In the study context, 'architectural design' encompassed high-rise buildings (nine or more floors) and walkups (up to five floors) with or without defensible space characteristics. The sites for each of the four categories of architectural design were selected to represent inhabitants of both high and low social stratification. The independent variable was architectural design and the dependent variable was crime. The study instrument included controlled variables such as inhabitants' interaction structure, resident dissatisfaction, perceived police presence, fear of victimization, readiness to defend, and avoidance behavior. Interviews were conducted with 715 residents of the sample areas. Comparison of individual sample areas revealed no influence of architectural design on the crime rate. The same result appeared when crime was related to the social strata. The attempt to differentiate the investigated areas according to the controlled variables by discriminant analysis yielded canonical correlation coefficients of .29 in Munich and .30 in Regensburg for those areas that differed most markedly in architectural design, i.e., high-rises without defensible space versus walkups with defensible space. Thus, the values of the discriminant function are so low that the investigated areas are not even describable through the controlled variables. It is surmised that the negative results do not exclude the possibility of a direct influence of architectural structure upon the variables of fear of crime, readiness to defend, and even others not included in this study, such as family interaction, childrearing styles dependent upon housing conditions, and other forms of socialization deficits. Tabular data, photographs, charts, and appendixes are provided. The bibliography includes over 80 references.