The findings of sleep studies in schizophrenia have remained inconsistent in the literature as exemplified by the recent controversy regarding reduced rapid eye movements (REM) latency in these patients. These inconsistencies can partly be explained by major methodological shortcomings in studies evaluating sleep in schizophrenia. Lack of standardized scoring and diagnostic methods in the earlier studies and more recently, the effect of neuroleptic treatment or its withdrawal, have confounded the sleep results. Data are presented to illustrate the effect of the presence of tardive dyskinesia or active psychotic symptoms that further skew the sleep polygraphic measurements in these patients. Studies in drug-naive patients can circumvent some of these confounds but then these studies are weakened by sampling bias. Available data suggest that previous duration of neuroleptic treatment, duration of neuroleptic withdrawal, presence of tardive dyskinesia, and severity of psychotic symptoms should be considered when interpreting REM sleep measures in schizophrenic patients.