The authors first explain the concepts of validity, reliability, and variance in the field of polygraphy. This is followed by an explanation of "levels of rules." The authors note that the problem of diagnosing deception begins with the selection of polygraph features for scoring, followed by determining how numbers should be assigned to those features. Then, decision rules that use those numbers must be formulated so that the best decisions result. Scientists who have examined the physiological data found 10 polygraph tracing features reliable for manual scoring. These features are in the broad physiological categories of respiration, which includes suppression, increase in cycle time, change in the inhalation/exhalation ratio, and baseline rise; electrodermal, which includes amplitude of phasic response, duration of response, and complexity of response; cardiovascular, which includes baseline amplitude increase and duration of response; and vasomotor, which involves a reduction of pulse wave amplitude. The authors discuss number assignment for these features and decision rules once numbers have been effectively assigned to the individual response comparisons. Regarding decisions rules, the authors identify three factors for setting cutting scores: the proportion of false positive errors, false negative errors, and inconclusive outcomes that the consumer can tolerate. The possibility of using available automated algorithms in deception diagnosis is also mentioned. 32 references