Time-Series Analysis of Behavioural Data: An Update

Abstract

Some recent developments in the use of interrupted time-series analysis (ITSA) are described with particular reference to the detection of effects with short data series such as those often encountered in applied behaviour analysis. The necessity to perform the sometimes troublesome model-identification procedure is questioned, and the likely incidence of Type 1 and 2 errors is discussed. Conclusions are drawn to suggest that ITSA may be safely applied to data that are typical of those collected in applied behaviour analysis.