Review of Polygraph Accuracy Research

Review of polygraph accuracy research reporting accuracy rates exceeding .90 using the Comparison Question Test format.

Systematic review of published polygraph accuracy studies using CQT methodology.

Honts and Peterson's review of polygraph accuracy research reported that CQT accuracy exceeded .90, consistent with findings by Raskin (2002) and the earlier Raskin and Podlesny (1979) estimates. This convergence across independent reviewers spanning two decades helped establish a scientific consensus that properly conducted CQT examinations reliably achieve high discrimination rates between truthful and deceptive examinees.

The convergence of accuracy estimates above .90 from multiple independent reviewers strengthens the scientific case for CQT validity. These consistent findings across different research groups and time periods provide robust evidence for the technique's reliability.

Background & Context

The 1997 work by Honts and Peterson represents a pivotal moment in the legal and scientific discourse surrounding polygraph validity. This document was filed as the "Brief of the Committee of Concerned Social Scientists as Amicus Curiae" in United States v. Scheffer, a landmark Supreme Court case addressing the admissibility of polygraph evidence in military courts-martial.

This brief emerged during a critical period when polygraph testing faced intense scrutiny in both scientific and legal communities. Between 1981 and 1997, the Department of Defense conducted over 400,000 polygraph examinations for counterintelligence, security, and criminal investigations, yet the evidentiary status of polygraph results remained deeply contested. The Committee of Concerned Social Scientists sought to inform the Supreme Court about the state of polygraph science by synthesizing existing accuracy research.

Rather than presenting new empirical findings, this amicus brief served as an authoritative review synthesizing decades of Comparison Question Test (CQT) research. Its timing and venue gave it significant influence on how courts and policymakers understood polygraph reliability throughout the late 1990s and early 2000s.

Research Design & Methodology

This work was not an original empirical research study but rather a systematic review and synthesis of existing polygraph accuracy literature prepared for legal proceedings. Honts and Peterson summarized the results of high quality polygraph studies in an amicus brief filed in U.S. v. Scheffer, and reported the accuracy rate as exceeding .900.

The methodology consisted of reviewing and aggregating findings from published CQT accuracy studies available in the scientific literature through 1997. The authors drew upon multiple independent sources to establish convergent validity for their accuracy estimates. Key criteria for inclusion likely emphasized studies employing proper CQT protocols with verified ground truth criteria, though the specific selection and weighting procedures were not detailed in subsequent citations to this work.

As a legal brief rather than a peer-reviewed article, this document lacked the methodological transparency typically required for academic publications—including detailed inclusion/exclusion criteria, quality assessment procedures, and formal meta-analytic statistics that would later become standard in polygraph reviews.

Results & Key Findings

Honts and Peterson (1997) reported the accuracy of polygraph studies as exceeding .90, establishing accuracy rates above 90% for properly conducted CQT examinations. This figure represented an aggregation across multiple high-quality studies examining the CQT's ability to discriminate truthful from deceptive examinees.

The brief's conclusions aligned remarkably well with convergent evidence from independent reviewers across different time periods:

• The findings were consistent with the .900 accuracy estimation of Raskin and Podlesny (1979), demonstrating stability over nearly two decades
• The accuracy estimation was later confirmed by Raskin and Honts (2002), who also reported polygraph accuracy exceeding .900
• This convergence across reviewers spanning 1979-2002 suggested robust, replicable findings rather than isolated claims

While three reviews reported results in the high .90s, the majority of systematic studies from this era (including OTA 1983, and later NRC 2002) produced accuracy estimates in the .80s to low .90s range, positioning Honts and Peterson's estimates at the higher end of the contemporary scientific consensus.

Discussion & Significance

The Honts and Peterson brief played a crucial role in establishing what the metadata describes as "scientific consensus" regarding CQT validity, though this characterization requires important qualification. The convergence of accuracy estimates above .90 from multiple independent reviewers (Raskin & Podlesny 1979, Honts & Peterson 1997, Raskin 2002) provided ammunition for polygraph proponents arguing for the technique's forensic utility and legal admissibility.

However, reviews of CQT accuracy have varied considerably, though they generally produced overall accuracy estimates of over 85%. The 2003 National Academy of Sciences review concluded there was "little basis for the expectation that a polygraph test could have extremely high accuracy", and found that 57 of approximately 80 research studies relied upon by polygraph proponents were significantly flawed. This suggests the "consensus" was far more contested than the brief's conclusions might imply.

The brief's influence extended well beyond the Scheffer case itself. For nearly a decade, it served as a frequently cited reference point in legal proceedings, policy debates, and academic discussions about polygraph validity. Starting in 1997, arguments detailing the case for and against legal admissibility of polygraph tests were published as exchanges in Modern Scientific Evidence, a legal volume series for judges and lawyers, with this brief providing foundational support for the pro-admissibility position.

Limitations & Considerations

As a legal brief rather than a peer-reviewed research article, this work lacks the methodological transparency and reproducibility standards expected of formal scientific publications. The specific studies included, quality assessment criteria, weighting procedures, and handling of conflicting evidence remain unspecified in available citations. This makes independent verification and critical evaluation difficult.

The accuracy claims face important contextual limitations. The Office of Technology Assessment (1983) suggested laboratory studies had average unweighted accuracy of .832 with inconclusive rate of .269, while field studies had accuracy of .847 with inconclusive rate of .042. The >.90 figure likely reflects selective aggregation of higher-quality or higher-performing studies rather than comprehensive averaging across all available evidence. Additionally, specific-incident polygraph testing could discern truth at "a level greater than chance, yet short of perfection," though actual field accuracy is almost certainly lower than laboratory studies suggest.

The broader scientific community remained—and remains—deeply divided about CQT validity, with critics arguing the technique rests on implausible theoretical assumptions and produces unacceptably high false positive rates, particularly in screening contexts where base rates of deception are low.

Practical Applications

For polygraph examiners and agencies relying on CQT testing in 1997, this brief provided authoritative support for their practices during a period of intense legal and scientific challenge. The >.90 accuracy claim offered a benchmark for quality control and examiner training, though practitioners should recognize this represents optimistic estimates under ideal conditions rather than guaranteed field performance.

For consumers and legal professionals, this work underscores the importance of understanding that accuracy estimates vary substantially based on methodological choices, study quality, and testing context. While convergent reviews suggesting accuracy in the 85-90% range provide some support for CQT utility in criminal investigations, these figures should not be interpreted as guaranteeing highly accurate results in individual cases—particularly given concerns about examiner variability, countermeasures, and the fundamental theoretical assumptions underlying comparison question methodology.