A TOTAL OF 102 SUBJECTS WERE EXAMINED ON THE POLYGRAPH AT THE REQUEST OF POLICE, DEFENSE ATTORNEYS, OR PROSECUTING ATTORNEYS IN UTAH AND NEVADA. PANELISTS USED IN THE STUDY EACH MADE DECISIONS ON ALL CASES IN THE SAMPLE, USING A 5-POSITION SCALE INDICATING THEIR DECISION AND THEIR CONFIDENCE IN IT. CASE INFORMATION PROVIDED TO THE EXAMINER PRIOR TO THE TEST SOMETIMES INCLUDED THE INITIAL POLICE REPORT OF THE INCIDENT BUT WAS MORE OFTEN LIMITED TO THAT PROVIDED BY A BRIEF TELEPHONE CONVERSATION WITH THE DEFENSE ATTORNEY. THE EFFECTS OF VARIOUS PERSONALITY AND SOCIOECONOMIC VARIABLES THAT MIGHT INFLUENCE AUTONOMIC RESPONSIVITY IN A REALISTIC DETECTION OF DECEPTION SITUATION WERE INVESTIGATED. FACTORS LIMITING THE GENERALIZABILITY OF THE STUDY ARE IDENTIFIED AS THE USE OF ONLY ONE POLYGRAPH EXAMINER, THE INADEQUACY OF THE CRITERIA AGAINST WHICH THE EXAMINER'S DECISIONS WERE COMPARED, AND THE SMALL NUMBER OF SUSPECTS DIAGNOSED AS TRUTHFUL BY THE POLYGRAPH AND INNOCENT BY EXTERNAL CRITERIA. IT IS CONCLUDED THAT, DESPITE LIMITATIONS, THE OVERALL RESULTS ARE CONSISTENT WITH THE PROPOSITION THAT CAREFULLY ADMINISTERED CONTROL QUESTION POLYGRAPH EXAMINATIONS ARE HIGHLY ACCURATE IN ASSESSING THE CREDIBILITY OF CRIMINAL SUSPECTS.