Sampling of speech pathology treatment activities: an evaluation of momentary and interval sampling procedures

Videotaped samples of aphasia treatment sessions were coded, using the Clinical Interaction Analysis System (CIAS), a 39-category system for recording the events that occur in clinician-patient interactions during aphasia treatment sessions. These coded records were then sampled according to various schedules and procedures and the fidelity with which each sampling schedule and procedure represented the content of the entire treatment record was evaluated. In addition, trained observers coded videotaped samples of treatment, using the CIAS with a number of sampling schedules and procedures. The fidelity with which these observers' records represented the content of the treatment sessions sampled was then evaluated. The results of the analysis indicated that momentary sampling at intervals distributed throughout the session generates more accurate records of session content than single longer samples taken from the session, unless those single samples comprise a major part of the session, and that sampling representativeness remains high even when only one event in ten is sampled, if sampled events are uniformly distributed throughout the session.