic (abnormal; O’Donohue et al., 1997). The other primary self-report questionnaire is the SFQ. The designers of the SFQ criticized the WSFQ as too vague and deficient in reliability (O’Donohue et al., 1997). The SFQ was designed to address the weakness of the WSFQ. In addition, the SFQ also focuses more on pedophilic fantasies rather than paraphilic in general (O’Donohue et al., 1997).

Again, these questionnaires have been criticized because sex offenders tend to underreport their sexual fantasies (Baumgartner et al., 2002). Sex offenders have systematic reasons for appearing socially desirable in the treatment and correctional context (Mills & Kroner, 2006). Therefore, it is critical to assess the degree to which sexual offenders actually exhibit response bias on self-report measures of sexual fantasies.

As a result of the continued criticism about the face validity of these specific self-report measures, this study sought not only to examine potential response bias with these measures but also examine their relationship with other behavioral and self-report measures of deviance. By examining other variables related to self-report measures of sexual fantasies, we hoped to provide concurrent validity for the measures in addition to assessing their tendency toward response bias. The issues for us are whether self-report measures of sexual fantasies are still valid indicators of sexual deviance, despite any tendency for respondents to underreport their level of sexual fantasy. For this study, we chose measures of sexual compulsivity and sexual sensation seeking to assess convergent validity because of their association with sexual deviance, and that they approached sexual deviance from different underling foundations (Williams, Scalora, & Huss, 1998). We believe there will be significant associations between each of the sexual fantasy measures and the measures of sexual deviance (sexual compulsivity and sexual sensation seeking), behavioral correlates, and a measure of social desirability.

Method

Participants

Our sample consisted of a group of offenders admitted to a state forensic facility (n = 469). The average age of these offenders was 34.5 (SD = 11.51) years, with the majority Caucasian (77.7%), followed by African American (15.7%), Hispanic (3.6%), Native American (1.9%), and Asian American or Other (1%). Offenders were diagnosed with dozens of different mental illnesses: pedophilia (n = 191), some specific form of schizophrenia (n = 56; that is, paranoid schizophrenia [n = 34], undifferentiated schizophrenia [n = 11], schizoaffective disorder [n = 6]), paraphilia not otherwise specified (NOS; n = 18), psychotic disorder NOS (n = 13), and alcohol abuse (n = 13) as the most common Axis I disorders. The most frequent Axis II disorders were personality disorder NOS (n = 88), antisocial personality disorder (n = 50), borderline intellectual functioning (n = 18), and mild mental retardation (n = 10). The current sample was further broken down by offense category with most of the participants being sexual offenders (n = 302). For purposes of the present article, we will only be focusing on the sexual offenders in this sample.

Procedure

Researchers collected demographic information on offenders from institutional records. Demographic information included extensive details concerning each offender’s mental health, medical history, criminal history, familial history, and sexual history. Offenders also responded to various psychological measures at the time of admission. The measures used in this study were the WSFQ, the SFQ, the Marlowe–Crowne Social Desirability Scale (MCSDS), Sexual Compulsivity Scale (SCS), and Sexual Sensation Seeking Scale (SSSS). All Institutional Review Board (IRB) regulations were adhered to in this study. Participants gave consent on admission for their information to be used for research purposes.

Materials

The MCSDS is a 33-item scale that has been used as a measure in attempts to control for response bias among a variance of psychological studies that utilize self-report measures. It began as a 50-item scale but was reduced to 33 items during primary testing by Marlowe and Crowne (Paulhus, 1991). The possible scores one may attain range from 0 to 33 with higher scores indicating a greater need for social approval (Paulhus, 1991). In the original testing in 1964, there was a reported mean of 15.5 (SD = 4.4) in an average group of young adults (Paulhus, 1991). There have been several other studies representing insignificant variations to this mean over time (Paulhus, 1991). Concerning reliability, the alpha coefficients in these studies ranged from .73 to .88 (Paulhus, 1991). In the original sample for Marlowe and Crowne, the reliability rating resulted in a .88 over 1 month for test–retest (Paulhus, 1991). In terms of validity, this scale was created to counteract social desirability for self-reports, and while showing only moderate correlations to original measures (such as Edwards Social Desirability Scale and Self-Deceptive Enhancement), it is the only scale independent of psychopathology (Paulhus, 1991). Cronbach’s alpha for the current sample was .84 for the Marlowe–Crowne.

Some of the typical questions used in the Marlowe–Crowne show simple scenarios that a person experiences more than once that pertain to a somewhat shameful decision or behavior (Crowne & Marlowe, 1960). For example, some of the situations in the scale are as follows: I never hesitate to go out of my way to help someone in trouble, I like to gossip at times, I am willing to admit when I make a mistake, and I am always courteous even to people who are disagreeable (Crowne & Marlowe, 1960). When responding to these situations, people will have a tendency to negate their behavior, when we all have engaged in behavior not considered ideal as far as these situations are concerned.

The WSFQ was first created to have an instrument to assess and standardize sexual fantasies (Wilson, 1988). The WSFQ contains 40 items that are split equally into 10 items per category. These four categories are fantasy subscales: Exploratory (i.e., exchanging partners or participating in an orgy), Intimate (i.e., kissing a loved one), Impersonal (i.e., engaging sexually with a stranger), and Sadomasochistic (i.e., forcing sex). The four fantasy subtypes and their items are used to determine where in the spectrum these fantasies lie through scoring—varying from normal to deviant. In a representative sample, strong correlations existed between the type of sexual behavior reported by the participants and the higher scores on the fantasy questionnaire, indicating that more fantasies led to more of these behaviors (Wilson, 1988). Similar strengths of reliability and validity have been found across different studies, representing the internal consistency and reliability of this measure (Wilson, 2010). One study recognizing its reliability, Skovran, Huss, and Scalora (2010), found reliable Cronbach’s alpha coefficients for each of the four subscales of the WSFQ: .84 for Exploratory, .92 for Intimate, .77 for Impersonal, and .81 for Sadomasochistic. This measure is known to display reliable results specifically among the male population (Baumgartner et al., 2002). The validity and reliability of this questionnaire have also been shown through its ability to be used cross-culturally and still delivering consistent and valid results (Wilson, 2010). Cronbach’s alpha for the current sample was .94 for the entire measure, .82 for the Exploratory subscale, .91 for the Intimate subscale, .78 for the Impersonal subscale, and .82 for Sadomasochistic subscale.

The SFQ is a scale that contains 90 items. The SFQ zones in on individual fantasies versus four broader subtypes such as the WSFQ (Skovran et al., 2010). Each item on the SFQ is answered with regard to frequency by the respondent: a one indicating an item never occurs, a two indicating an item occurs sometimes, or a three indicating an item occurs frequently (O’Donohue et al., 1997). Regarding reliability among the student and offender population in this primary study, Cronbach’s alpha coefficient ranged from .59 to .92—calculating .82 as the mean of the various scales (O’Donohue et al., 1997). In addition, test–retest coefficients averaged a mean of 1.84 and ranged from .72 to .84 (O’Donohue et al., 1997). Regarding validity of the SFQ, this measure showed results that it was valid due to the fact it measured what the test attempted to measure. In this primary study by O’Donohue et al. (1997), one of the hypotheses was that the percentage of participants who were sex offenders would have a higher score on the child scale of the SFQ than the student sample. This hypothesis was displayed in the original results (O’Donohue et al., 1997). The SFQ’s content is also considered to be valid. This was tested by originally containing an extra space where a respondent could write in any additional sexual fantasies, and it was determined that the SFQ asked about variations of the fantasies written by respondents, with the exception of 5.5% of the respondents’ answers (O’Donohue et al., 1997). Cronbach’s alpha for the current sample was .96. These results indicate the SFQ and its content are both valid.

The SCS is a measure to assess high-risk sexual behaviors, either a preoccupation with sexual acts or high frequency of deviant sexual acts (McBride, Reece, & Sanders, 2008). The SCS is a 10-item measure, where participants respond in the same manner as the SSSS, on a scale of 1 to 4, with a 4 being most like the participant (Gaither & Sellbom, 2003). The scores of this measure, therefore, range between 10 and 40, with a higher score showing a greater proclivity to be preoccupied with sexual acts and a lower score showing more normal preoccupation with sexual behavior (McBride et al., 2008). This scale maintains internal consistency with a Cronbach’s alpha coefficient of .84 in an average college-aged sample (Gaither & Sellbom, 2003). The SCS has displayed strong internal consistency across many studies, with Cronbach’s alpha coefficients ranging between .79 and .87 (McBride et al., 2008). Also, there were strong correlations between high SCS scores and sexual behaviors reported by these participants, representing strong validity (McBride et al., 2008). Cronbach’s alpha for the current sample was .92 for the entire measure.

The SSSS is a measure used to evaluate the desire to seek and engage in risky sexual encounters (i.e., sex with a stranger) to achieve an optimal level of arousal and fulfill the expectancies they are seeking (Gaither & Sellbom, 2003; Zuckerman, 2009). The SSSS was derived from the Sensation Seeking Scale to focus on measuring sexual behavior (Gaither & Sellbom, 2003; Kalichman et al., 1994). It is an 11-item measure, where participants may respond ranging from 1 to 4 (with 1 being least like the participant and 4 being most like the participant; Gaither & Sellbom, 2003). In reliability and validity studies regarding the SSSS, the alpha coefficients were .81 for women and .83 for men (Gaither & Sellbom, 2003). In addition to these control populations, Kalichman and Rompa (1995) found alpha coefficients of .79 and .81 for gay men and inner-city residents, respectively. This shows strong correlations over time and across populations depending on levels of sexual promiscuity. In addition, strong correlations between high SSSS scores and general sensation-seeking behaviors in participants were found across studies, whether this reported behavior was sexual or otherwise (Kalichman & Rompa, 1995). Cronbach’s alpha for the current sample was .84 for the entire measure.

Results

To assess the role of response bias on self-report measures of sexual fantasies, we conducted several sets of analyses. First, we performed simple bivariate correlations comparing scores on the sexual fantasy measures with scores on the Marlowe–Crowne. Pearson r correlations revealed a significant association between the WSFQ (r = −.412, p = .001) and the SFQ (r = −.316, p = .001). Second, we conducted additional correlations comparing the WSFQ and SFQ with other measures of sexual deviance, the SSSS and the SCS. Results suggest significant associations between the WSFQ and the SSSS (r = .682, p = .001) and the SCS (r = .613, p = .001). The SFQ demonstrates similar findings with the SSSS (r = .531, p = .001) and SCS (r = .586, p = .001). These analyses were followed by partial correlations with the Marlowe–Crowne as the covariate and still identified significant associations between the WSFQ and the SSSS (r = .635, p = .001) and the SCS (r = .518, p = .001) along with the SFQ and the SSSS (r = .460, p = .001) and the SCS (r = .529, p = .001; see Tables 1 and 2 for all correlations).

Table 1.

Correlations Between Measures of Sexual Fantasy and Measures of Sexual Deviance.

Measures	SFQ	SSSS	SCS
WSFQ	.729	.682	.613
SFQ		.531	.586
SSSS			.631
SCS

Note. WSFQ = Wilson Sex Fantasy Questionnaire; SFQ = Sexual Fantasy Questionnaire; SSSS = Sexual Sensation Seeking Scale; SCS = Sexual Compulsivity Scale.

Table 2.

Partial Correlations Between Measures With the Marlowe–Crowne as a Covariate.

Measures	SFQ	SSSS	SCS
WSFQ	.691	.635	.518
SFQ		.460	.529
SSSS			.487
SCS

Note. WSFQ = Wilson Sex Fantasy Questionnaire; SFQ = Sexual Fantasy Questionnaire; SSSS = Sexual Sensation Seeking Scale; SCS = Sexual Compulsivity Scale.

Finally, we examined the association between scores on self-report measures of fantasies and behavioral indications of sexual deviance. We examined offenders on several variables: unconventional sexual practices (present or absent), sexually inappropriate in school (present or absent), sex offender (present or absent), statement of responsibility (present or absent), number of victims of sexual assault, age of victim (youngest entered), sexual violence witnessed as a juvenile (present or absent), number of charges for sexual offenses, number of convictions for sexual offenses, sex abuse (present or absent), and their Marlowe–Crowne responses. Two regressions were run to examine the possibility of specific deviant behaviors and the validity of the Marlowe–Crowne indicating sexual deviance. These specific variables were chosen because they had been collected as part of the demographic information and intuitively appeared to be potential behavioral indications of sexual deviance. The intent of these regressions was to see whether the previous findings extended to behavioral indications that might be more readily discernable in clinical practice compared with the administration of a self-report measure. In Table 3, the WSFQ is the dependent variable, F(11, 236) = 4.75, p = .01. In Table 4, the SFQ is the dependent variable, F(11, 160) = 2.37, p = .01. Interestingly, these results indicated that sexual fantasies were not related to self-report measures of sexual deviance. Only the Marlowe–Crowne was a significant contributor in either regression. It should also be noted that there were no differences across mental health diagnoses.

Table 3.

Linear Regression WSFQ.

Variable	Beta	Significance
Engaged in Unconventional Sexual Practices	.073	.242
Sexually Inappropriate in School	.019	.755
Sex Offender	−.038	.528
Statement of Responsibility	.051	.404
Number of Victims of Sexual Assault	−.027	.660
Age of the victim (Youngest Entered)	.027	.656
Sexual Violence Witness as a Juvenile	−.017	.773
Number of Charges for Sexual Offenses	−.062	.506
Number of Convictions for Sexual Offenses	.013	.889
Sex Abuse	.055	.366
Marlowe–Crowne Total	−.374	.000

Note. WSFQ = Wilson Sex Fantasy Questionnaire.

Table 4.

Linear Regression SFQ.

Variable	Beta	Significance
Engaged in Unconventional Sexual Practices	.102	.196
Sexually Inappropriate in School	.126	.101
Sex Offender	.058	.449
Statement of Responsibility	.058	.451
Number of Victims of Sexual Assault	−.019	.810
Age of the Victim (Youngest Entered)	.035	.649
Sexual Violence Witness as a Juvenile	.086	.254
Number of Charges for Sexual Offenses	−.100	.306
Number of Convictions for Sexual Offenses	.081	.397
Sex Abuse	.088	.249
Marlowe–Crowne Total	−.219	.007

Note. SFQ = Sexual Fantasy Questionnaire.

Discussion

The results of the present study offer several important insights. We performed two sets of bivariate correlations initially between the sexual fantasy measures and the measure of response bias and the additional measures of sexual deviant responding. Results examining the association between the Marlowe–Crowne and the WSFQ and the SFQ indicate that both measures are influenced by social desirable responding. The significant, negative correlations suggest that higher scores on the Marlowe–Crowne are associated with lower scores on the sexual fantasy and thereby indicate that participants did produce lower scores on the sexual fantasy measures when they produced higher scores on the measure of social desirability. Therefore participants were trying to look less deviant in their self-reports of sexual fantasies, as expected.

The second set of bivariate correlations between the sexual fantasy measures and the other measures of sexual deviance clearly demonstrated a significant degree of concurrent validity for these self-report measures of sexual fantasies. All correlations were statistically significant at levels equal to or better than p < .001. These correlations were then duplicated using Marlowe–Crowne scores as a covariate in a series of partial correlations using measures of sexual fantasy and the other measures of sexual deviance. Results of the partial correlations were similar to the previous bivariate correlations in magnitude and the overall associated significance level (all p < .001). Overall, these results suggest that although responses on the sexual fantasy measures are associated attempts at social desirable responding, the level of social desirability does not suppress the robustness of the significant associations between sexual fantasies and self-report measures of sexual deviance.

Finally, the two linear regressions using the measures of sexual fantasy as separate dependent variable and multiple behavioral indications of sexual deviance, along with scores on the Marlowe–Crowne, as predictor variables, were calculated. Although both of these regressions were significant, the only statistically significant predictors were scores on the Marlowe–Crowne. None of the behavioral predictors of sexual deviance were significant predictors of scores on the self-report measures of sexual fantasy. Together, these results suggest significant influence of response bias in assessing sexual fantasies on self-report measures—without making self-report measures of this construct meaningless—but do not suggest similar findings for sexually deviant behaviors.

These results may have important implications. Although it appears that responses on these measures include some significant response bias, that bias is not sufficient enough to eliminate the association between sexual fantasies and related constructs such as sexual compulsivity and sexual sensation seeking. As a result, it appears that despite their transparent nature, self-report measures of sexual fantasies can be relied on to produce indications of sexual deviance. This finding is true for false negatives as well as true positives, given the correlational nature of the results simply tests for associations. There has been concern that although self-report measures of sexual fantasy may be more efficient and potentially more objective at assessing the level sexual deviance with regard to sexual fantasizing (Baumgartner et al., 2002), responses are more easily distorted by offenders, and thereby results are meaningless. However, the results of the present study suggest that this is not the case, and although response patterns are associated with social desirable responding, as one would expect given the potential perceived consequences by offenders and the sensitive nature of the topic, the level of social desirable responding does not invalidate responses on measures of sexual fantasy. Therefore, clinicians and researcher can feel more confident in the use of self-report measures for assessing sexual fantasies.

An unintended finding of the present study was the degree to which self-report measures of sexual fantasies relate to actual sexually deviant behaviors. The current results did not find any such relationship. Unfortunately, the literature is largely silent on this very question, so the degree to which these results are in keeping with other efforts is unclear. Although there was no relationship between selected deviant behaviors and scores on either self-report measure of sexual fantasies, this may be an issue of common method variance (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). In other words, it is common in psychology to see significant associations in variables that have similar methods of measurement such as two variables being measured using self-report measures. The lack of significant findings could be because of the manner in which the variables of interest were measured or that self-report measures are not associated with actual behavior. This is an especially important question because a connection has been established between deviant sexual predilections and future sexual recidivism (Hanson & Morton-Bourgon, 2005). Ultimately, professionals assess sexual deviance to identify the potential for particular sexual deviant behaviors. It may also be that the lack of temporal proximity between the measurements of the variables of interest also accounts for the lack of significant findings. Participants responded to the self-report measures on admission. The sexual deviant behaviors used to predict these scores occurred years or even decades prior. Sexual deviant behaviors occurring while institutionalized or even future sexual deviant behaviors were not available to compare with scores on the sexual fantasy measures.

Another finding of the current research suggests the two measures of sexual fantasies appear statistically similar in evaluation purposes. Representatively, the WSFQ and the SFQ are both valid and reliable assessment tools for sexual fantasies. Both measures showed significant relationships to a measure of response bias with the WSFQ indicating a numerically greater correlation between the two measures. Furthermore, both measures exhibited significant association with other measures of sexual deviance, although the WSFQ appeared consistently higher than the SFQ, even after controlling for response bias. Finally, both measures were not associated with actual deviant behaviors. In total, these results do not support the use of one measure over the other but do present some potentially interesting finings. Skovran et al. (2010) produced similar findings in a similar sample using both of these self-report measures of sexual fantasies. Furthermore, they suggested that convergent administration of both measures of sexual fantasy may be the most accurate manner to assess sexual fantasies (Skovran et al., 2010).

Finally, certain limitations need to be considered in this research. The use of institutional records could be one limitation regarding the behavioral data. Coders were unable to directly ask the offenders specific questions regarding their past deviant behaviors. Instead, the information was collected from the institutional files. Therefore, it is possible that certain information was never reported in the file although the offender exhibited or experienced these behaviors at one point prior to reporting. Another limitation could be using the Marlowe–Crowne itself. Although it has long been recognized as the gold standard in the measurement of social desirability, some have argued that the Balanced Inventory of Desirable Responding (Paulhus, 1984) has advanced the field by improving on a simple unitary conceptualization of social desirability and breaking it out into impression management and self-deception. Future research could be performed calculating the same statistical analyses, however, substituting the Marlowe–Crowne with the Balanced Inventory of Desirable Responding. A further criticism could be directed that the SSSS and the SCS need further validation as measures. Only Williams et al. (1998) have used these measures within forensic population, although they did indicate they exhibited appropriate psychometric properties in their examination.

Overall, this study intended to assess the degree to which self-report measures of sexual fantasies were prone to response bias. Although there was a clear association between a measure of response bias and scores from self-report measures of sexual fantasies, the measures still indicated utility by continuing to exhibit significant associations with other measures of sexual deviance. Sexual fantasies are frequently used to assess risk and treatment progress (Adherhold et al., 2011). Therefore, these results should be helpful in decreasing concern about their inappropriate use in clinical situations or even research settings. However, these results indicate that self-report measures of sexual fantasies do not relate to specific deviant behaviors. These results do indicate that both the SFQ and WSFQ are appropriate to use in assessing risk and treatment progress as they are only valid measurements of sexual fantasies and indications of sexual deviance in an offender. As deviant behavior is a strong predictor of recidivism, a sexual fantasy questionnaire should still be administered to offenders to seek whether any indications of sexual deviance exist to assist an offender in finding ways to decrease the risk of sexual recidivism. However, future studies need to address this use to substantiate a direct link between recidivism and sexual fantasizing.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Adherhold

Huss

M. T.

Palarea

R. P.

(2011). Sexuelle Phantasien Sexualstraftaten (The role of sexual fantasies in sexual offending). In Roberts

(Ed.), Gewaltphantasien (Violent fantasies: Between Welsten and realities): Zwischen Welsten und Wirklichkeiten (pp. 155-174). Frankfurt, Germany: Verlag für Polizeiwissenschaft.

Bartels

R. M.

Gannon

T. A.

(2011). Understanding the sexual fantasies of sex offenders and their correlates. Aggression and Violent Behavior, 16, 551-561.

Baumgartner

J. V.

Scalora

M. J.

Huss

M. T.

(2002). Assessment of the Wilson Sex Fantasy Questionnaire among child molesters and nonsexual forensic offenders. Sexual Abuse: Journal of Research and Treatment, 14, 19-30.

Beech

A. R.

Craig

L. A.

Browne

K. D.

(2009). Overview. In Beech

A. R.

Craig

L. A.

Browne

K. D.

(Eds.), Assessment and treatment of sex offenders: A handbook (pp. 1-12). West Sussex, UK: John Wiley.

Campbell

J. C.

(2007). Preface. In Campbell

J. C.

(Ed.), Assessing dangerousness: Violence by batterers and child abusers (2nd ed., pp. xi-xii). New York, NY: Springer.

Carnes

P. J.

(2001). Out of the shadows: Understanding sexual addiction (3rd ed.). Center City, MN: Hazelden Publishing.

Cortoni

(2009). Factors associated with sexual recidivism. In Beech

A. R.

Craig

L. A.

Browne

K. D.

(Eds.), Assessment and treatment of sex offenders: A handbook (pp. 39-52). West Sussex, UK: John Wiley.

Crowne

D. P.

Marlowe

(1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.

Delmonico

D. L.

Griffin

E. J.

(2002). The challenge of treating compulsive sex. Counselor: The Magazine for Addiction Professionals, 3, 14-20.

10.

Gaither

G. A.

Sellbom

(2003). The sexual seeking scale: Reliability and validity within a heterosexual college student sample. Journal of Personality Assessment, 81, 157-167.

11.

Hanson

R. K.

Morton-Bourgon

K. E.

(2005). The characteristics of persistent sexual offenders: A meta-analysis of recidivism studies. Journal of Consulting and Clinical Psychology, 73, 1154-1163.

12.

Hart

S. D.

Michie

Cooke

D. J.

(2007). Precision of actuarial risk assessment instruments: Evaluating the “margins of error” of group v. individual predictions of violence. British Journal of Psychiatry, 190, 60-65.

13.

Howitt

(2004). What is the role of fantasy in sex offending? Criminal Behaviour and Mental Health, 14, 182-188.

14.

Huss

M. T.

(2014). Forensic psychology: Research, clinical practice, and applications (2nd ed.). New York, NY: John Wiley.

15.

Kalichman

S. C.

Johnson

J. R.

Adair

Rompa

Multhauf

Kelly

J. A.

(1994). Sexual sensation seeking: Scale development and predicting AIDS-risk behavior among homosexually active men. Journal of Personality Assessment, 62, 385-397.

16.

Kalichman

S. C.

Rompa

(1995). Sexual sensation seeking and Sexual Compulsivity Scales: Reliability, validity, and predicting HIV risk behavior. Journal of Personality Assessment, 65, 586-601.

17.

Leedes

(1999). Theory and praxis: A heuristic for describing, evaluating, and intervening on sexual desire disorders when sexual expression interferes with humanistic expression. Sexual Addiction & Compulsivity, 6, 289-310.

18.

Leitenberg

Henning

(1995). Sexual fantasy. Psychological Bulletin, 117, 469-496.

19.

McBride

K. R.

Reece

Sanders

S. A.

(2008). Using the Sexual Compulsivity Scale to predict outcomes of sexual behavior in young adults. Sexual Addiction & Compulsivity, 15, 97-115.

20.

Mills

J. F.

Kroner

D. G.

(2006). Impression management and self-report among violent offenders. Journal of Interpersonal Violence, 21, 178-192.

21.

O’Donohue

Letourneau

Dowling

(1997). Development and preliminary validation of Paraphilic Sexual Fantasy Questionnaire. Sexual Abuse: Journal of Research and Treatment, 9, 167-178.

22.

Paulhus

D. L.

(1984). Two-component models of socially desirable responding. Journal of Personality and Social Psychology, 46, 598-609.

23.

Paulhus

D. L.

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

J. P.

Shaver

P. R.

Wrightsman

L. S.

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

24.

Podsakoff

P. M.

MacKenzie

S. B.

Lee

J. Y.

Podsakoff

N. P.

(2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879-903.

25.

Skovran

L. C.

Huss

M. T.

Scalora

M. J.

(2010). Sexual fantasies and sensation seeking among psychopathic sexual offenders. Psychology, Crime & Law, 16, 617-629.

26.

Williams

Scalora

Huss

M. T.

(1998, November). Sensation seeking and sexual compulsivity in sex offenders. Poster presented at the Joint Conference of the Nebraska Psychological Society and the Association for Psychological and Educational Research in Kansas, Emporia, KS.

27.

Wilson

G. D.

(1988). Measurement of sex fantasy. Sexual and Marital Therapy, 3, 45-55.

28.

Wilson

G. D.

(2010). The Sex Fantasy Questionnaire: An update. Sexual and Relationship Therapy, 25, 68-72.

29.

Woodworth

Freimuth

Hutton

E. L.

Carpenter

Agar

A. D.

Logan

(2013). High-risk sexual offenders: An examination of sexual fantasy, sexual paraphilia, psychopathy, and offence characteristics. International Journal of Law and Psychiatry, 36, 144-156.

30.

Zuckerman

(2009). Sensation seeking. In Leary

M. R.

Hoyle

R. H.

(Eds.), Handbook of individual differences in social behavior (pp. 455-465). New York, NY: Guilford Press.