The utility of electrodiagnostic testing in the evaluation of carpal tunnel syndrome (CTS) has been questioned. We studied patients who met the clinical criteria for CTS and compared patients who had normal nerve conduction studies (NCS) with patients who had abnormal NCS. We found that 25% of the CTS patients without confounding neurologic disorders had normal NCS with median palmar nerve stimulation. Patients with abnormal NCS were older and heavier and had more clinical features of CTS. NCS results could not be predicted accurately from clinical features by use of logistic regression models. This was especially true in clinically borderline cases. NCS did not predict the outcome of conservative management. We concluded that NCS provide independent information in the evaluation of suspected CTS, especially when fewer clinical criteria are present, but that NCS are not helpful in predicting the outcome of nonsurgical management.