The first and most obvious difficulty in conducting gang research is that most researchers are highly educated, middle-class persons, whereas most gang members are not. It takes a highly skilled ethnographer to overcome the initial hostility that is often inherent to interactions with gang members. In addition, many researchers have noted that gang members are notoriously unreliable as informants. Therefore, the collection of valid data through fieldwork is possible only after an extended period of contact during which trust is established. A more subtle problem is the degree to which the researcher's presence has a significant effect on the dynamics observed. Survey-based study designs have many of the same problems concerning trust and hostility. Survey research also entails two further problems: the sampling frame itself and the identification of the participants as gang members. The final data collection technique is based on information from police agencies. This approach raises the problems of identifying gang membership and classifying illegal events as gang-related. Another problem in gang research is inconsistency among definitions; this problem makes it difficult to make reliable comparisons across jurisdiction. Nevertheless, certain patterns that have emerged in empirical studies provides at least a minimal degree of confidence in some of the findings. Notes and 37 references