The bulk of Lykken's criticisms of polygraph validity rest on his claim that the accuracy of control-question polygraph tests is 84 percent for guilty suspects and 53 percent for innocent suspects. These figures are based on the results of three studies Lykken selected from the dozens of studies reported in the scientific literature. The Office of Technology Assessment (OTA) identified 10 field studies which met its criteria for validity, including the 3 studies selected by Lykken. Their findings indicated that polygraph test accuracy was 90 percent for guilty suspects and 80 percent for innocent suspects. The authors repeated the analysis using accuracies estimated from the OTA results and scientific laboratory studies. Assuming a base rate of guilt of 50 percent in the population tested, confidence ranged from 82 to 92 percent for deceptive outcomes and 88 to 97 percent for truthful outcomes. These accuracy rates are sufficient to make polygraph tests appropriate as evidence in investigations and court proceedings. In his criticism of the accuracy of urinalysis for drug consumption testing, Lykken did not acknowledge the procedure whereby positive tests from a large number of tests are subjected to a confirmatory test by independent procedures with greater precision, such that the false positive rate is reduced to near zero. 23 footnotes. For Lykken's arguments, see NCJ 105379.