This study had two purposes. The first purpose was to explore whether bullied students engage in specific protective behaviors that inhibit learning, put other students' safety at risk, or foster a negative school climate. Such behaviors were defined as avoidance (of places or individuals), truancy, physical fighting, and weapon carrying. The second purpose was to explore whether bullied students' behaviors varied by the type of bullying (direct, verbal, indirect, or cyber) endured.
This study relied upon a multistage, stratified sampling technique that is, for all intents and purposes, analogous to a simple random sample in terms of rare events forecasting. Prior to analysis, the NATIONAL CRIME VICTIMIZATION SURVEY: SCHOOL CRIME SUPPLEMENT, 2009 (ICPSR 28201) data was filtered to include only students who were ages 12 through 18, were enrolled in grades 6 through 12, were enrolled in school in the current school year, and were not homeschooled during this time. Students who did not fulfill one or more of these requirements for age (0 cases), grade (249 cases), enrollment (245 cases), and home schooling (203 cases) were deleted from the analysis. The final unweighted sample size was 4,326.
In the sample, each age category for students between 12 and 17 years old accounted for between 15 and 16 percent of the sample, and 18-year olds accounted for 9 percent. Males accounted for 51 percent of the sample and females 49 percent. Sixty-three percent of the sample was White, non-Hispanic; 12 percent was Black, non-Hispanic; 19 percent was of Hispanic origin; and 6 percent of the sample was of other, non-Hispanic race/ethnicities. Nine percent of the sample were 6th graders; 7th, 8th, and 9th graders each accounted for 15 percent of the sample; 16 percent were in 10th grade; and 11th and 12th graders each accounted for 14 percent. Household income ranged from less than $7,500 per year for 2 percent of the sample to over $75,000 per year for 20 percent of the sample. Twenty-seven percent of the sample lived in urban locations, 56 percent in suburban locations, and 17 percent in rural areas.
National Crime Victimization Survey (NCVS) household members aged 12 through 18.
NATIONAL CRIME VICTIMIZATION SURVEY: SCHOOL CRIME SUPPLEMENT, 2009 (ICPSR 28201)
program source code
The Stata syntax available as part of this data collection includes code to create 65 variables from the NATIONAL CRIME VICTIMIZATION SURVEY: SCHOOL CRIME SUPPLEMENT, 2009 (ICPSR 28201). One variable represents the subsample of cases selected from the 2009 SCS data file (as described in the Sampling metadata field). A total of 18 independent variables (describing direct, verbal, indirect, and cyber bullying behaviors), 4 dependent variables (describing response behaviors by those bullied), and 28 control variables (describing student and school characteristics) exist. The remaining 14 variables are dummy variables created for analysis purposes.