SUMMARYTwenty-one inpatients on a psychogeriatric unit were assigned randomly to a haloperidol or oxazepam treatment group. Drug effects were assessed with three psychiatric rating scales and direct behavioral observation. The behavioral microanalysis employed two separate systems of observation: aberrant behaviors were assessed by recording the frequency of behaviors per hour, and the activities assessment (eg social interaction, sleeping) sampled each resident's behaviors during two 10-min intervals per day. Results showed no differences between drug groups on any of the assessments, except for engagement in activities from the activities assessment. A modest non-significant decrease in symptoms was noted on all psychiatric assessments. The behavioral observations showed very low pretreatment levels of aberrant behaviors. Non-significant post-treatment decreases were observed for disruptive vocalizations, paranoid verbalization, and non-compliance; rates of physical aggression and psychomotor agitation increased somewhat. The potential role of behavioral microanalysis in nursing homes is discussed in light of new HCFA guidelines. , 1988;Cohen-Mansfield and Billing, 1986;Zimmer et al., 1984). Psychotropics, and most commonly neuroleptics, are frequently prescribed for these problems, although several literature reviews have concluded that indications of therapeutic efficacy of neuroleptic medications are modest at best (Devanand et al., 1988;Salzman, 1987;Schneider et al., 1990;Sunderland and Silver, 1988). In a recent meta-analysis, Schneider and colleagues (1990) determined that fewer than one in five patients prescribed neuroleptics will display an improvement in behavioral symptoms.