A Think Again Media Group publication
Crime Stumper

What does a polygraph actually measure?

Polygraph testing: the bodily signals behind the verdict

By Phil Springer, Editor · Published · Crime Stumper

Polygraph testing: the bodily signals behind the verdict

Polygraph testing begins with bodily measurements. The federal regulation defining the instrument specifies continuous, simultaneous recording of cardiovascular, respiratory and electrodermal patterns. That means changes involving circulation, breathing and the skin's electrical activity. The examiner uses those recordings to form an opinion about honesty. The distinction is built into the definition itself: one part describes what the equipment records, while another describes the purpose for which the recordings are used. A physiological trace and a judgment about deception are separate stages.

An early patent makes that separation visible in mechanical terms. Leonarde Keeler filed his application for an arterial blood-pressure recorder on July 30, 1925; U.S. Patent 1,788,434 was issued on January 13, 1931. Its stated aim was to record the cardiac cycle together with slower pressure oscillations. Keeler described tracing these fluctuations on moving paper and recording respiration alongside pressure curves. He also identified police investigations as an intended use. The patent documents an apparatus and its proposed applications, rather than a controlled measurement of how accurately it distinguished truthful people from deceptive ones.

The National Research Council examined that further step in its review of polygraph testing, The Polygraph and Lie Detection, published in 2003. Its opening chapter distinguishes the instrument from the test and the broader examination. Sensors transmit physiological data to analog or digital recording equipment. Interpretation then compares reactions to material relevant to the investigation with reactions to comparison material. The chart may be scored by an examiner, another evaluator or a computer. Each method still requires a rule connecting differences in bodily response to a conclusion about truthfulness.

The measured changes are not unique to lying. The Council's executive summary identifies fear of being judged deceptive and anxiety about the examination as influences on the same responses. A truthful person can experience those states. The committee also found little understanding of how strongly such factors affect outcomes in practical settings. Consequently, more sensitive equipment does not by itself resolve the central difficulty in polygraph testing: the recorded response can have more than one explanation.

The review nevertheless did not equate every result with random guessing. It used 57 studies to quantify accuracy, all concerning specific incidents, usually simulated crimes. For populations like those studied, and participants untrained in countermeasures, the committee concluded that specific-incident tests distinguished lying from truth telling above chance but below perfection. It also judged the studies' quality inadequate and their accuracy estimates likely too high for field use. Laboratory consequences were generally unlike the stakes of an actual investigation, while field research often lacked an independent way to establish what had happened.

Security screening added a different problem. In a known-incident investigation, the disputed act can be defined narrowly. Screening employees requires broader inquiries about conduct investigators may not know occurred. The committee found no clear, stable agreement across agencies about standards for judging some screening responses. It warned against carrying accuracy estimates from specific-incident research directly into that setting. Polygraph testing for an unexplained workplace offense and screening a workforce for hidden security threats therefore could not be treated as interchangeable applications.

The report illustrated the numerical difficulty with a hypothetical workforce of 10,000 containing 10 spies. Assuming a test more accurate than available research justified, a setting that detected eight spies also flagged 1,598 employees who were not spies. A setting with far fewer false alarms flagged 39 innocent employees but identified only two spies. Those were model results, not observations from a laboratory workforce. They showed why a seemingly impressive overall classification rate could coexist with missed threats or large numbers of innocent people requiring investigation.

A court dispute supplied another distinction between the examination and its legal use. In United States v. Scheffer, decided in 1998, an Air Force examiner reported no deception when airman Edward Scheffer denied drug use. A urine test detected methamphetamine. Scheffer maintained that he had not knowingly taken the drug and wanted the polygraph result admitted at his court-martial. Military Rule of Evidence 707 barred it. The Supreme Court upheld that exclusion against his constitutional challenge, citing the continuing disagreement over reliability. Its ruling concerned the military evidentiary rule; it did not establish one admissibility rule for every court.

The Council also separated detection accuracy from other possible effects of a testing program. Field reports and indirect evidence suggested examinations could deter misconduct or elicit admissions from people who believed the device would expose them. Those effects depended on beliefs about the procedure and could exist even when detection accuracy was limited. They could also encourage misplaced confidence in security. An admission obtained during an examination and an examiner's interpretation of a chart are different forms of information, requiring different assessments.

The Question

The lie detector has been sending people to prison and clearing spies for a century. What does a polygraph actually measure?

Show the answer

The Answer

Physiological responses. The National Academies found that specific-incident tests could distinguish lying from truth telling above chance, though well below perfection. It also found serious scientific limits, especially for security screening. Anxiety about being tested can move the needle. A technique that's susceptible to error is a shaky bet with electrodes attached.

Send this question to a friend

Read next

By clicking “Send me a stumper” you agree to Crime Stumper’s Terms of Use and Privacy Policy.