Design and Analysis of Time-Series Experiments (PB)

Hailed as a landmark in the development of experimental methods when itappeared in 1975 Design and Analysis of Time-Series Experiments is availableagain after several years of being out of print.Gene V Glass Victor L. Willson and John M. Gottman have carried forward thedesign and analysis of perhaps the most powerful and useful quasi-experimentaldesign identified by their mentors in the classic Campbell & Stanley text Experimental and Quasi-experimentalDesign for Research (1966). In an era when governments seek to resolve questions of experimental validity byfiat and the label Scientifically Based Research is appropriated for only certain privileged experimentaldesigns nothing could be more appropriate than to bring back the classic text that challenges doctrinaireopinions of proper causal analysis.Glass Willson & Gottman introduce and illustrate an armamentarium of interrupted time-series experimentaldesigns that offer some of the most powerful tools for discovering andvalidating causal relationships in social and education policy analysis.Drawing on the ground-breaking statistical analytic tools of Box & Jenkinsthe authors extend the comprehensive autoregressive-integrated-movingaverages(ARIMA) model to accommodate significance testing andestimation of the effects of interventions into real world time-series. Designsand full statistical analyses are richly illustrated with actual examples fromeducation behavioral psychology and sociology.