Twenty-six psychoneurotic inpatients received oxazepam (60-120 mg daily) or placebo for one to two weeks. The strongest evidence for the superiority of oxazepam over placebo was that a greater proportion of oxazepam-treated patients remained on the study the full two weeks than placebo-treated patients. Patients' self-evaluations on the MMPI also favored oxazepam. Behavioral ratings by clinicians and ward personnel on the BPRS, symptom checklist, and assessment of global disorder did not discriminate the active drug from placebo as well as self-ratings did. Most of the differences found favored oxazepam, although some failed to reach acceptable levels of significance. It is suggested that the relative similarity of neurotic individuals to normal individuals with respect to pathology (e.g. anxiety, tension, depression) and difficulties in life and that the subjectivity of experience and the absence of gross and easily demonstrable symptoms make evaluation of improvement difficult. Symptom measurement on behavioral rating devices is probably less effective in discriminating drug-placebo differences in neurotic syndromes than in psychotic syndromes.