Four hundred and eleven workers from 4 different companies participated in a worksite screening program designed, in part, to estimate the prevalence of carpal tunnel syndrome (CTS). Each worker completed a discomfort survey and underwent limited nerve conduction testing of the median and ulnar sensory nerves in both wrists. The discomfort survey included a hand diagram which allowed subjects to shade in area(s) affected by numbness, burning, tingling, or pain. The discomfort survey also asked each worker to indicate whether she or he had experienced neuropathic symptoms (i.e., numbness, burning, tingling, or pain) in the wrist, hand or fingers of each hand, without regard to localization (i.e., median versus ulnar versus radial distribution), and also nocturnal occurrence of symptoms. Analyses involved comparing hand diagram scores and non localized wrist/hand/finger symptoms with electrodiagnostic test results. All configurations of hand diagram scores of the dominant hands had a statistically significant association with electrophysiologically determined median nerve dysfunction, but so did non localized symptom reports. The sensitivity, specificity, and positive predicted values of hand diagrams were poorer than those reported previously. While some test performance characteristics of hand diagrams were better than those for non localized distal extremity symptoms consistent with CTS, some were worse. Overall, our data suggest that hand diagrams are no better than using a questionnaire to determine if workers have experienced symptoms consistent with CTS in their wrists, hands or fingers without regard to localization. The choice of screening tool would depend on the goal of screening, in particular, whether it is more desirable to have slightly higher sensitivity or positive predictive value.