Methods of time sampling: A reappraisal of momentary time sampling and partial interval recording.

The authors compared two ways to record behavior: momentary time sampling (MTS) and partial interval recording (PIR).

They asked which method gives truer duration data and which spots relative change best.

The tests were run in a lab with adults who followed scripts to create known behavior patterns.

MTS gave duration numbers that sat close to the real clock time.

PIR missed the mark on absolute time, but it caught when behavior went up or down.

Still, at high rates PIR shrank the size of the change it reported.

Gardenier et al. (2004) later saw the same PIR inflation in kids with autism who showed stereotypy. Their field data matched the 1986 lab warning that PIR overestimates duration.

Sisson et al. (1993) added a caution: if behavior splits into many short bursts, even MTS can mislead. The 1986 paper did not test fragmented patterns, so the later note sharpens the picture rather than clashes with it.

Iwata et al. (1990) moved MTS into real classrooms and found 15-second samples worked well except for very brief acts. Together these studies build a rule: use MTS for accurate time shares, but stay alert when behavior is fleeting or chopped into bits.

Pick MTS when your treatment goal states a strict percentage of time, like 'on-task for 80% of the period'. Pick PIR only when you need a quick alert that behavior rose or fell, not the exact minutes. If you must track fast, broken behavior, shrink the MTS interval or collect continuous data instead of trusting a slow sample.