Concurrent use of two direct observation instruments (momentary time sample of task engagement and event recording of discrete student responses) were evaluated in six middle school resource rooms using either of two reading programs. Findings tended to be program specific with differences lost or diluted when data were combined across programs. (Author/DB)