From a survey of 131 municipal police departments in 9 States, 79 responses were received. The survey was conducted in the late summer of 1991. Researchers examined how 15 independent variables influenced police perceptions of the gang problem, using hierarchical regression. City size and number of gangs described the background in which police-gang interactions occurred. Other variables described the largest gang's traits and the frequency of their criminal activities. The traits of the city's largest gang included members' ages, predominant race, out-of-town affiliations, longevity in the jurisdiction, number of active members, and level of organizational sophistication. Criminal involvements were operationalized in terms of level of drug use, reliance on drug sales, expressive property crime, instrumental property crime, violence, and propensity to carry guns and other weapons. Results show that gang size and level of organizational sophistication had a significant effect on police perceptions of the overall severity of the gang problem within their jurisdiction. Variables that described the frequency with which the gang committed various criminal acts had no significant impact when gang traits and related factors were controlled. The authors identify several areas that require additional examination before estimates of the severity of the gang problem can be accurately made. 1 table and 68 references