Possible Sources of the Difference in the Severity of the Prior Information Effect on Polygraph Scoring Found in Three Published Studies

This 2025 study by Ginton compares three published studies examining confirmation bias effects in polygraph scoring to understand why they found different magnitudes of prior information effects. The analysis concludes that practical impact of examiner bias may affect less than 5% of real-world polygraph examinations.

Comparative analysis of three confirmation bias studies (Elaad et al. 1994, Ginton 2019, Krapohl & Dutton 2018) examining how prior case information influences polygraph chart interpretation and scoring decisions.

Ginton's 2025 comparative analysis examines discrepancies between three studies investigating confirmation bias in polygraph scoring, where examiners' prior knowledge of case details influences chart interpretation. The comparison reveals that while all three studies found evidence of bias effects, the magnitude varied considerably depending on study methodology and examiner samples. The analysis suggests that prior information primarily affects scoring when physiological responses are ambiguous, but has minimal impact when charts show clear deception or truthfulness indicators. Ginton concludes that real-world adverse effects may be limited to less than 5% of examinations, and suggests that using conservative inconclusive scoring zones could further mitigate bias impact.

The findings suggest that confirmation bias in polygraph scoring, while present, may have more limited practical impact than some critics argue, supporting the continued use of blind scoring protocols and conservative inconclusive ranges in operational practice.

Background & Context

Confirmation bias in forensic examination has emerged as a critical concern across multiple disciplines, from fingerprint analysis to polygraph testing. This 2025 study by Avital Ginton compares three Comparison Question Test (CQT) polygraph studies—Elaad et al. (1994), Ginton (2019), and Krapohl & Dutton (2018)—each indicating a different degree of the prior information effect in triggering confirmation bias while scoring polygraph examinations. The variation in findings across these three studies has created confusion about the practical severity of bias effects in real-world polygraph practice.

The "prior information effect" refers to how knowledge of case details, suspect background, or investigative findings can unconsciously influence an examiner's interpretation of polygraph charts. Critics of the CQT have long argued that this contamination represents a fundamental flaw making polygraph results unreliable. However, the magnitude of this effect in actual practice has been disputed, with some studies showing dramatic bias and others finding minimal impact.

This comparative analysis is particularly timely given ongoing debates about polygraph admissibility in legal proceedings and the profession's efforts to establish quality control standards. Understanding why these three methodologically similar studies produced different results has important implications for blind scoring protocols, examiner training, and the actual risk posed by prior information in field practice.

Research Design & Methodology

The paper compares three studies examining confirmation bias in polygraph scoring. In the original Elaad et al. (1994) study, two experiments manipulated prior expectations to determine whether and to what extent polygraph examiners are affected by their prior expectations when analyzing and interpreting polygraph charts. Examiners scored charts after being told examinees had confessed or been proven innocent.

The Krapohl & Dutton (2018) study recruited volunteer scorers through the American Polygraph Association. Examiners blindly evaluated 15 polygraph cases, some ambiguous and others which clearly indicated truthfulness or deception, with half told all cases were confirmed deceptive while the other half were told all were verified truthful, finding that on average, polygraph scores and decisions were shifted in the direction of the biasing information.

The Ginton (2019) reanalysis took an "applied psychology" perspective, re-evaluating the 1994 data to estimate real-world impact. The reanalysis found that the estimated impact of prior expectations on real-life polygraph decisions would in fact adversely impact just 3% of the total volume of polygraph examinations, suggesting an applied perspective rather than being satisfied with the basic science approach.

Results & Key Findings

Ginton concluded that the sample of examiners in Krapohl and Dutton's study (2018) was a better representative of the examiners' population; however, the robust effect found in it is somewhat questionable since an unspecified part of it could be attributed to uncontrolled contamination of a conformity factor, leaving the findings of the other two studies, which indicate a smaller effect.

The comparative analysis revealed several key patterns across the three studies:

• Prior expectations affected examiners' judgments when polygraph charts did not include clear indications of guilt or innocence, but when objective physiological evidence included strong indications, the effect diminished
• The adverse effect in practice may concern less than 5% of the specific event-related CQT examinations
• Using the conservative inconclusive zone of +/-5 in numerical scoring might mitigate the prior information impact by reducing the possibility that it may change results from Deception Indicated (DI) to No Deception Indicated (NDI)
• With such cut scores, at most, the effect would be shown in entering or leaving the Inconclusive zone, which is less problematic in increasing the number of potential errors, though the danger of being affected by prior information is still there and should be paid attention to

Discussion & Significance

Ginton's analysis suggests that methodological differences—particularly regarding examiner sampling, the presence of conformity factors, and the ambiguity of test materials—account for the divergent findings. The Krapohl & Dutton study's use of American Polygraph Association volunteers may have introduced social conformity pressures, inflating the apparent bias effect beyond what occurs in routine practice. This is a critical distinction: laboratory demonstrations of bias susceptibility may not reflect the frequency or magnitude of bias in actual field conditions.

The finding that research evidence indicates that prior information affects only a small percentage of the total volume of field CQT tests challenges the narrative that confirmation bias fatally undermines polygraph validity. However, this should not be interpreted as dismissing the concern entirely. Even a 3-5% error rate attributable to bias represents real cases where innocent individuals might be incorrectly judged deceptive or guilty parties assessed as truthful.

The analysis highlights an important tension in applied behavioral science research: laboratory findings demonstrating a phenomenon's existence don't automatically translate into estimates of its real-world frequency or impact. This has broader implications beyond polygraphy for how forensic sciences interpret and apply psychological research on cognitive biases.

Limitations & Considerations

The primary limitation acknowledged is that the robust effect found in Krapohl and Dutton's study is somewhat questionable since an unspecified part of it could be attributed to uncontrolled contamination of a conformity factor. This uncertainty makes it difficult to determine which study provides the most accurate estimate of real-world bias effects. The possibility that social conformity influenced the Krapohl & Dutton results suggests that volunteer-based research designs may not be ideal for studying examiner bias.

Additionally, all three studies examined primarily inconclusive or ambiguous charts. Prior expectations affected examiners' judgments when charts did not include clear indications of guilt or innocence, meaning the findings may not generalize to the majority of field cases where physiological responses are more definitive. The proportion of truly ambiguous cases in actual practice remains an important but unanswered question.

Practical Applications

For polygraph practitioners, this analysis provides both reassurance and caution. The evidence suggests that blind scoring protocols and strict information management—while important—may not need to be implemented with the urgency that laboratory studies alone might suggest. Using the conservative inconclusive zone of +/-5 in numerical scoring might mitigate the prior information impact by reducing the possibility that it may change results from Deception Indicated to No Deception Indicated and vice versa.

However, the danger of being affected by prior information is still there and should be paid attention to. Quality assurance programs should focus on the subset of cases most vulnerable to bias: those with ambiguous physiological data where examiner judgment plays a larger role. Independent blind scoring may be most valuable precisely in these borderline cases where prior information could tip the decision. The analysis ultimately supports a risk-management approach rather than categorical rejection of CQT methodology based on bias concerns.