Computer-assisted behavioral observation data were collected from 66 nursing home residents while they received assistance during personal care routines from a certified nursing assistant (CNA). Data were collected both before and after a comprehensive behavior-management and communication-skills training program was delivered to the nursing staff. A total of 30 residents showed 6 or more episodes of disruptive vocalization or other forms of agitation during baseline observations, and timed-event sequential analysis methods were used to test whether certain CNA behaviors either elicited or prevented onsets of agitation in these 30 residents. Simple verbal prompts used by CNAs during personal care routines before staff training were found to elicit agitation onset (p=.03), whereas positive statements to the resident were found to reduce the likelihood of agitation onset (p=.02). In a previous analysis of the effects of staff training, we found that rates of resident agitation were significantly lower after training compared with baseline and that CNAs increased their rates of positive statements to residents. In the present study, timed-event sequential analyses indicated that the verbal prompts used by CNAs during personal care routines were no longer associated with an increased rate of resident agitation after staff training among the 20 residents who showed 6 or more episodes of agitation both before and after staff training. These findings suggest that the staff training program improved the quality of CNA verbal prompts. The sequential analysis of observational data provides an important method for studying interpersonal interactions, including those of nursing home residents and the nursing staff who care for them.