Is the Guilty Knowledge Polygraph Technique Applicable in Criminal Investigations? A Review of FBI Case Records

John A. Podlesny's 1993 FBI case-record review examined whether the Guilty Knowledge Technique (GKT) — a polygraph method that detects crime-specific knowledge rather than direct deception — could realistically replace the Comparison Question Test in FBI criminal investigations. By analyzing closed case files for the presence of sufficient, operationally viable crime-scene details, the study found that the GKT could have been applied in only approximately 13.1% of examined cases (95% CI: 4.9%–21.4%), with no basis for treating it as a general substitute for existing methods.

Retrospective review of closed FBI criminal case files in which CQT examinations had previously been administered, assessing each case for the presence of sufficient, containable crime-specific facts ("keys") necessary to construct a valid GKT examination, with findings reported as proportional estimates with 95% confidence intervals.

Podlesny conducted a unique field-based assessment of GKT applicability by reviewing actual closed FBI cases where polygraph examinations had been conducted. The study addressed the critical gap between laboratory demonstrations of GKT effectiveness and real-world implementation challenges. Each case was evaluated for the availability of sufficient 'operable case facts' - discrete, crime-specific details that could serve as GKT stimulus items without contamination. The research revealed significant practical constraints that limit GKT deployment in actual criminal investigations, contrasting with the technique's strong theoretical foundation and laboratory performance.

This study provided crucial empirical evidence about the practical limitations of implementing GKT in real criminal investigations, informing policy decisions about polygraph technique selection and highlighting the need for techniques that can function effectively within actual investigative constraints.

Background & Context

By the early 1990s, the Guilty Knowledge Technique (GKT) — also called the Concealed Information Test (CIT) — had attracted considerable scientific interest as a theoretically elegant alternative to the widely used Comparison Question Test (CQT). The Guilty Knowledge Technique involves the detection of knowledge about a crime as a means for inferring deception, and it had been promoted as an alternative or replacement for the commonly used Control Question Technique in investigative polygraph examinations. Rather than asking whether a suspect "did it," the GKT probes whether a suspect recognizes specific crime-scene details that only a true perpetrator would know.

Despite its scientific appeal, a critical gap existed between theory and practice. Guilty knowledge tests had rarely been used in actual investigations, and some authors had suggested that practical considerations prevent their widespread use. Advocates praised the technique's logical foundation and its strong performance in controlled laboratory studies, but skeptics questioned whether real criminal cases would ever supply the precise, containable crime-scene details the GKT requires to function properly.

John A. Podlesny, a researcher embedded within the FBI's operational polygraph program, was uniquely positioned to answer this question empirically. By turning to actual closed FBI case files rather than hypothetical scenarios, his 1993 study delivered one of the first rigorous, data-driven assessments of the GKT's practical reach — a contribution that would become a touchstone in the field for decades.

Research Design & Methodology

This study reviewed a sample of closed criminal cases in which control-question polygraph examinations had been used, to assess the possible applicability of the Guilty Knowledge Technique in those cases. By anchoring the analysis in real, completed FBI investigations, Podlesny avoided the inherent artificiality of laboratory simulations and instead tested GKT applicability against the messy, information-constrained reality of fieldwork.

The core analytical task was to evaluate each case file for the presence of sufficient "operable case facts" — discrete, crime-specific details that could serve as GKT stimulus items. It is difficult to identify a sufficient number of salient features of an event that can be used to formulate proper GKT questions. A proper GKT question refers to a specific feature of the event that is very likely to be noticed by a guilty person, and it is crucial that guilty individuals not only notice the designated feature, but also remember it during the polygraph investigation, which may occur long after the event. Each case was assessed against the minimum threshold of operationally viable GKT items needed to construct a valid examination.

Key design elements of the study included:

• Data source: Closed FBI criminal case files in which CQT examinations had previously been administered
• Unit of analysis: Individual polygraph examination cases, assessed for GKT "key" availability
• Applicability criterion: Whether sufficient, uncontaminated, crime-specific details existed to formulate a valid GKT
• Statistical reporting: Proportional estimates with a 95% confidence interval to reflect sampling uncertainty
• Publication venue: Originally published in the FBI Crime Laboratory Digest (July 1993); later reprinted in Polygraph, Vol. 23, No. 1 (1994), pp. 85–94

The retrospective case-review design was methodologically appropriate for the study's purpose: rather than staging artificial crimes, it asked whether the GKT could have been deployed within the actual investigative constraints that FBI examiners face every day.

Results & Key Findings

The headline finding was stark: guilty knowledge tests could have been feasibly used in only approximately 13.1% of the examined cases. The current study found that guilty-knowledge tests might have been used in 13.1 percent of the examinations, with a 95% confidence interval for similarly selected examinations of 4.9 percent to 21.4 percent. This wide confidence interval itself underscores how difficult it is to generalize across the heterogeneous landscape of FBI criminal investigations.

The remaining cases did not simply lack ideal GKT facts — most were structurally incompatible with the technique. A large proportion of the remaining examinations were conducted under circumstances that would not be amenable to the use of guilty-knowledge tests, even with much effort, and the results indicate that the Guilty Knowledge Technique might be applied in a small proportion of FBI cases, but there is no basis to consider it as a general replacement for current methods.

Key quantitative and qualitative findings included:

• ~13.1% of cases had sufficient operable facts to construct a valid GKT examination
• 95% CI of 4.9%–21.4% — reflecting genuine uncertainty about broader applicability
• The majority of cases lacked the minimum number of containable, crime-specific details required
• No basis found for treating the GKT as a general substitute for direct-questioning (CQT) methods
• Recent reports of novel physiological parameters and methods of analysis for use with guilty-knowledge tests may be of limited value in criminal investigations.

Podlesny also flagged that his estimates should be considered a somewhat conservative lower bound, because incomplete file records may have obscured cases where GKT keys were technically present but not documented. In difficult cases (otherwise a CQT would not have been necessary), a CIT would have been possible in approximately 10% of the cases, and this figure should be regarded as the lower bound estimate as it is based on incomplete file records.

Discussion & Significance

Podlesny's findings entered a live scientific debate. Laboratory research had made the GKT look highly promising — mock crime studies revealed average detection rates of 80.6% for guilty examinees and 95.9% for the innocent, and in 11 of these studies no false positive identifications were observed. Yet impressive laboratory accuracy means little if the technique cannot be applied in actual investigations. Podlesny's study provided the crucial reality check, demonstrating that the operational preconditions for GKT administration are met in only a narrow slice of real FBI casework.

The study's significance extended well beyond the FBI. It catalyzed a richer understanding of the gap between a technique's theoretical validity and its operational feasibility. The results indicate that the guilty knowledge technique cannot be used in the majority of FBI criminal investigations in which deception detection is used, and the guilty knowledge technique should not be used as a general alternative to direct-questioning methods; however, where applicable, guilty knowledge techniques may be used as a supplement to other methods, possibly adding protection against false positives.

Scholars continued to engage with Podlesny's data even a decade later. Some reanalysts argued that even a ~13% applicability rate — in cases difficult enough to require polygraph — represents a meaningful proportion of law enforcement use cases. In a regular forensic context, the applicability of the CIT seems amply enough to yield a positive cost/benefit analysis, and the large-scale use of the CIT in Japan has shown that successful implementation is indeed possible. This cross-national contrast became increasingly important as Japan built an extensive, operationally successful GKT infrastructure while U.S. investigators rarely attempted the technique.

Limitations & Considerations

The study's retrospective design introduced inherent constraints. Because the case files were assembled for CQT examinations — not for GKT planning — the documentation of crime-scene details may have been systematically incomplete. Facts that would enable a GKT might have existed but gone unrecorded, meaning the 13.1% estimate may understate true potential applicability. The narrow confidence interval remains wide enough to encompass a fourfold range of possible outcomes (4.9% to 21.4%), reflecting real uncertainty about what a broader or differently selected sample would reveal.

Additional limitations acknowledged by researchers engaging with this work include:

• Sample scope: Cases were drawn specifically from FBI investigations — a federal agency with distinct case types — limiting generalizability to state, local, or international law enforcement
• Information leakage risk: Once salient features of the event have been identified, it is very difficult to keep them from being leaked to innocent examinees, and leakage of the critical items might put innocent suspects in great danger because knowledge of the critical items might be sufficient for producing differential responding to these items
• Retrospective bias: Cases reviewed were already closed, potentially skewing toward case types where CQT was clearly the dominant tool
• Evolving investigative practices: The study predates coordinated crime-scene analysis protocols that might systematically identify GKT-eligible facts earlier in an investigation

Practical Applications

Podlesny's findings carry clear operational guidance for polygraph examiners and investigators: the GKT should be treated as a supplementary tool in a specialized subset of cases, not a universal replacement for comparison-question methods. For U.S. law enforcement agencies to increase the frequency of use of the CIT, there must be a coordinated effort by investigators and polygraph examiners to facilitate its use in criminal investigations. This means identifying crime-scene "keys" early and systematically, and rigorously protecting those details from media or investigative leakage before examination.

The study also has important implications for examiner training and case triage. Examiners reviewing case assignments should actively assess GKT eligibility at intake — not as an afterthought. The GKT is particularly useful when a second polygraph examination is required, such as in a child abduction leading to a homicide investigation. For examiners operating within agencies that already conduct systematic crime-scene analysis, this research argues that even partial adoption of GKT protocols — in that eligible minority of cases — can meaningfully strengthen investigative outcomes and reduce false-positive risk for innocent suspects.