This guidance post draws on the peer-reviewed literature summarised in Nelson (2015) and the British Polygraph Society’s Standards of Practice. Full citations appear in the Sources list at the end of this page.
Key points
- Polygraph testing is probabilistic, not deterministic — it can produce incorrect results.[1]
- Published diagnostic accuracy averages 0.89 (95 % CI 0.83–0.95), with false-positive and false-negative rates both above zero.[2]
- An inconclusive outcome is a legitimate scientific result, not a “failure”.[3]
- Multiple studies have found no significant difference in accuracy for individuals with psychopathic traits.[4]
- The National Research Council (2003) found no credible scientific evidence that countermeasures easily defeat polygraph and a trained examiner.[5]
The short answer: yes, a polygraph can be wrong
A polygraph examination measures physiological responses — such as changes in skin conductance, blood pressure, and respiration — and uses validated scoring methods to classify an examinee’s responses as indicative of truth-telling or deception. It does not, however, read minds. Like any diagnostic procedure in science or medicine, it is probabilistic: it provides a well-founded estimate of the likelihood of deception, not an absolute certainty.[1]
This means that errors do occur. Responsible practitioners — and the professional standards that govern them — acknowledge this. The British Polygraph Society’s Standards of Practice require that members must not overstate the reliability of polygraph results.[6] Understanding the nature and frequency of errors is essential for anyone who commissions, undergoes, or relies upon a polygraph examination.
Types of error
Polygraph outcomes fall into three categories — truthful, deceptive, or inconclusive. Errors arise when a result does not match the examinee’s actual condition. There are two principal error types.
False positive (truthful person classified as deceptive)
A false positive occurs when a truthful examinee produces physiological responses that the scoring algorithm classifies as indicative of deception. In the diagnostic context, the published false-positive rate is 0.12; in the screening context, it is 0.14.[7] In practical terms, this means that roughly 12–14 out of every 100 truthful examinees may be incorrectly classified as deceptive, depending on the context.
False negative (deceptive person classified as truthful)
A false negative occurs when a deceptive examinee’s responses do not reach the scoring threshold for a deceptive classification. The published false-negative rate is 0.08 for diagnostic examinations and 0.11 for screening examinations.[8]
Inconclusive results
An inconclusive outcome is not, strictly speaking, an error — it is a third category in which the data do not support a classification in either direction. Inconclusive rates reported in the literature are 0.13 for truthful examinees and 0.09 for deceptive examinees in the diagnostic context.[9] We address what inconclusive results mean in a dedicated section below.
Published accuracy rates
The most comprehensive quantitative summary of polygraph accuracy available in the peer-reviewed literature is the 2011 meta-analysis conducted by the American Polygraph Association (APA), reported and discussed at length in Nelson (2015). It synthesised data across two broad testing contexts.
Accuracy at a glance
| Measure | Diagnostic | Screening |
|---|---|---|
| Mean accuracy | 0.89 (CI 0.83–0.95) | 0.85 (CI 0.77–0.93) |
| Sensitivity (correct deceptive) | 0.84 (CI 0.73–0.93) | 0.77 (CI 0.60–0.90) |
| Specificity (correct truthful) | 0.77 (CI 0.65–0.85) | 0.72 (CI 0.63–0.81) |
| PPV | 0.89 (CI 0.81–0.99) | — |
| NPV | 0.91 (CI 0.82–0.99) | — |
| False-positive rate | 0.12 | 0.14 |
| False-negative rate | 0.08 | 0.11 |
| Studies included | 24 (n = 8,975) | 14 (n = 1,194) |
Data from the APA 2011 meta-analysis as reported in Nelson (2015).[2]
Several points merit emphasis. First, the confidence intervals show that accuracy is not a single fixed number; it varies across studies and techniques. Second, sensitivity (the ability to detect deception correctly) exceeds specificity (the ability to classify truthful examinees correctly) in both contexts, meaning that the test is somewhat more likely to err against a truthful person than to miss a deceptive one. Third, the positive predictive value (PPV) and negative predictive value (NPV) in diagnostic testing are both above 0.89, indicating that when a result is rendered, it is correct the large majority of the time.[10]
What affects accuracy
Several factors influence whether a given examination achieves accuracy rates consistent with the published research.
Examiner training and competence
The quality of the examiner is a critical variable. The BPS Standards of Practice require members to use evidence-based validated techniques.[11] Proper training ensures that question formulation, pre-test procedures, data collection, and scoring are all conducted in accordance with published protocols. Poor training or deviation from validated methods undermines the basis on which published accuracy figures rest.
Validated techniques and scoring
Not all polygraph question formats are equally well supported by the evidence. The BPS Standards of Practice set explicit validation thresholds: evidentiary testing requires accuracy of at least 90 %, paired testing at least 86 %, investigative testing at least 80 %, and screening testing must perform significantly above chance.[12] Techniques that have not been validated to these standards should not be relied upon.
Examinee suitability
Certain medical conditions, acute psychological states, or pharmacological effects may affect an examinee’s physiological responsivity. A competent examiner assesses suitability before proceeding and may decline to conduct an examination if conditions are not met. This protects against unreliable data being scored and reported.
Instrumentation
Modern computerised polygraph instruments record multiple physiological channels with greater precision and consistency than earlier analogue systems. The use of properly calibrated, maintained equipment is a prerequisite for valid data collection.
Common myths debunked
Myth 1: “Psychopaths can beat the polygraph”
This is one of the most persistent claims about polygraph testing, yet the evidence does not support it. Three separate studies — Barland & Raskin (1975), Raskin & Hare (1978), and Patrick & Iacono (1989) — all found no significant differences in polygraph accuracy for individuals with psychopathic personality traits.[4] The NRC (2003) likewise concluded that there is no scientific evidence that any particular personality type affects polygraph accuracy.[13]
Myth 2: “Being nervous will make you fail”
It is entirely normal to be nervous during a polygraph examination, and validated testing formats are designed to account for this. Comparison question techniques (such as probable-lie or directed-lie formats) include comparison questions that are expected to produce a physiological response in truthful examinees. The scoring system evaluates relative differences between responses to relevant and comparison questions, rather than absolute levels of arousal. General nervousness affects responses to all questions approximately equally and is therefore controlled for in the comparison process.
Myth 3: “Countermeasures easily defeat the polygraph”
Countermeasures are deliberate attempts by an examinee to manipulate their physiological responses to influence the test outcome. Despite popular claims, the NRC (2003) found no credible scientific evidence supporting the assertion that examinees can be easily trained to defeat both the polygraph and a trained examiner. In the NRC’s own words: “claims that it is easy to train examinees to beat both the polygraph and trained examiners require scientific supporting evidence to be credible. However, we are not aware of any such research.”[5]
Moreover, truthful examinees who attempt to use countermeasures increase their risk of being incorrectly classified as deceptive — a false-positive outcome.[14]
What an inconclusive result means
An inconclusive result means that the physiological data did not meet the decision threshold for either a truthful or a deceptive classification. It is not a “failure” and should not be treated as an indication of deception.
The BPS Standards of Practice recognise inconclusive as a valid result.[3] In the diagnostic literature, approximately 13 % of truthful examinees and 9 % of deceptive examinees receive inconclusive outcomes.[9] The inconclusive category exists precisely because the scientific method demands that when data are ambiguous, the honest response is to withhold a classification rather than force one that may be wrong.
If you receive an inconclusive result, this may mean that a re-examination is appropriate, that the particular question set was not well suited to you, or simply that your physiological responses on that occasion did not produce sufficiently clear differentiation. Your examiner should explain the outcome to you and discuss any next steps.
What to do if you think your result was wrong
If you believe that your polygraph result does not accurately reflect the truth, you have several options.
1. Ask your examiner for an explanation
A qualified examiner should be willing to explain the basis for the result, including the technique used, the scoring methodology, and how the data were interpreted. Under the BPS Standards of Practice, a result is not final until it has been documented.[15]
2. Request a quality-assurance review
Many polygraph examiners operate within organisations that have quality-assurance processes, including independent chart review by a second qualified examiner. You may ask whether such a review is available.
3. Seek a second opinion or re-examination
You may commission a separate examination from a different qualified examiner. When doing so, look for an examiner who uses evidence-based validated techniques and who adheres to published professional standards.
4. Understand your rights
If your examination was conducted by a BPS member and you believe the examiner acted improperly or failed to follow professional standards, you can consult the Your Rights page for further information, or refer to the Complaints, Discipline and Appeals Procedure.
Important context
A polygraph result — whether truthful, deceptive, or inconclusive — is one piece of information, not a verdict. In most professional and legal contexts, polygraph outcomes are considered alongside other evidence and are not determinative on their own. No responsible examiner or commissioning body should treat a single polygraph result as conclusive proof of truth or deception.
Glossary
- APA
- American Polygraph Association. A US-based professional membership body for polygraph examiners; conducted the 2011 meta-analysis cited in this article.
- BPS
- British Polygraph Society. A UK unincorporated membership body for polygraph professionals, founded in 2017.
- CI (Confidence Interval)
- A statistical range within which the true value of a measure is expected to fall at a given probability level (usually 95 %). Wider intervals indicate greater uncertainty.
- Comparison Question
- A question in a polygraph test format (such as CQT) designed to elicit a physiological response from truthful examinees. It serves as a baseline against which responses to relevant questions are compared.
- Countermeasures
- Deliberate physical or mental actions taken by an examinee during a polygraph test in an attempt to manipulate physiological responses and influence the outcome.
- CQT (Comparison Question Test)
- A widely used polygraph testing format that compares physiological responses to relevant questions against responses to comparison questions.
- DLC (Directed-Lie Comparison)
- A type of comparison question in which the examinee is directed to give a response known to be untruthful, providing a known-lie baseline for comparison.
- EDA (Electrodermal Activity)
- Changes in the electrical conductivity of the skin, typically measured at the fingertips, caused by sweat-gland activity. One of the physiological channels recorded by a polygraph instrument.
- False Negative (FN)
- An outcome in which a deceptive examinee is incorrectly classified as truthful.
- False Positive (FP)
- An outcome in which a truthful examinee is incorrectly classified as deceptive.
- Inconclusive
- A polygraph result in which the physiological data do not meet the scoring threshold for either a truthful or a deceptive classification.
- Meta-analysis
- A statistical method that combines the results of multiple independent studies to produce a more precise overall estimate of an effect or measure.
- NPV (Negative Predictive Value)
- The probability that an examinee classified as truthful is genuinely truthful. An NPV of 0.91 means that 91 % of truthful classifications are correct.
- NRC
- National Research Council (United States). Published a major review of polygraph research in 2003.
- PLC (Probable-Lie Comparison)
- A type of comparison question designed so that most examinees are likely to respond deceptively to it, providing a probable-lie baseline for comparison.
- PPV (Positive Predictive Value)
- The probability that an examinee classified as deceptive is genuinely deceptive. A PPV of 0.89 means that 89 % of deceptive classifications are correct.
- Sensitivity
- The proportion of truly deceptive examinees who are correctly identified as deceptive by the test.
- Specificity
- The proportion of truly truthful examinees who are correctly identified as truthful by the test.
Sources
- Nelson, R. (2015). ‘Scientific Basis for Polygraph Testing’, Polygraph, 44(1), pp. 28–61. Available at: polygraph.org (PDF).
- British Polygraph Society (2020). Standards of Practice, v1.0. Available at: polygraph.org.uk.
This post is a general-information guide for members of the public and commissioning organisations. It does not constitute legal advice. Polygraph results should always be considered within their proper professional and legal context. If you require specific advice about a polygraph examination, consult a qualified examiner or seek independent legal counsel. The British Polygraph Society is an unincorporated UK membership body and is not a statutory regulator.