ected through the use of a variety of procedures. Included in these procedures were (1) surveys of (a) victimization, (b) atti tudes, (c) crime reporting rates, and (d) perceptions, (2) interviews and recorded data, and (3) abstracting departmental data.
Analysis of variance (ANOVA) procedures were applied to the collected data in the testing of the hypotheses. Statisti cal significance was established at the p<.05 level of prob ability for each hypothesis tested, as the acceptance/rejection criterion.
The basic concept of the methodology for the Kansas City Preventive Patrol Experiment was sound. The hypotheses were reasonable, the data collection effort was thorough and valid, and the quantitative procedures were consistent with standards for the social sciences.
The major problem with the methodology of the Experiment was its attempt to measure the effects of patrol variations as isolated phenomena. The level of criminal activity, the level of reporting of such activity, public perceptions of police activity, and so forth are the products of a variety of fac tors. The variation of just one of these factors patrol activity variation in the Kansas City Experiment, and the mea surement of the consequent effects implicitly (1) presumes that each of the other variables remained unchanged during the per iod of the experiment, and (2) ignores the potential effects which might result of other factors were manipulated along with the level and type of patrol activity. Thus, while the Experi ment measured the results that it set out to measure, the methodological defect just described creates serious questions about any efforts to either (1) draw conclusions from the find ings, or (2) in any way generalize the results beyond their narrow methodological base.
The results of the Kansas City Experiment found that patrol type and activity level had no statistically significant effects on "the level of crime, citizens' attit...