Data were derived from three principal sources: (1) field observations in each neighborhood over 14 months in 1976-77, (2) telephone surveys of random samples of residents in each neighborhood conducted in the fall of 1977, and (3) official crime rates in the neighborhoods for 1976 for the crimes of rape, assault, robbery, and burglary. Interviews designed to provide street-level knowledge of neighborhood characteristics and problems were conducted with residents, officials, and community leaders in each neighborhood. In addition to the qualitative information from the field observations, the demographic characteristics of the neighborhoods were compared using data provided by respondents to the telephone survey. Results showed several anomalies. Citizens' perceptions of dangerous areas in their neighborhoods matched for the most part official records of crimes committed there. However, assessments of neighborhoods' specific crime problems and personal risks did not consistently correspond with official statistics. Consequently, it is contended that citizens' perceptions of crime are shaped not so much by the neighborhood conditions reflected in the crime statistics, but rather by the level of incivility in their communities. Indicators of incivility are abandoned buildings, vandalism, drug use, and loitering teenagers. A correspondence between levels of fear and concern about incivility is demonstrated. It is suggested that fear of crime is triggered by a broad range of neighborhood conditions; and that attempts to understand and control that fear should look beyond serious crime incidents as the sole source of the problem. Tables, figures, footnotes, and 15 references are provided.