Time Series Methods in Applied Social Research

Abstract

Listen

Translate article

Time series methods represent an important research tool for applied social psychologists. Consider interrupted time series designs, in which one assesses the effect of some intervention on an outcome measured repeatedly over time. Interrupted time series designs have been employed to address such varied questions as: the effect of drunk driving laws (Hilton, 1984; Shore & Maguin, 1988; West, Hepworth, McCall, & Reich, 1989), mandatory seat belt laws (Wagenaar, Maybee, & Sullivan, 1988), and an increase in the legal minimum drinking age (Wagenaar, 1982) on traffic fatalities; the effect of incentive payment (Wagner, Rubin, & Callahan, 1988) and of merit pay (Pearce, Stevenson, & Perry, 1985) on productivity; the effect of publishing offenders’ names in a newspaper on the frequency of shoplifting and drunk driving (Ross & White, 1987); the effect of mandatory sentencing on firearms violence (Loftin, Heumann, & McDowall, 1983); the effect of TV public service announcements on inquiries to social service agencies (McAbee & Cafferty, 1982); and even the effect of introducing meditation groups on the crime in an area (Dillbeck et al., 1987). In addition, related time series methods, which focus not on the effects of some abrupt intervention but rather on the covariation between two ongoing time series, have addressed the relationship between such variables as: air pollution and psychological state (Bullinger, 1989); air pollution, weather, and violent crime (Rotton & Frey, 1985); economic conditions and alcohol-related traffic fatalities (Wagenaar & Streff, 1989); alcohol use and suicide rates (Norstrom, 1988); expenditures on cigarette advertising and cigarette consumption (Chetwynd, Coope, Brodie, & Wells, 1988); and historical economic data and lynchings (Hepworth & West, 1988). As one indication of the potential importance of time series methods, consider that Cook and Campbell (1979) dedicated two chapters of their book on quasi-experimentation to the topic of interrupted time series, and part of another chapter to methods for assessing the relationship between two ongoing time series.