Observer error in the use of momentary time sampling and partial interval recording

Abstract

The accuracy of 60 first-year psychology degree students using either the method of momentary time sampling (MTS) or partial interval recording (PIR) was estimated by agreement measures in an independent groups study. The purpose of the study was to investigate possible method differences in a design that maximized their comparability. After training, observers were randomly allocated to MTS or PIR methods and, after further practice, recorded either one, two or three behaviours (namely, reading, writing, hand-clasping) from a constructed 10-minute videotape of human studying behaviour from which a criterion record had been derived. Results showed MTS introduced significantly less error into observers' records than PIR across all levels of complexity. However, there was no generally significant increase in error with increasing complexity and PIR showed relatively more accuracy in recording writing behaviour. Despite the apparent support offered to practitioners for the use of MTS in behavioural investigations, the experimenters qualify their findings and indicate the need for a more extensive series of experiments comparing these methods.