Analyzing Policy Impact: Selection of a Linear Trend Model

In: Policy studies journal: an international journal of public policy, Volume 8, Issue 6, p. 928-941

Abstract

Policy analysts & politicians have in recent years shown great interest in assessing both short- & long-term consequences of public policies. Recent time-trend studies have attempted to depict the time dimension of policy consequences through extensions of regression techniques. Three linear trend models that have been used to depict policy impact through time-series analyses are critically examined. In a quasi-experimental time-series design, the models are applied to time-series of per capita state expenditures for large cities, in 8 states, over a 20-year period. Differences in degrees of multicollinearity & autocorrelation inherent in the three models are discussed, & the model providing the most conservative coefficient estimates is identified. 5 Tables, 3 Models. Modified HA.