Home › Polygraph Research › A prospective study of the impact of polygraphy…
Catalogue entry · Sex Offender Management (PCSOT)
A prospective study of the impact of polygraphy on high risk behaviours in adult sex offenders
Daniel E. Sosnowski — Sexual Abuse: A Journal of Research and Treatment,
At Time 1, 97% of polygraphed sex offenders disclosed an average of 2.45 high-risk behaviors unknown to supervising probation officers (76 total); at Time 2, disclosures dropped significantly (t = 2.55, p = .019) and the polygraph pass rate rose from 29% to 71% (chi square = 12.82, p < .001), indicating both a disclosure and a behavioral deterrence effect following direct polygraph experience.
Abstract
This prospective study by Grubin, Madsen, Parsons, Sosnowski, and Warberg (2004) examined whether polygraph testing would result in adult sex offenders on community supervision programs engaging in fewer high-risk behaviors, using a two-group experimental design with examinations at three and six months. The study found that 97% of offenders who attended their first polygraph disclosed an average of 2.45 high-risk behaviors previously unknown to their probation officers, and that repeated testing was associated with a statistically significant reduction in such behaviors and a marked increase in proactive disclosure to supervisors.
Methodology
Prospective controlled field study with 50 adult male sex offenders in community treatment across three English probation areas, randomly allocated to Polygraph Aware and Polygraph Unaware groups, with Comparison Question Technique (CQT) polygraph examinations at 3 months and 6 months, using Static-99 for risk stratification and paired t-tests and chi-square analyses for outcome measurement.
Detailed summary
This landmark UK Home Office-funded study was the first to prospectively examine whether polygraph testing reduces high-risk behaviors in sex offenders using a controlled design. Fifty adult males in English community treatment programs were allocated to polygraph-aware and polygraph-unaware conditions, with 3-4 individualized risk behaviors identified per participant. Using the CQT administered by professional examiners, the study tracked disclosures and behaviors at baseline, 3 months, and 6 months. This research was particularly significant as it was conducted in England where polygraphy was not yet used in criminal justice settings, and it informed what would later become national policy.
Implications for polygraph practice
This study provided crucial evidence for policy development regarding polygraph use in sex offender management and represents a significant methodological advance in PCSOT research through its prospective design with comparison groups.
Comprehensive study analysis
An in-depth, original analysis of this research study's methodology, findings, and significance for the polygraph profession.
Background & Context
By the early 2000s, post-conviction sex offender polygraph testing had become an established — if contested — feature of community supervision programs in the United States. Critics attacked not only the evidence base for polygraph use, but also the scientific standing of polygraphy more generally, with much of the argument focused on investigative settings such as crime detection and pre-employment screening, which are only partly relevant to post-conviction applications. In the post-conviction context, the emphasis shifts: accuracy of detection becomes secondary to the therapeutic and supervisory utility of disclosures made during the examination process.
Advocates of post-conviction polygraph examinations claim that the tool enables clinicians to obtain more reliable sexual histories and more accurate offence behavior descriptions, both of which assist in overcoming denial and improving assessment of treatment need and risk of reoffending. However, published research data supporting these claims was limited at the time. Prior studies — while promising — tended to be retrospective, non-peer-reviewed, or lacked the control group designs necessary to draw firm causal conclusions about polygraph's behavioral impact on offenders.
This study examined whether polygraph testing would result in sex offenders engaging in fewer high-risk behaviors — a question that had never before been addressed using a prospective design with comparison groups. The research was conducted in England, where at the time polygraphy was not used at all by police, probation, or prison services, making this work especially significant in informing what would later become national policy. The study was funded by a UK Home Office Innovation Grant and represents a landmark collaboration between British and American polygraph researchers.
Research Design & Methodology
Fifty adult male sex offenders taking part in community treatment programs were allocated into two groups: "Polygraph Aware" subjects were told they would receive a polygraph examination in three months regarding their high-risk behaviors, while "Polygraph Unaware" subjects were told only that their behavior would be reviewed in three months. Relevant behaviors for each subject were established at baseline interviews, following which both groups were polygraphed at three months, and all subjects were polygraphed again at six months.
The sample included participants drawn from three distinct English probation service areas. The 50 participants were all adult males enrolled in community-based sex offender treatment programs. The sample was composed of 45 men who had offended against children and 5 whose victims were adult females, with ages ranging from 22 to 67 (mean age 41). Risk stratification was conducted using the well-validated Static-99 actuarial instrument as well as the dynamic risk factor scale developed by Hanson and Harris, providing an objective baseline for comparing groups.
Key elements of the research design included:
- Instrument: The Comparison Question Technique (CQT), administered by two American Polygraph Association–qualified examiners with specialist expertise in sex offender testing
- Individualized risk behaviors: 3–4 specific high-risk behaviors were identified per participant by their treatment providers (e.g., unsupervised contact with children, use of pornography)
- Multi-stage disclosure tracking: Disclosures were recorded at four distinct points — the pre-examination research interview, the polygraph pre-test interview, the polygraph examination itself, and the post-test interview
- Prospective follow-up: Measurements taken at baseline, 3 months (Time 1), and 6 months (Time 2)
- Statistical methods: Paired-sample t-tests, chi-square analyses, and ANOVA (F-tests) to compare group differences and change over time
Participants were fully informed that information obtained during the polygraph examinations would be shared with their treatment providers and probation officers, an ethically important transparency condition that also strengthened the ecological validity of the study's findings.
Results & Key Findings
The headline result: an almost universal concealment of high-risk behaviors from supervisors. Thirty-two subjects (64%) attended the first polygraph examination, with 31 (97%) disclosing an average of 2.45 high-risk behaviors each previously unknown to supervising probation officers. This finding alone — that probation staff were unaware of the vast majority of ongoing risk behaviors — powerfully demonstrated the intelligence gap that polygraph testing could address.
Key statistical findings across both time points:
- Time 1 total disclosures: 76 high-risk behaviors disclosed by 31 of 32 participants; mean of 2.45 behaviors per offender (SD = 1.7, range 0–7)
- Polygraph failure rate at Time 1: 78% of subjects (25 of 32) failed the polygraph; of those, 80% (n=20) made further disclosures in the post-test interview
- No group effect: There was no statistically significant difference in disclosed behaviors between the Polygraph Aware and Polygraph Unaware groups at Time 1 (F = .613, p = .44), indicating that advance knowledge of the test did not reduce risky behavior before the first examination
- Time 2 improvement: Twenty-one subjects (42%) completed the second polygraph test, with 71% disclosing an average of 1.57 behaviors, a significant decrease compared with the first test.
- Pass rate improvement: 71% of subjects passed the polygraph at Time 2, compared with just 29% at Time 1 — a highly significant improvement (chi square = 12.82, p < .001)
- Proactive supervisor disclosure: Of the 15 subjects who reported high-risk behaviors at Time 2, 9 (60%) had already disclosed these to their supervisors, compared with just 1 who had done so at Time 1.
The improvement between Time 1 and Time 2 was statistically robust, with the paired t-test yielding t = 2.55, p = .019. Offender self-report feedback reinforced these behavioral findings: 20 of 21 completers reported that polygraph examination helped them avoid reoffending, 57% said that knowledge of the impending test caused them to inhibit their behavior, and 52% said it resulted in them reporting more information to their supervising probation officers.
Discussion & Significance
Perhaps the most clinically significant dimension of this study is what the initial findings revealed about the invisible landscape of risk in supervised sex offender populations. The fact that nearly all participants were engaging in meaningful risk behaviors — ranging from masturbation to deviant fantasies (84%) to unsupervised contact with children (28%) — while these behaviors were almost entirely unknown to their probation officers, underscores a profound supervision gap. The study concluded that polygraph testing resulted in offenders engaging in less high-risk behavior, although the possibility that offenders fabricated reports of high-risk behaviours to satisfy examiners is also considered; similarly, offenders seemed to be more honest with their supervisors, but this only occurred after experience of the test itself.
The authors made a conceptually important contribution by reframing the role of polygraph in this context. Rather than functioning as a binary "lie detector," the authors argued that in post-conviction settings the tool is better understood as a "truth facilitator" — a structural mechanism that creates conditions under which offenders are more likely to disclose clinically relevant information to treatment providers and supervisors. This reframing has significant implications for how PCSOT programs are evaluated and justified, shifting the evidentiary standard from detection accuracy to therapeutic and supervisory utility.
The behavioral deterrence effect observed between Time 1 and Time 2 is especially noteworthy — but so is the timing. The deterrent impact of anticipated polygraph testing only emerged after offenders had directly experienced the examination process. This suggests that the threat of a future polygraph test alone is insufficient to change behavior; it is the lived experience of the test — its rigor, its physiological immediacy, and the interpersonal accountability it creates — that produces behavioral inhibition. This finding has direct implications for program design, suggesting that simply notifying offenders of upcoming tests may not be sufficient without ensuring they actually undergo them.
Limitations & Considerations
The authors were admirably candid about the study's limitations, and these deserve careful attention from any practitioner or researcher drawing on the findings:
- High attrition rate: Only 64% of enrolled participants attended the first polygraph, and only 42% completed both. The sample of completers may systematically over-represent motivated, treatment-engaged offenders.
- Self-report reliance: Almost all behavioral data depended on offender self-report, which could include fabricated disclosures ("defensive lying") to explain a failed examination or satisfy the examiner. The authors could verify accuracy only in exceptional cases.
- No ground truth for behavioral change: The apparent reduction in high-risk behaviors at Time 2 could reflect actual behavioral change, reduced disclosure motivation, or improved ability to beat the test — these possibilities cannot be distinguished within this design.
- No Polygraph Unaware group at Time 2: Because virtually all participants disclosed risk behaviors at Time 1, all were told to expect a second test, collapsing the experimental groups and eliminating the comparison condition.
- Volunteer bias and voluntary participation: The study was explicitly voluntary, which may have filtered out the most resistant or highest-risk participants from the outset — those who declined to participate initially represented 57% of those approached.
- Small final sample: With only 21 completers at Time 2, statistical power was limited and findings should be interpreted with caution regarding generalizability.
The small numbers, possible sample bias, and the self-report nature of the study mean that one must be cautious in interpreting these results. The authors themselves noted that the self-selecting group of completers were likely the most motivated subset, and that extending polygraph to a broader population of offenders would require some form of structured compulsion to achieve representative coverage.
Practical Applications
For practitioners working in community-based sex offender management, this study delivers several operationally important insights. First, it provides compelling empirical grounds for incorporating periodic polygraph examinations into supervision protocols: the sheer volume of undisclosed risk behaviors identified at Time 1 — behaviors entirely invisible to probation staff — illustrates that standard supervision alone is insufficient to capture ongoing risk. Second, the significant increase in proactive supervisor disclosure between Time 1 and Time 2 suggests that repeated polygraph testing may reshape the supervisory relationship itself, encouraging offenders to communicate concerns to their probation officers rather than concealing them until an examination forces disclosure.
The finding that behavioral change required direct polygraph experience — not just forewarning — has practical design implications for PCSOT programs. Offenders seemed to be more honest with their supervisors, but this only occurred after experience of the test itself. This points toward the importance of ensuring high-risk offenders actually undergo their first examination rather than merely being notified of upcoming tests. The study also supports the policy conclusion reached by its authors: that voluntary participation in polygraph programs will inevitably miss the highest-risk, least-motivated individuals, suggesting that mandated testing frameworks — such as those later implemented in England under the Offender Management Act 2007 — are necessary to achieve the population-level benefits this research suggests are achievable.
The analysis above is original editorial content based on our review of this research. For the complete study including full data, methodology details, and author discussion, access the original publication below.
Related research
Other studies in this category that may be of interest.
A Survey of the Views of Catholic Safeguarding Coordinators about the Inclusion…
[002] (2022)Exploring the implications of item-level data in sex offenders’ maintenance polygraph results
[003] (2021)Why Polygraph Testing Does Not Consistently Lead to Reduced Recidivism
[004] (2020 • UK POLICE EVALUATION)An Evaluation of Polygraph Testing by Police to Manage Individuals Convicted or…
[005] (2020)Improving the Quality of Sexual History Disclosure: Emphasis on a Polygraph Examination
[006] (2020 • KOREAN STUDY)Improving the Quality of Sexual History Disclosure on Sex Offenders: Emphasis on…
Join Our Examiner Network
APA-trained examiners using validated techniques can apply to join the LieDetectorTest.com network.
Keep reading the ledger.
Every peer-reviewed study on polygraph and deception detection we track — catalogued, searchable and citable.

