Respondents who reported gangs in 1996 were asked for specific demographic information: age, sex, and race/ethnicity (see Appendix A). All responses were in the form of percentages. Demographic categories were grouped as follows:

• Age: younger than 15, 15-17, 18-24, and older than 24.

• Sex: male and female.

• Race/Ethnicity: African-American, Hispanic, Caucasian, Asian, and "other."

Gang activity within each demographic category was analyzed using the following variables: area type, geographic region (as defined by Crime in the United States, 1996: Uniform Crime Reports, Federal Bureau of Investigation, 1997; see Appendix E), and population. In addition, each demographic category was grouped by jurisdictions that reported their gang problem began before 1990 ("older" gang jurisdictions) and those that reported their problem began between 1990 and 1996 ("newer" gang jurisdictions). This grouping allowed for the comparison of characteristics of gang members in older gang jurisdictions with those in newer gang jurisdictions.

In order for a response to be considered in the analysis, the sum of the percentages within each category was required to equal 100 percent. Although most agencies responded appropriately, a small percentage (8 percent for age, 1 percent for sex, and 9 percent for race/ethnicity) of the responses did not total 100 percent and were subsequently excluded from the analysis. The excluded data were examined to determine if there was some systematic problem that contributed to the miscalculations. It was concluded that human error was primarily responsible. Additionally, summary data were not included for categories in which there were fewer than 20 observations, unless otherwise noted.

An important limitation to interpreting the responses to the demographic questions and the question regarding the percentage of migrant gang members is that jurisdictions only provided average percentages in their responses. These percentages do not reflect differences in the size of the gangs across the reporting jurisdictions. To account for this important factor, the percentages were weighted by the total number of gang members reported in each jurisdiction (the total number of gangs was used for the question regarding multiethnic/multiracial gangs). Agency responses must have met the following criteria in order to be considered in the analysis: the sum of the percentages within each category must have totaled 100 percent, and the agency must have reported a total number of gang members or, for the question regarding multiethnic/multiracial gangs, a total number of gangs. These requirements decreased the number of agencies included in the analysis but increased the reliability of the translation of percentages into actual numbers of gangs or gang members.

It is important to be able to make comparisons about gang activity across jurisdictions utilizing the variables previously discussed (area type, geographic region, population, and year of onset). Therefore, weighted percentages were used only when estimating aggregate percentages for gangs or gang members. The remainder of the analyses were conducted by comparing averages, regardless of the number of gangs or gang members reported, for each jurisdiction. This procedure allowed differences across jurisdictions to be observed without being skewed by jurisdictions with a disproportionate number of gangs or gang members. When appropriate, tables are included in the following analysis to illustrate the differences between weighted and unweighted averages.