Crime incident data from the 27 public housing developments were collected over a 24-month period ending in March 1975. Analysis of variance was used to test for relationships between housing types and crime and between demographic groupings and crime. The research design provided for consideration of temporal fluctuations. Spatial analysis of crime incidents was conducted through the use of maps and contingency coefficients. The research design was essentially pre-experimental and exploratory, with few specific hypotheses tested. The study reports that (1) demographic groups were stronger than architectural types in their association with crime levels; (2) offenders tend to move from a higher income census tract to commit a crime in a lower income census tract; and (3) four significant canonical factors emerged when the criterion and predictor sets were analyzed simultaneously. Further research is needed on the composition of the demographic groupings used and on alternative methods for creating such groups before confidence can be placed in the validity of statistically significant relationships found in using this variable. The conclusions related to offender travel require further analysis. Without the two-way tables from which the chi-squares were computed, support for the conclusions is difficult to find. The high contingency coefficient might result from large diagonal values or the large number of empty cells. The use of canonical correlation was improper because of the smallness of the sample size. Overall, little confidence can be given to study results, except for the one-way analysis of variance reported. The problem researched is complex and requires a large number of variables. See NCJ 40752 for the original report.