Constructing a Quasi-Controlled Environment

Abstract

Three alternative techniques are utilized for evaluation of a law enforcement experiment. (I) forecasting holding the experimental variables constant to compare with the observed time series, (2) interrupted time series analysis, and (3) contrasting observed experimental results with a controlled environment. Each of the first two have theoretical or practical weaknesses noted in the literature of econometrics and sociometries. The third approach is to construct, from a set of 72 alternatives, a control group appropriate to the circum stances of the test for the purpose of evaluation. A technique is presented for selecting the best quasi-control group. This technique should find suitable application in situations in which the first two alternatives would be inappropriate. Test results do not differ sig nificantly among the three alternatives.