How a Polygraph Test Actually Works — From the Examiner's Chair

Almost everything the public believes about polygraph comes from television, and television gets one thing badly wrong: it puts the drama in the machine. In reality the instrument is the least interesting part of the process. What determines whether an examination is worth anything is the hour of conversation that happens before a single sensor is attached.

What Is Actually Being Measured

A polygraph records three channels of physiological data. Pneumograph tubes across the chest and abdomen record respiration. An electrodermal sensor on the fingers records changes in skin conductivity driven by sweat gland activity. A cardiovascular cuff records relative blood pressure and pulse. None of these is a lie. They are involuntary responses associated with psychological arousal, and arousal has many causes.

The examiner's job is to construct a set of questions where one particular cause — concern about deception on a specific issue — is the only plausible explanation for a consistent pattern of response. That is a question-design problem, not an equipment problem.

The Pre-Test Interview

In a properly run examination the pre-test phase takes longer than the test. Every question is reviewed with the examinee word for word. There are no surprise questions, no ambushes, and no ambiguity about what a term means. If the relevant question asks whether someone took money from a register, both parties must agree on what "took" means before the charts are ever collected.

This phase also establishes whether the examinee is a suitable subject at all. Acute illness, intoxication, severe sleep deprivation, certain medications and certain psychiatric conditions can render results uninterpretable. A competent examiner will decline to test rather than produce a chart that cannot be defended.

Single Issue Beats Multiple Issue, Every Time

The most common error in commercially available polygraph is trying to answer four questions in one sitting. Accuracy is highest when a test addresses one clearly defined incident with one clearly defined behaviour. "Did you take the missing $4,000 from the safe on 12 March" is testable. "Have you ever stolen from an employer" is a conversation, not a test.

What a Result Means

Charts are scored numerically against the physiological reactions to comparison questions, producing a call of "no deception indicated," "deception indicated," or "inconclusive." Inconclusive is a legitimate outcome and an honest examiner reports it rather than guessing. Anyone promising a definitive answer on every subject is promising something the methodology does not deliver.

The Honest Summary

A polygraph is a well-validated investigative tool with real limits. It is very good at resolving a narrow factual dispute with a cooperative subject. It is poor at general fishing expeditions. It does not read minds and it is not infallible. Used correctly, and by someone who will tell you when the answer is unclear, it resolves questions that would otherwise sit unanswered for years.

About the author

Dan Ribacoff is a licensed private investigator and the founder of International Investigative Group, a certified polygraph examiner, certified firearms instructor, GIA Graduate Gemologist and the author of I, Spy: How to Be Your Own Private Investigator. He works from offices in New York, Boca Raton and London, and has appeared as a polygraph expert on American and British television.