Are polygraph tests accurate? What science says

  //science-technology.news-articles.net/content/2 .. -polygraph-tests-accurate-what-science-says.html
  Published in Science and Technology on Saturday, August 30th 2025 at 16:56 GMT

Polygraphs Under Scrutiny: What the Science Really Says About Their Accuracy

By a research journalist
August 30, 2025

When a suspect sits in a chair and the technician presses a pen against a pulse‑monitoring cuff, the image of a polygraph test often conjures up a dramatic courtroom moment. In popular culture the “lie detector” is treated as a near‑infallible tool that can separate truth from deception. The reality, however, is far messier. A recent KTLA piece titled “Are Polygraph Tests Accurate? What Science Says” dives deep into the scientific literature, FBI guidelines, and courtroom rulings to paint a nuanced picture of the device’s strengths and limits. Below is a thorough recap of that analysis, supplemented by insights from the additional links cited in the article.

1. The Mechanics of the Polygraph

A polygraph records several physiological responses—heart rate, blood pressure, respiration, and skin conductance—while a subject answers a series of questions. The underlying hypothesis is that lying elicits a detectable “stress response” that can be distinguished from truthful answering. The technician interprets the data in real time, marking “genuine” vs. “non‑genuine” responses. Yet the process is inherently subjective: there is no objective, quantitative threshold that guarantees a correct verdict.

The KTLA article explains that the “classic” polygraph is based on the “psychophysiological model” developed by John B. Watson in the 1920s. Subsequent iterations, such as the “multivariate” or “cross‑modal” methods, aim to reduce examiner bias but still rely on the same physiological premise.

2. What the Scientific Community Claims

2.1 Accuracy in Controlled Settings

A pivotal study from the University of Texas (2014) is highlighted in the piece: over 400 participants underwent a “controlled‑truth” protocol. Researchers reported an average “accuracy” of roughly 71 % when experts judged the answers. This figure sits somewhere between the commonly cited “80–90 %” range and the more conservative “60‑65 %” range presented by other scholars.

Critics note that many of these studies suffer from selection bias—subjects are often volunteers who are not representative of the general population, and the “truthful” answers are scripted. The KTLA article cites the 2011 National Research Council (NRC) report “The Polygraph: An Evaluation of the Evidence,” which concluded that the polygraph’s sensitivity (true‑positive rate) is about 70 %, while the specificity (true‑negative rate) hovers around 63 %. This means that roughly one in four true liars and one in three true truth‑tellers can be misclassified.

2.2 Real‑World Accuracy

The article references a 2015 FBI internal study where 3,000+ polygraph results were cross‑checked against post‑test investigations. The FBI reported an overall accuracy rate of 75 % for detecting deception in criminal investigations. However, the investigators’ own assessment revealed that “many of the cases were complex, and the examiner’s conclusions often hinged on interpretive judgment rather than hard data.” Importantly, the FBI’s own guidelines caution that polygraph results are non‑admissible in federal courts—yet are frequently used in internal investigations and hiring.

3. Legal and Institutional Standards

The piece explains that most U.S. jurisdictions, including California where KTLA is based, forbid the use of polygraph evidence in criminal trials. This stems from a 1988 Supreme Court ruling that the Daubert standard applied to polygraph data, effectively limiting its admissibility due to questions of reliability.

Despite this, agencies like the FBI and the Department of Justice (DoJ) have formalized standards for polygraph use. The FBI’s “Polygraph Examination Rules” include stringent examiner training (a minimum of 150 hours of practice), and a requirement that each examination be recorded and reviewed by an independent examiner. However, even the FBI’s own guidelines acknowledge the possibility of false positives and advise that results be considered only as a “tool of evidence” rather than conclusive proof.

The article also points readers to a linked CNN piece, “Why Law Enforcement Still Uses Polygraphs,” which cites that about 70 % of federal agencies use polygraphs for hiring and background checks. This prevalence underscores the disconnect between scientific uncertainty and policy practice.

4. Common Misconceptions Debunked

4.1 “The Polygraph Is 99 % Accurate”

The KTLA article cites a myth‑buster from the American Polygraph Association (APA). The APA claims a 95 % accuracy figure, but independent reviewers note that this number comes from a small sample of “well‑trained” examiners and does not account for examiners’ own bias.

4.2 “The Polygraph Can Detect All Lies”

The piece quotes Dr. Daniel Lykken, a psychologist who has studied deception. Lykken argues that the polygraph cannot detect “every lie”—the device is most effective against “intentional deception” but may fail against subtle or accidental misstatements. Lykken emphasizes that the physiological markers can also be triggered by emotions other than deception, such as anxiety, pain, or excitement.

5. Emerging Technologies and Future Directions

One of the article’s most hopeful sections covers emerging research in machine learning and biometric analytics. A 2023 study from MIT’s Media Lab introduced a neural‑network model that analyzes polygraph data streams in real time, reportedly boosting accuracy to 85 % in lab tests. However, the KTLA piece cautions that such systems still require human oversight and have not been validated across diverse populations.

The article also highlights a link to an academic paper in Nature Communications that examined electroencephalography (EEG) signals during deception. The authors found that combining EEG with traditional polygraph metrics improved detection rates by 12 %. While promising, this technique is still in the experimental stage and not widely available for law‑enforcement use.

6. Bottom Line for Practitioners

The KTLA article concludes that polygraphs should be regarded as low‑certainty tools that can assist in investigations but should not be treated as definitive evidence. The FBI’s own guidelines echo this sentiment: “Polygraph results are best used as a starting point for deeper inquiry, not as a final verdict.” For employers, the key takeaway is that polygraph‑based hiring decisions carry a risk of defamation if the test is inaccurate. The article advises companies to combine polygraph data with structured interviews, reference checks, and, where possible, technological verification (e.g., biometric ID checks).

7. How to Stay Informed

The piece encourages readers to consult reputable sources such as the National Research Council, the FBI’s Polygraph Examination Rules, and peer‑reviewed journals in psychology and forensic science. It also links to an American Psychological Association (APA) statement that advises caution in interpreting polygraph results.

TL;DR

• Polygraphs are not infallible: Accuracy hovers around 70 % in lab studies and 75 % in FBI internal reports, but real‑world conditions lower confidence levels.
• Legal restrictions: Most U.S. courts disallow polygraph evidence due to reliability concerns.
• Policy usage: About 70 % of federal agencies still employ polygraphs for hiring and background checks, despite scientific uncertainty.
• Future tech: Machine learning and EEG integration show promise but are not yet mainstream.
• Practical advice: Treat polygraph results as a supplement to other investigative tools, not a conclusive verdict.

The KTLA article’s comprehensive review reminds us that the intersection of science, law, and public perception is fraught with complexity. While the allure of a simple, objective lie detector is strong, the reality demands a nuanced, evidence‑based approach to interpreting polygraph data.