Comparison Question Test: Accuracy, Court Use, and Employer Limits

The comparison question test is the polygraph format used in almost every criminal investigation, security clearance interview, and workplace inquiry that involves a lie detector. It works by measuring involuntary physiological reactions to three categories of questions and comparing how strongly the examinee reacts to questions about the specific matter under investigation versus vague questions about past behavior. The theory is simple: a person lying about the incident will react more intensely to incident-specific questions, while a truthful person will react more intensely to the broad behavioral questions. A 2003 National Research Council review found the format performs better than chance but well short of reliable.¹

The Three Types of Questions

Every comparison question test is built around three categories, and the contrast between them is what the examiner interprets.

Irrelevant questions are neutral prompts with no emotional weight, like “Is today Wednesday?” or “Are you sitting in a chair?” They usually open the sequence because reactions to the very first question in any series are considered unreliable, and they reappear throughout the test to let the examinee’s baseline settle between meaningful questions.²

Relevant questions go directly at the issue. In a theft case, a relevant question might be “Did you take the missing deposit?” The examiner expects a deceptive person to find this threatening. A truthful person, the theory says, should not react as strongly because the question poses no real danger.²

Comparison questions are intentionally broad and cover past behavior most people feel at least slightly guilty about. In a burglary investigation, a comparison question might be “Have you ever stolen anything?” The idea is that a truthful examinee, unconcerned about the relevant questions, will fixate on the comparison questions and show a stronger reaction to them. A deceptive examinee’s attention stays locked on the relevant questions.²

Probable-Lie vs. Directed-Lie

Traditional comparison questions are probable-lie questions. The examiner phrases them so broadly that most people have probably done the thing, then steers the examinee toward denying it. The examinee is never explicitly told to lie; the denial is nearly inevitable.

A newer variant uses directed-lie comparison questions, where the examiner openly instructs the examinee to answer “no” to something they both know calls for a “yes.” For example: “During the first 20 years of your life, did you ever tell even one lie?” The examinee is told to deny it, producing a known, deliberate lie as the comparison baseline.² Supporters say it is more standardized because it removes guesswork about whether the examinee actually lied. Critics say it changes the psychological dynamic in ways that are not well understood.

What the Polygraph Measures

The instrument tracks several involuntary bodily responses at once. No one can consciously suppress all of them, which is why the test uses multiple channels rather than a single reading.

Pneumograph tubes wrap around the chest and abdomen and record breathing patterns. The examiner watches for changes in rate, depth, and regularity. A classic stress sign is a brief suppression of breathing followed by a deep recovery breath shortly after a question.³

Electrodermal sensors on the fingertips or palms measure skin conductance. Small increases in sweat gland activity change electrical conductivity across the skin. This channel tends to be the most reactive and often produces the clearest data for scoring.

A cardiovascular cuff, similar to a standard blood pressure cuff, tracks heart rate and relative blood pressure. The examiner looks for pulse spikes or shifts in blood volume that line up with specific questions.³

Modern computerized systems often add activity sensor pads under the seat, armrests, and feet. These catch subtle or obvious movements the examiner might miss, and their main purpose is detecting deliberate physical countermeasures like pressing a toe against the floor or clenching muscles to spike the readings on comparison questions.⁴

The Pre-Test Interview

The pre-test interview usually takes longer than the recording phase, and it shapes the outcome more than most examinees realize. The examiner gathers background information, discusses the circumstances of the investigation, and explains how the instrument works. Building rapport is part of it. So is positioning the examinee psychologically, because the examiner wants a truthful person to leave the pre-test more worried about the comparison questions than the relevant ones.

Every question that will appear on the recorded test is reviewed word by word beforehand. Nothing should be a surprise. A startled reaction to an unexpected question looks identical on the chart to a deceptive reaction, so the examiner and examinee agree on the exact phrasing, and the examinee confirms they can answer each one with a clear “yes” or “no.”

Health matters too. The Department of Justice identifies fatigue, obesity, heart disease, respiratory difficulties, and abnormal blood pressure as physical conditions that can distort readings. Mental health conditions that impair understanding of the questions, and undetected alcohol or drug use, present similar problems.⁵ Under federal regulations implementing the Employee Polygraph Protection Act, a test cannot proceed if a physician has provided written evidence that the examinee has a medical or psychological condition that could cause abnormal responses.⁶

How Scoring Produces a Result

After the recording phase, the examiner assigns numerical values to the differences between relevant and comparison reactions. The standard method uses a seven-position scale from +3 to −3, where zero means no distinguishable difference between the two.³ Each pair of relevant and comparison questions is scored across all three physiological channels for every chart collected.

Positive scores mean the comparison question produced the stronger reaction. Negative scores mean the relevant question did. The individual scores are totaled across all charts into a single cumulative number.

A cumulative score of +6 or higher gets a finding of No Deception Indicated. A score of −6 or lower produces Deception Indicated. Anything between the two is Inconclusive, meaning the data did not clearly point either way.³ Lower cutoffs like plus or minus three reduce the inconclusive rate but raise the error rate.

If a deep breath distorts the pneumograph tracing just before a question, that channel may be excluded from scoring for that question. Movements that distort more than two consecutive pulses on the cardiovascular channel can also render data unscorable.³

How Accurate Is It

The most comprehensive independent review of polygraph science came from the National Research Council in 2003, and its conclusions still frame the debate. After reviewing 57 studies, the council found polygraph testing detects deception at rates well above chance but well below perfection. The median accuracy index across laboratory and field studies was 0.86, on a scale where 0.50 is random guessing and 1.0 is flawless detection.¹

The range is where the problem lives. In some studies, when the false positive rate was held at roughly 10 percent, sensitivity for correctly identifying deceptive examinees ran anywhere from 43 to 100 percent. A truthful person’s experience of the test depends heavily on who administers it and under what conditions.¹

The council was especially skeptical about screening applications, where there is no specific incident and the examinee faces vague questions about broad categories of behavior. It concluded that the evidence did not support polygraph screening at an accuracy index of 0.90 or above, and that generalizing from specific-incident studies to screening was not justified. The underlying problem is that the psychological states associated with deception also appear in truthful people who are anxious, afraid, or unsettled by the testing situation. The council also noted that most of the studies it reviewed fell below the quality level typically required for funding by the National Science Foundation or the National Institutes of Health.⁷

Can the Results Be Used in Court

Rarely, and almost never in a way that decides a case. No Federal Rule of Evidence specifically addresses polygraph admissibility, so federal courts evaluate the evidence under general standards for expert testimony.⁸

The governing framework comes from the Supreme Court’s 1993 decision in Daubert v. Merrell Dow Pharmaceuticals, which requires trial judges to assess whether expert testimony rests on genuine scientific knowledge and would assist the jury. Judges consider whether the technique has been tested, subjected to peer review, has a known error rate, and has gained general acceptance in the scientific community.⁸ Polygraph evidence struggles on all four, particularly general acceptance, because the scientific community remains divided.

The Supreme Court addressed the question directly in United States v. Scheffer (1998), upholding Military Rule of Evidence 707, which flatly bars polygraph evidence from courts-martial. The Court held that excluding polygraph results does not violate a defendant’s right to present a defense, citing three concerns: the lack of scientific consensus about reliability, the risk that jurors would give excessive weight to an examiner’s opinion on the ultimate question of credibility, and the likelihood of time-consuming side disputes about the test itself rather than guilt or innocence.⁹

Military Rule of Evidence 707 is still in force. It makes polygraph results, an examiner’s opinion, and any reference to an offer to take, refusal to take, or taking of a polygraph completely inadmissible in courts-martial. Statements made during a polygraph session may still be admitted if they are independently admissible.¹⁰

State courts vary. Some exclude polygraph evidence entirely. Others allow it when both parties stipulate to admission before the test. A smaller number let judges admit results at their discretion under Daubert or older reliability standards. In practice, polygraph results are excluded far more often than admitted, and no jurisdiction treats them as dispositive.

Can a Private Employer Make You Take One

Almost never. The Employee Polygraph Protection Act, codified at 29 U.S.C. §§ 2001–2009, prohibits most private employers from requiring, requesting, or even suggesting that an employee or applicant take a polygraph test.¹¹ The ban covers pre-employment screening and current-employee testing, and it extends to using, accepting, or asking about results from a test taken elsewhere.

The Narrow Exceptions

Employers whose primary business is armored car services, security alarm installation or maintenance, or employing uniformed or plainclothes security personnel may polygraph prospective employees. Employers authorized to handle controlled substances listed on federal drug schedules may test employees in connection with drug-related investigations.¹¹

A separate exemption lets private employers test current employees during an ongoing investigation into workplace theft or other economic loss, but only if the employer has reasonable suspicion that the specific employee was involved and that employee had access to the property. The employer must give a written statement before the test explaining the specific incident and the basis for suspecting that employee.¹²

Your Rights If a Test Is Allowed

Even where an exemption applies, the regulations impose strict protections. You can stop the test at any time. No employer may fire, discipline, or deny a promotion to someone solely because they refused. Any adverse employment action tied to a refusal requires additional supporting evidence.⁶

Employers must give written notice at least 48 hours before the test, excluding weekends and holidays, in a language the examinee understands. The notice must explain the instrument, the right to stop the test, the right to consult an attorney before each phase, the right to review all questions in writing beforehand, and the limits on adverse action.⁶ Certain subjects are off limits entirely: the examiner may not ask about religious beliefs, racial matters, political beliefs, sexual behavior, or labor union activities.

An employer who violates the Act faces a civil penalty of up to $26,262 per violation, adjusted periodically for inflation from the original statutory cap of $10,000.¹³ One significant boundary: the Act does not apply to federal, state, or local government employers, so public-sector agencies can require polygraph testing without these restrictions.

Security Clearance and Federal Use

Federal law enforcement and intelligence agencies are the heaviest users of polygraph testing. Within the Intelligence Community, examinations are governed by Intelligence Community Policy Guidance 704.6, which sets three tiers based on the level of access involved.¹⁴

Counterintelligence scope covers espionage, sabotage, terrorism, unauthorized disclosure of classified information, unreported foreign contacts, and deliberate damage to government information systems.
Expanded scope, also called full scope, covers everything in the counterintelligence scope plus criminal conduct, drug involvement, and falsification of security forms.
Specific-issue examinations target a single concern, such as a particular allegation of unauthorized disclosure, and may be used alongside either of the other two.

Counterintelligence and expanded-scope examinations can be given during initial vetting and repeated at periodic or irregular intervals as part of reinvestigations or continuous evaluation. Agencies like the CIA, NSA, and FBI each set their own policies for which positions require which tier, but expanded scope is standard for the most sensitive compartmented information.

Post-Conviction Sex Offender Monitoring

One other place the comparison question test shows up is probation and parole. Programs in many states use polygraph testing to monitor convicted sex offenders after release. The American Polygraph Association’s model policy for Post-Conviction Sex Offender Testing treats the polygraph as a decision-support tool meant to help supervision professionals assess risk, not as a standalone determination of compliance.¹⁵

Testing in this context falls into several categories: examinations focused on the original offense when the offender denies key allegations, sexual history examinations probing unreported past behavior, and periodic maintenance examinations conducted roughly every 6 to 12 months to check compliance with supervision conditions. Examiners must use a recognized comparison question technique with published evidence of validity.¹⁵

1
The National Academies Press. The Polygraph and Lie Detection – Chapter 5: The Accuracy of Polygraph Tests
2
The National Academies Press. The Polygraph and Lie Detection – Appendix A: Polygraph Questioning and Techniques
3
American Polygraph Association. Polygraph, 1999, 28(1)
4
Lafayette Instrument. Activity Sensor Pads User’s Manual
5
U.S. Department of Justice. Criminal Resource Manual 261 – Polygraphs Examination Variables
6
eCFR. 29 CFR Part 801 – Application of the Employee Polygraph Protection Act of 1988
7
The National Academies Press. The Polygraph and Lie Detection – Executive Summary
8
U.S. Department of Justice. Criminal Resource Manual 262 – Polygraphs Introduction at Trial
9
Justia Law. United States v. Scheffer, 523 U.S. 303 (1998)
10
Joint Service Committee on Military Justice. Military Rules of Evidence
11
Office of the Law Revision Counsel. 29 U.S.C. Chapter 22 – Employee Polygraph Protection
12
U.S. Department of Labor. Fact Sheet 36 – Employee Polygraph Protection Act of 1988
13
U.S. Department of Labor. Civil Money Penalty Inflation Adjustments
14
Office of the Director of National Intelligence. Intelligence Community Policy Guidance 704.6 – Polygraph Program
15
American Polygraph Association. Model Policy for Post-Conviction Sex Offender Testing