How Long Men Look at Images May Reveal Hidden Patterns of Sexual Arousal

A new study of civilly committed men in a secure treatment program has found that a simple, noninvasive computer task—measuring how long someone looks at an image—produces patterns that closely mirror the results of penile plethysmography, the most widely used physiological test of male sexual arousal. The research, published in Archives of Sexual Behavior, offers some of the strongest evidence yet that viewing time measures can serve as meaningful complements to a procedure long criticized as invasive, uncomfortable, and difficult to standardize.

Penile plethysmography, or PPG, records changes in penile circumference while a participant views or listens to sexual stimuli, providing a direct physiological index of arousal. It is regarded by many researchers as the most specific measure of sexual response in men and is used internationally for diagnosis, treatment planning, and risk management in forensic settings. Meta-analyses have underscored its predictive power: arousal to child stimuli has been identified as one of the strongest single predictors of sexual recidivism, and phallometric assessments reliably distinguish individuals who have committed sexual offenses from control groups. Yet the method has persistent drawbacks, including its invasiveness, the discomfort it can provoke, and a striking lack of standardization in stimuli, equipment, and interpretation across laboratories and programs.

Viewing time paradigms offer an alternative route to the same clinical question. In these tasks, participants view and rate images while software unobtrusively records how long they spend on each one. Longer viewing times are inferred to reflect greater sexual interest, on the theory that stimuli of higher sexual relevance demand more extensive cognitive processing. Although viewing time was originally developed to study sexual orientation, it has become a common forensic tool for assessing sexual interest in children, and prior research shows it can reliably distinguish men who have committed sexual offenses against children from those who have not. What remained unclear—until now—was how well viewing time results align with the physiological classifications produced by PPG, particularly across clinically meaningful subgroups.

The new study, led by Melinda R. Ward and Timothy G. Thornton of the Minnesota Sex Offender Program and Ashley E. Thompson of the University of Minnesota Duluth, examined 146 adult males who had been civilly committed as sexually dangerous persons or as having sexual psychopathic personality following incarceration. All participants had completed both a PPG assessment and a computerized viewing time assessment, the Affinity 2.5, within 30 days of one another. The researchers drew on archival data collected between August 2017 and August 2023 as part of routine treatment programming, applying strict quality controls, including an embedded attention task during PPG testing and the exclusion of assessments with poor engagement or technical problems.

The PPG data were used to compute a deviance differential, a score reflecting the difference between a person’s standardized arousal to adult stimuli and to child stimuli. Participants were then classified into three groups: child arousal, those with greater physiological arousal to prepubescent or pubescent stimuli; adult arousal, those with greater arousal to adult stimuli; and undifferentiated arousal, those showing relatively similar responses to both. The undifferentiated group, often overlooked in prior research, was deliberately included to provide preliminary insight into its characteristics. The classification thresholds used, at plus or minus 0.30, match those routinely applied in civil commitment programs across the United States, lending the findings direct ecological validity.

The viewing time task presented 80 standardized images of clothed individuals spanning four developmental stages and both sexes, with participants rating how sexually attractive or unattractive they found each image. Rather than analyzing raw viewing times, the researchers used a mean-rank procedure that captures each person’s relative processing latency across stimulus categories, reducing the influence of individual differences in overall response speed. Guided by incentive motivation theory, which holds that sexual stimuli acquire salience through conditioning and form motivational hierarchies operating at both automatic and controlled levels, the team predicted that viewing time patterns would diverge across the three arousal groups in ways that mirrored their physiological profiles.

That is largely what the data showed. After applying a Bonferroni correction across six statistical tests, significant group differences emerged for female stimuli only. Men in the child arousal group spent relatively longer viewing images of small child and pre-juvenile females, while men in the adult arousal group spent relatively longer viewing images of adult women. The effect sizes were substantial, with planned contrasts between the child and adult arousal groups yielding standardized differences of 0.82 for adult female stimuli and 0.93 for small child female stimuli. A planned contrast for pre-juvenile female stimuli, though its omnibus test was not significant, also reached significance with a moderate effect size. The undifferentiated group consistently fell between the two extremes, and its viewing times for female stimuli exceeded those for male stimuli, suggesting that sexual orientation toward women, rather than age preference, was the dominant organizing feature of their interest patterns.

The restriction of significant findings to female stimuli is itself informative. The authors point to the well-documented phenomenon of sex-specificity in male sexual interest, in which attraction to a given sex remains consistent across levels of sexual maturity. Recent multidimensional scaling work supports a model in which men’s sexual interest is organized along two intersecting dimensions—the sex of the target and sexual maturity—with maturity-based preferences layered onto an underlying sex-based orientation. Because nearly 70 percent of the sample identified as heterosexual and most had committed offenses primarily against females, the absence of group differences for male stimuli is consistent with this two-dimensional structure of male sexual interest.

The convergence between physiological arousal and cognitive viewing time carries practical weight for forensic assessment. Professional guidelines, including those of the Association for the Treatment and Prevention of Sexual Abuse, caution that clinical decisions should never rest on PPG results alone and recommend integrating multiple sources of information. The new findings suggest that viewing time measures can corroborate PPG-derived classifications, offering a less intrusive option when physiological testing is unavailable, inappropriate, or declined by a client. Prior work has also hinted at predictive value: one study found that men with higher visual reaction time scores to child stimuli were significantly more likely to be arrested and charged for new sexual offenses, with reoffense concentrated in the first six months after release.

The authors are careful to note the limits of both methods. Viewing time tasks can potentially be manipulated, particularly by clients in adversarial settings who have access to detailed task descriptions online, and participants may attempt to suppress arousal or accelerate their ratings for strategic reasons—a concern heightened in a civilly committed population seeking to demonstrate reduced risk. Stimulus characteristics also differ between the two assessments, notably the absence of coercive narrative content in the viewing time task, and neither measure should be interpreted as a direct indicator of offending behavior. Generalizability is further constrained by the sample, which consisted entirely of adult males in a single secure program. The authors call for replication across lower-risk offender groups and community samples, and for multi-method viewing time approaches that incorporate self-report ratings and additional latency indices. Even so, the study concludes that the three-category deviance differential classification captures psychologically meaningful profiles of sexual interest, strengthening the case for multimodal assessment in forensic and clinical practice.

Subject of Research: Concordance between penile plethysmography-derived arousal classifications and viewing time measures of sexual interest in civilly committed adult males

Article Title: Sexual Interest and Arousal in Males Who Are Civilly Committed: Exploring Overlap and Distinctions

Article References: Ward, M. R., Thompson, A. E., & Thornton, T. G. (2026). Sexual Interest and Arousal in Males Who Are Civilly Committed: Exploring Overlap and Distinctions. Archives of Sexual Behavior. https://doi.org/10.1007/s10508-026-03547-6

Image Credits: AI Generated

DOI: 10.1007/s10508-026-03547-6

Keywords: penile plethysmography, viewing time, sexual interest, sexual arousal, deviance differential, forensic assessment, civil commitment, sexual offending, Affinity assessment, sexual deviance, risk assessment, Archives of Sexual Behavior