The brief, in using Larson's own model for determining police presence in experimental beats (i.e., visibility operationalized as mileage covered in an average 8-hour beat), shows that the levels of police presence in proactive and reactive beats were disparate. Larson's formula is used to support this position by deriving the total mileage covered in an average 8-hour shift in each of the experimental conditions. Findings show that the differences in experimental conditions were so extreme that substantively interesting and statistically significant results would have absorbed much of the random and systematic biases, if preventative patrol was causally related to the dependent measures. This was especially true for pair-wise analyses of reactive versus proactive conditions. The experimental conditions were different enough to expect significant results if variability in police presence did indeed affect the dependent measures. In his critique of the experiment, Larson did not report similar findings because he failed to correctly apply his own model. If social policy decisions are to rely on program evaluation, applied social scientists must be very secure that design, implementation and analyses are both reliable and valid. Three references are cited.