Diagnostic vs. Screening Polygraphs: Sensitivity & Specificity

Why Polygraph Tests Are Not All the Same

Polygraph testing encompasses a range of test types, each designed for specific operational contexts with distinct statistical properties, question formats, and decision rules. At the highest level, polygraph examinations are classified into two primary categories: diagnostic tests and screening tests [1]Verified Practical Polygraph: How to Parse Categorical Results for Test Questions of Diagnostic and Screening Polygraphs
Establishes that the Grand Total Rule applies to diagnostic and Subtotal Score Rule to screening examinations. Diagnostic tests are conducted in response to known problems, known incidents, or known allegations, while screening tests are conducted in the absence of a known problem and are intended to search for possible problems [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations.

Understanding the fundamental differences between these two categories has profound implications for how test results are interpreted, what decisions can be made based on those results, and how errors are managed. The American Polygraph Association (APA) has long recognized this distinction in its standards and practices. Raymond Nelson's "Five Minute Science Lesson" column in the APA Magazine has explored many of the critical statistical concepts underpinning these distinctions across multiple issues [3]Verified Five Minute Science Lesson (column series)
Confirms Raymond Nelson's recurring column in APA Magazine covering statistical foundations of polygraph science, covering topics from test accuracy metrics to successive hurdles screening and measurement theory.

The APA's 2011 meta-analytic survey — the most comprehensive review of polygraph accuracy to date — analyzed 38 studies that satisfied qualitative and quantitative requirements for inclusion, involving 32 different samples and 45 different experiments, with 295 scorers providing 11,737 scored results of 3,723 examinations [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. This landmark study established differentiated accuracy benchmarks for diagnostic and screening polygraph techniques that remain the foundation of evidence-based polygraph practice.

For practitioners, understanding these concepts is essential for selecting the appropriate test format, interpreting results accurately, and communicating findings to stakeholders. Our guide to validated polygraph techniques and APA compliance provides additional context on how these standards are applied in practice.

Purpose and Context of Diagnostic Polygraphs

A diagnostic polygraph is administered when there is a known or alleged specific incident under investigation. The test is designed to determine whether a particular individual is being deceptive about their involvement in, or knowledge of, that specific event [1]Verified Practical Polygraph: How to Parse Categorical Results for Test Questions of Diagnostic and Screening Polygraphs
Establishes that the Grand Total Rule applies to diagnostic and Subtotal Score Rule to screening examinations. Common scenarios include criminal investigations involving specific crimes (theft, assault, homicide), workplace theft investigations where inventory discrepancies have been identified, false accusation cases, infidelity investigations involving specific incidents, and insurance fraud cases.

The defining characteristic of a diagnostic test is its specificity of focus. Because the examiner and the examinee both know exactly what incident is being addressed, the relevant questions can be crafted with high precision. For example, rather than asking a broad question like "Have you ever stolen anything from an employer?" a diagnostic test would ask something like "Did you take the $5,000 that went missing from the safe on March 15th?" This specificity enables cleaner physiological responses and more straightforward interpretation of results. For more on how specific incidents are investigated through polygraph, see our guide on marital trust issues and lie detector testing.

Question Structure in Diagnostic Tests

Diagnostic polygraph examinations typically employ the Comparison Question Test (CQT) — also historically called the Control Question Test. The CQT was first developed in 1947 and remains the most widely used forensic polygraph method [4]Verified Do 'lie detectors' work? What psychological science says about polygraphs
Confirms the CQT was first developed in 1947 and provides context on polygraph use in law enforcement and screening. In this format, the test includes three types of questions:

Relevant questions directly address the specific incident under investigation. Comparison questions are broader questions designed to evoke a physiological response in truthful examinees. Irrelevant questions are neutral baseline questions that are not scored.

The physiological responses to relevant questions are compared against responses to comparison questions. A deceptive examinee is expected to show stronger reactions to relevant questions, while a truthful examinee is expected to show relatively stronger reactions to comparison questions [4]Verified Do 'lie detectors' work? What psychological science says about polygraphs
Confirms the CQT was first developed in 1947 and provides context on polygraph use in law enforcement and screening. Research by Shaw (2012) found little to no meaningful difference in detection accuracy between directed lie and probable lie comparison questions, supporting the scientific validity of directed lie comparison questions in both diagnostic and screening contexts [5]Verified The Empirical Basis for the Use of Directed Lie Comparison Questions in Diagnostic and Screening Polygraphs
Confirms no meaningful difference in detection accuracy between directed lie and probable lie comparison questions.

Single-issue diagnostic tests focus all relevant questions on one specific topic, which concentrates the examinee's psychological attention and typically produces cleaner data. For diagnostic examinations, the Grand Total Rule should be used when all relevant questions address the same issue, as detailed in Blalock's (2019) practical guidance on parsing categorical results [6]Verified A Validation and Reliability Study of Counterintelligence Screening Tests
Found significantly better than chance detection rates for counterintelligence screening with higher inconclusive rates than specific-issue tests. The 3-position scale scoring method provides a standardized framework for numerical evaluation.

Accuracy of Diagnostic Polygraphs

The APA's 2011 Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques found that event-specific (single issue) diagnostic testing techniques produced an aggregated decision accuracy rate of.890 (95% confidence interval:.829–.951), with a combined inconclusive rate of.110 [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. These findings were based on comparison question techniques in which the criterion variance of multiple relevant questions is assumed to be non-independent.

A 2021 comprehensive meta-analysis by Honts et al., published in Applied Cognitive Psychology, analyzed 138 datasets of Comparison Question Test studies and found that reviews have "generally produced overall accuracy estimates of over 85%" [7]Verified A comprehensive meta-analysis of the comparison question polygraph test
Confirms reviews have generally produced overall accuracy estimates of over 85% for the CQT across 138 datasets. Crewson's (2001) Department of Defense study found that for field polygraph examinations, diagnostic accuracy (.88) was comparable to established medical imaging technologies including MRI, CT scan, and ultrasound [8]Verified A comparative analysis of polygraph with other screening and diagnostic tools
Confirms field screening sensitivity of.59 and specificity of.92; inter-rater reliability kappa of.77 exceeding physicians. The inter-rater reliability among polygraph examiners (kappa =.77) exceeded that of physicians (.56) and approached psychologists (.79) [8]Verified A comparative analysis of polygraph with other screening and diagnostic tools
Confirms field screening sensitivity of.59 and specificity of.92; inter-rater reliability kappa of.77 exceeding physicians.

As Amsel (2013) confirmed in a comparative review published in European Polygraph, polygraph diagnostic accuracy was on par with widely accepted medical diagnostic tools [9]Verified Comparative Review of Polygraph and Other Diagnostic Tools and Methods
Confirms polygraph diagnostic accuracy (.88) comparable to MRI, CT scan, and ultrasound. Nelson's (2015) scientific basis review reported that event-specific diagnostic polygraphs provide a mean accuracy of.89 with a 95% confidence range from.83 to.95 [10]Verified Scientific Basis for Polygraph Testing
Confirms mean diagnostic accuracy of.89 (CI:.83-.95) and screening accuracy of.85 (CI:.77-.93); details response independence assumptions.

Several factors contribute to the generally higher accuracy of diagnostic tests compared to screening tests: the examinee's attention is focused on a single well-defined issue, relevant questions can be precisely targeted, the pre-test interview can thoroughly explore the specific incident, response dependence among related questions amplifies correct classification, and base rates of deception in diagnostic populations are often higher, improving positive predictive value. Our guide on evaluating polygraph data through automated scoring explains how these numerical scores are processed.

Purpose and Context of Screening Polygraphs

Screening polygraphs are fundamentally different from diagnostic tests in their operational purpose. Rather than investigating a known incident, screening tests are administered proactively to assess risk in the absence of any specific allegation or event [1]Verified Practical Polygraph: How to Parse Categorical Results for Test Questions of Diagnostic and Screening Polygraphs
Establishes that the Grand Total Rule applies to diagnostic and Subtotal Score Rule to screening examinations. The goal is to identify potential problems, risks, or undisclosed behaviors before they manifest as actual incidents.

Screening polygraphs are widely used in pre-employment screening for law enforcement agencies, national security clearance processes for intelligence agencies, post-conviction sex offender testing (PCSOT) programs, periodic security reinvestigations for employees with access to classified information, and corporate pre-employment screening. Agencies including the FBI, CIA, CBP, and DEA all use variations of screening polygraph examinations, making this category arguably the most consequential in terms of the volume of people affected [4]Verified Do 'lie detectors' work? What psychological science says about polygraphs
Confirms the CQT was first developed in 1947 and provides context on polygraph use in law enforcement and screening. For more on the PCSOT application, see the Georgia Sex Offender Polygraph Program guide.

Because screening tests are not tied to a specific incident, the questions must necessarily be broader. The National Research Council (2003) noted that screening raises particular theoretical issues because "the relevant questions must be generic" when the examiner does not have a specific event to ask about [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations. Such generic questions create more ambiguity for examinees than specific questions, which can affect response clarity.

Multi-Issue Screening: The Operational Reality

Most screening polygraph examinations are multi-issue tests — they address several distinct behavioral categories within a single examination session. For example, a pre-employment screening for a federal law enforcement position might cover illegal drug use, criminal history, falsification of application materials, foreign contacts or allegiances, and financial integrity.

Each of these topics represents a separate "issue" within the screening test, and each generates its own subtotal score that contributes to the overall examination outcome. For screening examinations, Blalock (2019) established that the Subtotal Score Rule is appropriate when relevant questions address independent issues [6]Verified A Validation and Reliability Study of Counterintelligence Screening Tests
Found significantly better than chance detection rates for counterintelligence screening with higher inconclusive rates than specific-issue tests. This multi-issue structure creates unique statistical challenges related to response independence and multiplicity effects.

Barland's (1985) validation study of counterintelligence screening tests for the US military found significantly better than chance detection rates, though with higher inconclusive rates than specific-issue tests [12]Verified Relevant-Irrelevant Screening Format Accuracy
Concluded that RIR screening format decision accuracy is substantially poorer than other validated screening formats. Meanwhile, Krapohl and Rosales (2014) concluded that the decision accuracy of the Relevant-Irrelevant (RIR) screening format is substantially poorer than other validated screening formats, providing evidence-based justification for agencies to adopt superior testing methods [13]Verified Decision Accuracy for the Relevant-Irrelevant Screening Test: Influence of an Algorithm on Human Decision-Making
Investigated whether algorithmic support could resolve interrater reliability problems in RI screening tests. Krapohl and Goodson (2015) further investigated whether algorithmic support could resolve fundamental reliability issues in RI screening tests [14]Verified Studies of the accuracy of security screening polygraph examinations
Found 94.9% accuracy for innocent examinees but only 43.2% for guilty examinees in security screening.

Accuracy of Screening Polygraphs

The APA's 2011 meta-analysis found that polygraph techniques in which multiple issues were encompassed by the relevant questions produced an aggregated decision accuracy of 85% (confidence interval: 77%–93%) with an inconclusive rate of 13% [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. Nelson (2015) confirmed these figures, reporting that multi-issue screening polygraphs provide accuracy rates with a mean of.85 and a 95% confidence range of.77 to.93 [10]Verified Scientific Basis for Polygraph Testing
Confirms mean diagnostic accuracy of.89 (CI:.83-.95) and screening accuracy of.85 (CI:.77-.93); details response independence assumptions.

Crewson's (2001) Department of Defense study provided additional granularity for screening contexts, finding that for field screening assessments, polygraph demonstrated a specificity of.92 but a sensitivity of.59 [8]Verified A comparative analysis of polygraph with other screening and diagnostic tools
Confirms field screening sensitivity of.59 and specificity of.92; inter-rater reliability kappa of.77 exceeding physicians. This asymmetry — high specificity but lower sensitivity in screening — was also observed by Barland, Honts, and Barger (1989), whose security screening study found 94.9% accuracy for innocent examinees but only 43.2% accuracy for guilty examinees [15]Verified 2015 Update to the APA 2011 meta-analytic survey of validated polygraph techniques
Confirms and updates accuracy figures from the original 2011 meta-analysis; addresses base rate computations.

Several structural factors contribute to the challenges in screening accuracy: broader questions produce less focused physiological responses, multiple issues can divide attention and dilute response clarity, lower base rates of deception in screening populations affect predictive value, and response fatigue from longer test sessions can lead to physiological habituation. The APA acknowledged the paucity of available research in the screening area and endorsed the use of a successive-hurdles approach to minimize errors [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. For strategies to address these challenges, see our guide on minimizing false positives and negatives in polygraph testing.

Defining the Core Metrics of Polygraph Accuracy

Sensitivity and specificity are the two fundamental metrics used to evaluate the accuracy of any binary classification test — whether medical, forensic, or psychophysiological. In polygraph testing, these metrics define how well the test distinguishes between deceptive and truthful examinees. The National Research Council (2003) emphasized that data should be reported in a way that makes it possible to calculate both sensitivity and specificity, preferably at multiple thresholds [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations.

Sensitivity (the true positive rate) measures the probability that the polygraph correctly identifies a deceptive individual as deceptive. Mathematically: Sensitivity = True Positives / (True Positives + False Negatives). A test with high sensitivity catches most deceptive individuals. If a diagnostic polygraph has a sensitivity of 90%, it means that out of 100 genuinely deceptive examinees, approximately 90 will be correctly classified as deceptive, while 10 will be incorrectly classified as non-deceptive (false negatives). In investigative contexts, false negatives are particularly concerning because they allow guilty individuals to evade detection.

Specificity (the true negative rate) measures the probability that the polygraph correctly identifies a truthful individual as truthful. Mathematically: Specificity = True Negatives / (True Negatives + False Positives). Crewson (2001) found that field polygraph screening demonstrated notably high specificity (.92), meaning the test is particularly effective at correctly clearing truthful individuals [8]Verified A comparative analysis of polygraph with other screening and diagnostic tools
Confirms field screening sensitivity of.59 and specificity of.92; inter-rater reliability kappa of.77 exceeding physicians. For more on how baseline testing helps establish accurate response patterns, see our dedicated guide.

The Sensitivity-Specificity Trade-Off

In virtually all classification tests, there is an inherent trade-off between sensitivity and specificity. Adjusting the decision threshold to catch more deceptive individuals (increasing sensitivity) will inevitably cause more truthful individuals to be misclassified (decreasing specificity), and vice versa. This relationship is visualized using Receiver Operating Characteristic (ROC) curves, which plot sensitivity against the false positive rate (1 - specificity) across all possible thresholds [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations.

The NRC (2003) report found that the median accuracy index among comparison question tests was 0.85, with an interquartile range from 0.83 to 0.90 [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations. This accuracy index represents the area under the ROC curve, a standardized measure of overall discriminative ability. Signal detection theory distinguishes two independent features contributing to diagnostic performance: test accuracy (overlap of score distributions) and the decision threshold used by the examiner [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations.

Polygraph decision rules must balance competing priorities based on the operational context. In criminal investigations, the consequences of a false negative may justify a threshold that favors sensitivity. In pre-employment screening, the consequences of false positives may justify a threshold favoring specificity. Professional organizations like the APA provide guidelines for cut scores and decision rules that attempt to optimize this balance. Understanding the role of p-values in polygraph testing helps examiners apply these thresholds correctly.

The polygraph reaction window of 10–35 seconds is another critical factor in accurate scoring, as it defines the physiological timeframe within which responses are evaluated against these thresholds.

Why Prevalence of Deception Matters for Test Interpretation

One of the most important concepts in polygraph science is the base rate problem. The base rate refers to the prevalence of deception within the population being examined. Even a test with excellent sensitivity and specificity can produce misleading results when the base rate of deception is very low [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations.

The NRC (2003) demonstrated this powerfully: if the proportion of major security risks in a screened population equals 1 in 1,000 or less, even with optimistic assessments of polygraph accuracy, each correctly identified spy or terrorist would be accompanied by at least hundreds of nondeceptive examinees mislabeled as deceptive [11]Verified The Polygraph and Lie Detection
Confirms median accuracy index of 0.85 for CQT with interquartile range of 0.83-0.90; addresses base rate effects and screening limitations. This mathematical reality is driven by Bayes' theorem and is not a flaw in the test itself — it applies to all diagnostic procedures.

To illustrate: consider a screening polygraph with 90% sensitivity and 90% specificity. In a high base rate diagnostic context (50% deception rate) among 1,000 examinees: 450 true positives, 50 false negatives, 450 true negatives, 50 false positives — yielding a positive predictive value of 90%. In a low base rate screening context (5% deception rate): 45 true positives, 5 false negatives, 855 true negatives, 95 false positives — yielding a positive predictive value of only 32%.

Nelson explored these Bayesian computations in the APA Magazine, showing how different base rates and accuracy levels interact to determine the trustworthiness of specific outcomes [3]Verified Five Minute Science Lesson (column series)
Confirms Raymond Nelson's recurring column in APA Magazine covering statistical foundations of polygraph science. Ginton's work demonstrated through Israeli Police data that the confidence with which a deceptive outcome (DI) can be trusted to be correct is about 0.65, while the posterior probability of a truthful outcome (NDI) being correct reaches 0.98 [16]Verified Criterion Accuracy of Multiple-issue Polygraphs (Multiplicity Effects Study)
Confirms multiplicity effects in serial testing reduce specificity; Bonferroni and Šidák corrections mitigate cumulative error.

This mathematical reality has significant implications for how screening results should be used operationally. A "significant reactions" finding in a screening test should be treated as a flag for further investigation — perhaps through a more focused diagnostic follow-up, additional questioning through techniques like the PEACE model and cognitive interviewing, or corroborating evidence review.

What Is Response Independence?

Response independence is a statistical assumption that an examinee's physiological reaction to one test question does not influence their physiological reaction to another question. In scientific testing, independence refers to the assumption that the criterion variance of each individual test stimulus is not affected by and does not affect the criterion variance of other test stimuli [17]Verified Bonferroni and Šidák Corrections for Multiplicity Effects with Subtotal Scores of Comparison Question Polygraph Tests
Describes application of Bonferroni and Šidák corrections to polygraph subtotal scores for multi-issue tests. This assumption has different implications and different validity depending on whether the test is diagnostic or screening in nature.

Nelson (2015) clarified that relevant questions of event-specific diagnostic polygraphs are constructed with the assumption of non-independent criterion variance — the relevant questions have a common or shared source of response variance because the external criterion states are shared [10]Verified Scientific Basis for Polygraph Testing
Confirms mean diagnostic accuracy of.89 (CI:.83-.95) and screening accuracy of.85 (CI:.77-.93); details response independence assumptions. In contrast, screening polygraphs are constructed with relevant questions describing distinct behaviors for which the external criterion states are assumed to vary independently [10]Verified Scientific Basis for Polygraph Testing
Confirms mean diagnostic accuracy of.89 (CI:.83-.95) and screening accuracy of.85 (CI:.77-.93); details response independence assumptions.

Response Dependence in Diagnostic Tests

In diagnostic polygraph examinations, the relevant questions all address the same specific incident from different angles. For example, in a theft investigation: "Did you take the missing cash?" "Do you know where the missing $2,000 is?" "Were you involved in the theft on January 10th?" Because all these questions target the same underlying truth, the physiological responses are not independent.

This response dependence in diagnostic tests is actually advantageous. The APA's meta-analysis explicitly noted that comparison question techniques intended for event-specific diagnostic testing assume non-independent criterion variance and achieved the highest accuracy rates [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. When a deceptive examinee responds consistently across multiple relevant questions about the same incident, the aggregate score is pushed further from the truthful pattern, making deception easier to detect. Conversely, a truthful examinee shows consistently weaker responses to relevant questions relative to comparison questions.

This is one key reason why single-issue diagnostic tests achieve higher accuracy (89%) than multi-issue tests (85%) [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. The natural dependence among responses to related questions amplifies the signal the examiner seeks to detect. For understanding how this affects practical scoring, see our guide on global evaluation in polygraph.

Response Independence in Screening Tests

In multi-issue screening polygraphs, each issue represents a separate behavioral domain. An examinee might be completely truthful about drug use but deceptive about an undisclosed arrest. Their physiological response to the drug question has no inherent relationship to their response to the criminal history question. Both multi-facet and multi-issue examinations are interpreted with decision rules based on an assumption of criterion independence among the relevant questions [17]Verified Bonferroni and Šidák Corrections for Multiplicity Effects with Subtotal Scores of Comparison Question Polygraph Tests
Describes application of Bonferroni and Šidák corrections to polygraph subtotal scores for multi-issue tests.

However, Nelson (2015) noted that there is evidence that response variance for screening questions is not actually fully independent, citing research by Barland, Honts, and Barger (1989) and Podlesny and Truslow (1993) [10]Verified Scientific Basis for Polygraph Testing
Confirms mean diagnostic accuracy of.89 (CI:.83-.95) and screening accuracy of.85 (CI:.77-.93); details response independence assumptions. For this reason, field practices do not permit both positive and negative test results within a single examination [10]Verified Scientific Basis for Polygraph Testing
Confirms mean diagnostic accuracy of.89 (CI:.83-.95) and screening accuracy of.85 (CI:.77-.93); details response independence assumptions.

When responses to different screening topics are treated as independent, the statistical properties of the test change: each issue functions as a separate mini-test with its own sensitivity and specificity, the overall probability of at least one false positive increases, and statistical corrections may be needed to maintain the desired overall error rate. Understanding how the psychology of lying influences physiological responses across different question types provides additional context for these dynamics.

Understanding Cumulative False Positive Risk

Testing multiple issues in a single screening session creates a statistical phenomenon known as multiplicity effects — the cumulative increase in false positive probability when multiple comparisons are made. As Nelson and colleagues have demonstrated, analytic results from polygraph testing involve multiplicity effects whenever multiple statistical comparisons are used to draw conclusions [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control.

Research on the serial single-issue (SSI) approach to multiple-issue screening found that serial testing can potentially decrease overall test accuracy with a disproportional increase in false positive errors and increased inconclusive results among innocent persons [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control. The greatest observed effect from the SSI approach was a large reduction in test specificity [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control. While some advantages may exist when testing one question at a time, available evidence does not support the complete abandonment of omnibus analysis methods or multiple-issue screening techniques [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control.

The overall effect is that multiple-issue tests can have correspondingly less precision and accuracy compared to single-issue exams [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control. This is a fundamental reason why the APA distinguishes between diagnostic and screening accuracy benchmarks and why the successive hurdles approach — following a screening test with a single-issue diagnostic test before reaching a positive conclusion — is recommended [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations.

Bonferroni and Šidák Corrections in Polygraph Practice

To control cumulative error rates in multi-issue testing, statistical corrections can be applied. The two most commonly referenced in polygraph literature are the Bonferroni and Šidák corrections [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control.

The Bonferroni correction divides the desired significance level by the number of tests being conducted [19]Verified Šidák correction
Explains the Šidák correction is derived by assuming individual tests are independent; slightly less conservative than Bonferroni. For example, if you are testing five separate issues at an alpha of.05, each individual test would need to reach significance at.05/5 =.01 to maintain the desired family-wise error rate. The Šidák correction provides a slightly less conservative alternative and is derived by assuming that individual tests are independent.

Nelson (2015) described how these corrections can be applied either to desired alpha thresholds or to computed statistical results in polygraph testing [18]Verified Bonferroni correction
Explains the Bonferroni correction method for multiple comparisons and family-wise error rate control. The Empirical Scoring System (ESS) and its multinomial update (ESS-M) incorporate multiplicity corrections for cut-scores, providing more objective and stable analysis of psychophysiological outcomes [17]Verified Bonferroni and Šidák Corrections for Multiplicity Effects with Subtotal Scores of Comparison Question Polygraph Tests
Describes application of Bonferroni and Šidák corrections to polygraph subtotal scores for multi-issue tests.

For examiners, understanding these corrections is essential when interpreting multi-issue screening results. The European Polygraph Association's training standards incorporate these statistical principles, and our guide on evaluating polygraph data covers practical implementation.

Key Differences at a Glance

Diagnostic polygraphs are conducted in response to a known incident or allegation and focus on a single specific issue. They use the assumption of non-independent criterion variance, are scored using grand total rules, and achieve an aggregated decision accuracy of 89% (CI: 83%–95%) with an 11% inconclusive rate [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. They benefit from higher base rates of deception, focused questioning, and response dependence that amplifies detection signals.

Screening polygraphs are conducted in the absence of a known problem to search for possible undisclosed issues. They address multiple independent behavioral domains, use the assumption of independent criterion variance, are scored using subtotal score rules [6]Verified A Validation and Reliability Study of Counterintelligence Screening Tests
Found significantly better than chance detection rates for counterintelligence screening with higher inconclusive rates than specific-issue tests, and achieve an aggregated decision accuracy of 85% (CI: 77%–93%) with a 13% inconclusive rate [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. They face challenges from lower base rates, broader questions, multiplicity effects, and the need for statistical corrections.

Both test types record the same physiological channels — respiratory, electrodermal, and cardiovascular activity — and both employ comparison questions. The critical differences lie in their question structure, scoring assumptions, and the statistical frameworks applied to interpret results. Crewson (2001) demonstrated that polygraph's overall combined accuracy was comparable to MRI, CT scan, and ultrasound [8]Verified A comparative analysis of polygraph with other screening and diagnostic tools
Confirms field screening sensitivity of.59 and specificity of.92; inter-rater reliability kappa of.77 exceeding physicians, and Amsel (2013) confirmed this comparison across diagnostic categories [9]Verified Comparative Review of Polygraph and Other Diagnostic Tools and Methods
Confirms polygraph diagnostic accuracy (.88) comparable to MRI, CT scan, and ultrasound.

Selecting the Right Test Format

Choosing between diagnostic and screening formats is not merely a procedural decision — it determines the statistical framework that governs every aspect of test administration, scoring, and interpretation. Examiners must select the format that matches their operational context. When investigating a specific incident, a single-issue diagnostic test provides the highest accuracy and most defensible results. When assessing general risk or suitability, a multi-issue screening format is appropriate, but examiners must account for the additional statistical challenges.

The APA Standards of Practice require that specific-issue testing techniques meet at least 90% accuracy to be considered validated for evidentiary use [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. For screening techniques, the APA endorsed their use when research indicates accuracy significantly greater than chance, with the successive hurdles approach recommended to minimize errors [2]Verified Meta-Analytic Survey of Criterion Accuracy of Validated Polygraph Techniques
Confirms 89% diagnostic accuracy (CI: 83-95%) and 85% multi-issue accuracy (CI: 77-93%) from 38 studies and 3,723 examinations. Examiners preparing for certification should master these distinctions, and mentorship programs provide invaluable guidance.

Communicating Results to Stakeholders

Understanding the statistical distinctions between diagnostic and screening tests is critical when communicating results. A "significant reactions" finding on a multi-issue screening test carries different inferential weight than the same finding on a single-issue diagnostic test. Examiners must convey the base rate context, explain the possibility of false positives in low-prevalence screening populations, and recommend appropriate follow-up procedures.

For legal professionals, these distinctions are vital when evaluating polygraph evidence. The history of polygraph licensing laws shows how different jurisdictions have approached the admissibility question, and understanding sensitivity, specificity, and positive predictive value is essential for any legal challenge or defense of polygraph evidence.

The APA Standard 1.8.3 requires that probabilistic information be provided along with categorical test results for evidentiary examinations [3]Verified Five Minute Science Lesson (column series)
Confirms Raymond Nelson's recurring column in APA Magazine covering statistical foundations of polygraph science, ensuring that stakeholders receive the complete statistical picture rather than a simple pass/fail outcome. For guidelines on what happens after testing in specific legal contexts, see our EFCC polygraph results guide.