A polygraph, often incorrectly referred to as a lie detector test, is a pseudoscientific device or procedure that measures and records several physiological indicators such as blood pressure, pulse, respiration, and skin conductivity while a person is asked and answers a series of questions. The belief underpinning the use of the polygraph is that deceptive answers will produce physiological responses that can be differentiated from those associated with non-deceptive answers; however, there are no specific physiological reactions associated with lying, making it difficult to identify factors that separate those who are lying from those who are telling the truth.

In some countries, polygraphs are used as an interrogation tool with criminal suspects or candidates for sensitive public or private sector employment. Some United States law enforcement and federal government agencies, as well as many police departments, use polygraph examinations to interrogate suspects and screen new employees. Within the US federal government, a polygraph examination is also referred to as a psychophysiological detection of deception examination.

Assessments of polygraphy by scientific and government bodies generally suggest that polygraphs are highly inaccurate, may easily be defeated by countermeasures, and are an imperfect or invalid means of assessing truthfulness. A comprehensive 2003 review by the National Academy of Sciences of existing research concluded that there was "little basis for the expectation that a polygraph test could have extremely high accuracy", while the American Psychological Association has stated that "most psychologists agree that there is little evidence that polygraph tests can accurately detect lies." For this reason, the use of polygraphs to detect lies is considered a form of either pseudoscience or junk science.

The examiner typically begins polygraph test sessions with a pre-test interview to gain some preliminary information which will later be used to develop diagnostic questions. Then the tester will explain how the polygraph is supposed to work, emphasizing that it can detect lies and that it is important to answer truthfully. Then a "stim test" is often conducted: the subject is asked to deliberately lie and then the tester reports that he was able to detect this lie. Guilty subjects are likely to become more anxious when they are reminded of the test's validity. However, there are risks of innocent subjects being equally or more anxious than the guilty. Then the actual test starts. Some of the questions asked are "irrelevant" ("Is your name Fred?"), others are "diagnostic" questions, and the remainder are the "relevant questions" that the tester is really interested in. The different types of questions alternate. The test is passed if the physiological responses to the diagnostic questions are larger than those during the relevant questions.

Criticisms have been given regarding the validity of the administration of the Control Question Technique. The CQT may be vulnerable to being conducted in an interrogation-like fashion. This kind of interrogation style would elicit a nervous response from innocent and guilty suspects alike. There are several other ways of administering the questions.

An alternative is the Guilty Knowledge Test (GKT), or the Concealed Information Test, which is used in Japan. The administration of this test is given to prevent potential errors that may arise from the questioning style. The test is usually conducted by a tester with no knowledge of the crime or circumstances in question. The administrator tests the participant on their knowledge of the crime that would not be known to an innocent person. For example: "Was the crime committed with a .45 or a 9 mm?" The questions are in multiple choice and the participant is rated on how they react to the correct answer. If they react strongly to the guilty information, then proponents of the test believe that it is likely that they know facts relevant to the case. This administration is considered more valid by supporters of the test because it contains many safeguards to avoid the risk of the administrator influencing the results.

Assessments of polygraphy by scientific and government bodies generally suggest that polygraphs are inaccurate, may be defeated by countermeasures, and are an imperfect or invalid means of assessing truthfulness. Despite claims that polygraph tests are between 80% and 90% accurate by advocates, the National Research Council has found no evidence of effectiveness. In particular, studies have indicated that the relevant–irrelevant questioning technique is not ideal, as many innocent subjects exert a heightened physiological reaction to the crime-relevant questions. The American Psychological Association states "Most psychologists agree that there is little evidence that polygraph tests can accurately detect lies."

In 2002, a review by the National Research Council found that, in populations "untrained in countermeasures, specific-incident polygraph tests can discriminate lying from truth telling at rates well above chance, though well below perfection". The review also warns against generalization from these findings to justify the use of polygraphs—"polygraph accuracy for screening purposes is almost certainly lower than what can be achieved by specific-incident polygraph tests in the field"—and notes some examinees may be able to take countermeasures to produce deceptive results.