Abstract

Objectives To assess the between-school variation in suicide ideation and to estimate the contribution of school-level attributes, student-level characteristics, and 2 cross-level interactions (school by student) to student suicide ideation. Study design A secondary analysis of the California Healthy Kids Survey in 2 large and representative samples of California high schools and students: 2009-2011 and 2011-2013. This is a population sample of all public high school students (grades 9 and 11) in California. Analyses were first conducted on surveys administered in the 2011-2013 academic years to 790 schools with 345 203 students and replicated on surveys administered in 2009-2011 to 860 schools with 406 313 students. Results School-level suicide ideation rates ranged between 4% and 67%. with a median of 19.3% and mean of 20.0% (SD, 5.7%). Student suicide ideation was explained by student-level characteristics (R-2 = .20) and to a larger extent by school-level attributes (R-2 = .55). Student-level characteristics predictive of suicide ideation included, sex. ethnic and racial affiliation, victimization, and perceptions of school climate. In both samples. school size and average level of academic achievement were not associated with rates of school suicide ideation. Schools with a larger number of girls and higher levels of victimization had higher rates of suicide ideation in both samples. The hypotheses regarding cross-level interactions were not confirmed. Conclusions Differences among schools in student suicide ideation are meaningful. The findings suggest an emphasis on the role of schools in prevention programs, public health campaigns to reduce suicide, multilevel research. and theory development.