A polygraph test works by recording involuntary body signals — heart rate, blood pressure, breathing, and skin sweat — while an examiner asks a scripted series of yes-or-no questions, then interpreting the pattern of stress responses to estimate whether you were truthful. The machine does not detect lies. It detects physiological arousal, and arousal can come from fear, anger, confusion, or the simple pressure of being tested. That gap between what the device measures and what people assume it measures shapes everything about how polygraph results should be read.
What the Machine Records
Three sets of sensors feed the polygraph. A blood pressure cuff on your upper arm tracks pulse and blood pressure. Two rubber tubes called pneumographs stretch across your chest and abdomen to record the depth and rhythm of your breathing. Small metal plates on your fingertips measure electrodermal activity, the slight rise in skin conductivity that happens when you sweat, even imperceptibly.
All four channels record at once, onto a moving chart in older machines or a computer display in modern ones. The examiner looks for spikes or shifts that line up with particular questions. The theory behind the whole setup is that lying creates an emotional load your autonomic nervous system reveals through these signals, whether you want it to or not. The theory is intuitive. It is also, as research has repeatedly shown, shaky.
The Three Phases of an Exam
A polygraph session usually runs 90 minutes to two hours and moves through three distinct stages.
Pre-Test Interview
Before any sensors go on, the examiner spends a substantial chunk of time talking with you. They explain the procedure, gather background, and walk through every question that will be asked. Nothing on the test is supposed to be a surprise. The examiner also uses this phase to observe your baseline demeanor and to construct the comparison questions that will anchor the scoring later.
In-Test Phase
With the sensors attached, the examiner asks the prepared questions in a set sequence while the polygraph records continuously. You answer each one with a simple “yes” or “no.” The same question set is typically run at least three times with brief pauses between rounds, so the examiner can separate a genuine recurring pattern from a stray spike caused by an itch or a noise outside the room.
Post-Test Analysis
When the questioning ends, the examiner scores the recorded data and reaches one of three conclusions: no deception indicated, deception indicated, or inconclusive. Many examiners then discuss the results with you. In law enforcement settings, that conversation often becomes an informal interrogation. If the examiner says the chart shows deception, they may press for an explanation, hoping to produce an admission. In practice, the post-test confrontation is frequently the part of the exam that generates usable information for investigators, not the chart itself.
Two Very Different Questioning Techniques
“Polygraph” is a single word for two methods that work on different logic.
Comparison Question Test
The Comparison Question Test (CQT) dominates criminal investigations and employment screening. It mixes three question types: relevant questions about the specific issue under investigation, irrelevant questions meant to establish a neutral baseline, and comparison questions that probe vaguely about past misdeeds unrelated to the case.
The comparison questions do the heavy lifting. The examiner designs them to be uncomfortable to answer — things like “Have you ever lied to someone who trusted you?” The scoring theory assumes a truthful person will react more strongly to those broad questions, because they are harder to answer with full confidence, while a deceptive person will react more strongly to the relevant questions that hit their actual secret.
The weakness is built in. An innocent person terrified of being falsely accused may react powerfully to the relevant questions too. The CQT essentially asks the examiner to tell the fear of being caught apart from the fear of being wrongly accused, using the same physiological signals.
Concealed Information Test
The Concealed Information Test (CIT), sometimes called the Guilty Knowledge Test, works differently. Instead of accusatory questions, it tests whether you recognize details only the perpetrator would know. If investigators know a stolen item was hidden in a blue bag, the examiner might ask whether it was in a red bag, a green bag, a blue bag, or a black bag. You answer “no” to all four. Someone who committed the crime typically shows a stronger reaction to the correct answer.
The CIT rests on recognition memory, which is better understood than the stress-based assumptions behind the CQT, and the FBI Law Enforcement Bulletin has noted its theoretical foundation may be strong enough to meet federal standards for admitting scientific evidence. The catch is that it only works when investigators hold crime-specific details that have never been shared with the suspect or the public. That narrows its use, and the weaker-grounded CQT remains the default in real-world testing.
How Accurate Are the Results
Accuracy figures vary dramatically depending on the source. The American Polygraph Association has cited research showing accuracy between 80 and 98 percent. Independent reviews are far less flattering.
A comprehensive analysis by the congressional Office of Technology Assessment found that in field studies of the comparison question technique, correct identification of guilty subjects averaged 86.3 percent, while correct identification of innocent subjects averaged only 76 percent. The false positive rate — innocent people labeled as deceptive — averaged 19.1 percent. Roughly one in five truthful people who sit for the test will be told they failed. In laboratory studies, the numbers dropped further: 63.7 percent for correct guilty detections and 57.9 percent for correct innocent detections, barely above chance.
The American Psychological Association’s position, drawn from the weight of research, is that there is little evidence polygraph tests can accurately detect lies. One prominent researcher summarized the underlying problem bluntly: there is no unique physiological reaction to deception, which means the premise of the technology rests on an unproven assumption.
Can the Test Be Beaten
Countermeasures exist, and their existence is itself a reliability problem. The commonly described techniques involve artificially inflating your response to the comparison questions so they overshadow your response to the relevant ones. Biting your tongue, pressing your toes into the floor, or doing mental arithmetic during comparison questions can all generate a spike that makes the comparison baseline look hotter than the questions actually under investigation.
How reliably these techniques work in practice is debated, and some research suggests trained subjects can defeat the CQT at rates well above chance. What is not seriously debated is that a test whose results can be manipulated by a motivated subject carries a built-in vulnerability, even when the stakes attached to it are serious.
Whether Polygraph Results Are Allowed in Court
Federal courts have historically excluded polygraph evidence, and that position has not meaningfully shifted. In 1998, the Supreme Court upheld a military rule that categorically barred polygraph evidence from court-martial proceedings, ruling the exclusion did not violate a defendant’s right to present a defense.
In civilian federal courts, admissibility runs through the standard set in Daubert v. Merrell Dow Pharmaceuticals (1993), which asks trial judges to evaluate whether expert testimony rests on sound scientific methodology, including testing, peer review, known error rates, and general acceptance. Polygraphs struggle on every one of those criteria. The Department of Justice’s guidance to federal prosecutors notes that for many years the courts of appeals have upheld the exclusion of polygraph evidence. A few circuits have moved from a flat ban to case-by-case consideration since Daubert, but even those rarely admit the results in practice. State courts vary, though most follow the same restrictive approach.
Where the Test Still Gets Used
Scientific skepticism and courtroom exclusion have not retired the polygraph. Two settings keep it in heavy rotation.
Federal intelligence agencies use polygraphs routinely for security clearances. The National Reconnaissance Office, for example, requires every employee to complete and pass a counterintelligence polygraph regardless of position, and other intelligence community agencies impose similar requirements. For these agencies, the polygraph functions less as a scientific instrument than as a structured interrogation. The pre-test and post-test conversations sometimes produce admissions that a background investigation alone would miss, and that is the operational value.
Many state and local law enforcement agencies also polygraph police applicants during hiring, usually framed as a check on the accuracy of information in the applicant’s background questionnaire. Failing the test often disqualifies the candidate, even though the same results would be inadmissible if that candidate were later a defendant in court.
Your Rights If a Private Employer Asks
If you work in the private sector, the Employee Polygraph Protection Act (EPPA) makes it illegal for most employers to require, request, or even suggest that an employee or applicant take a lie detector test. Employers also cannot fire, discipline, or refuse to hire you for declining one.
The EPPA applies to private employers engaged in interstate commerce. Federal, state, and local government employers are completely exempt, which is why a city police department can polygraph every applicant and a private company generally cannot.
The law carves out narrow exceptions. An employer can request a polygraph from an employee during an ongoing investigation of a specific incident that caused economic loss, when the employee had access to the property involved and the employer has a reasonable, documented suspicion of involvement — and only after providing a written statement that identifies the loss and the basis for suspicion. Security firms providing armored car, security guard, or alarm services can polygraph prospective employees who would have access to protected facilities. Pharmaceutical companies authorized to handle controlled substances can polygraph prospective employees who would have direct access to those substances, and current employees in investigations involving them. Even where an exemption applies, strict procedural rules govern how the test is conducted and how results may be used.
Violations carry civil penalties up to $26,262 per violation under current federal regulations, and affected employees can sue within three years for reinstatement, back pay, and attorney’s fees.
What a Result Actually Tells You
The polygraph sits in an unusual place. Courts reject it. Scientists distrust it. Federal agencies and many employers use it anyway, largely because the examination process itself — not the machine — sometimes produces useful admissions. If you are facing a polygraph for a job, a clearance, or a legal matter, the thing to hold on to is what the device is doing: measuring your nervous system’s reaction to stress. It is not reading your mind. A failed polygraph is not proof of dishonesty, and a passed one is not proof of truthfulness.