This study investigated the frequency of distortion that occurs when raw score patterns of Affinity 2.0, a viewing time measure designed to assess sexual interest, are converted to ipsative scores. Eighty-eight percent of a sample of ninety-nine non-pedophilic, exclusively heterosexual males' profiles sustained some degree of distortion. The study also applied Brown's (2005) approach to predicting distortion with this sample of males' responses. Brown's techniques were largely ineffective in predicting males' distortion scores.

Given that the Affinity 2.0 is a largely untested instrument that purports to measure sexual attraction, it was unknown whether or not the instrument is temporally stable and would provide a characteristic response curve for normal heterosexual male sexual interest. This study examined the Affinity 2.0, a newly standardized viewing time (VT) instrument that purports to measure sexual interest. More specifically, the Affinity 2.0 was examined in regard to its temporal stability and ability to generate a characteristic curve for non-pedophilic, exclusively heterosexual male interests. This examination was carried out by administering the Affinity 2.0 twice on a sample of 88 self-reportedly non-pedophilic, exclusively heterosexual men. The results of this study find the Affinity 2.0 to be mildly to moderately temporally stable and capable of rendering a characteristic sexual interest curve. Subjects created a typical profile that showed longest viewing time to images of adult and juvenile females with low viewing times to all other categories of attractors. Implications for norm-referenced decision making are discussed.

The purpose of the current study was to evaluate the temporal stability of a normal heterosexual female response to the Affinity 2.0, a newly standardized viewing time (VT) instrument that purports to measure sexual interest. Participants were 120 female undergraduate and graduate students from a private university (mean age = 21.67 years) who met inclusion criteria of non-pedophilic interest/history, identified as “Exclusively Heterosexual" on the Kinsey Scale, and who took the Affinity 2.0 both at test and at retest (approximately two weeks later). Participants also filled out a questionnaire following the retest composed of demographic questions, a shortened version of the Marlowe-Crowne Desirability Scale-10 [M-C 2(10)], and declared their sexual interest on the Kinsey Scale. Pearson correlation coefficients (PPMCC), Spearman's Rho correlation coefficients, and a Chi-Square Goodness of Fit Test were all utilized to assess the temporal stability of the sample's response to Affinity 2.0. All PPMCC and Spearman's Rho correlations for VT were statistically significant at the p < .01 level; while practical significance for PPMCC could only be considered moderate as best (range of r =.28 - .63), Spearman's Rho correlations (range of r =.41 - .65) were stronger and approached practical significance. The existence of an overall distinct and discernable VT response pattern was apparent as the sample demonstrated identical categorical VT preferences both test and retest. These results were consistent with results from previous VT studies (Quinsey, Rice, Grant & Reid, 1993; Wright & Adams, 1994; Quinsey, Ketsetzis, Earls, & Karamanoukian, 1996). In addition, this VT pattern was found to be stable from test to retest as assessed by a Chi-Square Goodness of Fit Test. An unexpected finding was the presence of elevated VT for adult and juvenile female images. Implications, limitations, and future studies are discussed.

This study investigated distortion that occurs when raw scores are converted to ipsative scores on Affinity 2.0, a relatively new instrument for assessing sexual interest. Using a sample of 146 non-offending, heterosexual females, this study examined the characteristics of distorted ipsative score profiles and attempted to develop an algorithm to identify such distortions. A method was developed for defining distortion objectively. Of the 146 profiles,125 were found to contain some degree of distortion. Several hypotheses were formulated as to variables that might be related to distortion. These relationships were examined using Pearson Product Moment Correlations. Several statistically significant, but weak, correlations were found. An interaction effect was calculated for four of these variables, and was found to have a moderately strong correlation with distortion (r = .530, p <.01). An algorithm for identifying distortion was developed using this interaction effect. Several cut-off scores were tested. The most effective cut-off only correctly identified 42.9% of the significantly distorted profiles. Implications and limitations of the results are discussed, and directions for future research are provided.