The study presented competing confirmatory factor analytic models to test the relative fit of one- and two-factor models of citizens' perceptions of disorder and crime in their communities. The results show that the two-factor model is inappropriate because of a high correlation between perceptions of disorder and crime. The primary implication of this finding is that broken windows theory and order maintenance policing need to be reconsidered. Broken windows, or order maintenance policing, is premised on the idea that citizens recognize disorder as a problem apart from crime and that reducing disorder will cause a decline in fear. The results of the current study refuted the underlying logic of order maintenance policing by demonstrating that survey respondents did not, in fact, distinguish between disorder and crime. Broken windows theory, and policing strategies based on this perspective, predicts that citizens' perceptions of disorder in their communities cause fear and social withdrawal, which thereby opens the streets for serious predatory crime. A key assumption, therefore, is that the primary exogenous variable in the broken windows process (disorder) and its outcome variable (crime) are actually separate constructs. The empirical evidence regarding the discriminate validity of measures of disorder and crime, however, was mixed. This result also has implications both for the actual effectiveness of order maintenance policing in terms of crime reduction and for citizens' satisfaction with police agencies that focus more on disorder than on serious crime. Data for this study came from a 2003 mail survey conducted to gather information on eastern Washington residents’ fear of crime and perceptions of crime and disorder in their neighborhoods, using a random sample for each of 21 municipalities. Tables, references, appendices A-C