A growing number of states and school districts use school climate assessments in progress reporting systems and are interested in incorporating these assessments into accountability systems. This analysis of response data from middle school students and teachers on the California School Climate, Health, and Learning Survey examines the appropriateness of the survey for identifying needs and monitoring changes in school climate. It analyzes survey response data for 730,160 grade 7 students from 2004/05 through 2010/11 in 1,117 California middle schools and 16,255 teachers from 2008/09 through 2010/11 in 812 California middle schools to examine the following research questions:
- What school climate domains in middle school do the survey items measure?;
- How reliably do the survey items measure school climate at the respondent and school levels?;
- How stable are the school-level climate measures?; and
- Are the school climate measures related to student achievement and discipline infractions?
Three analytic strategies were used: exploratory/confirmatory factor analysis, estimation of respondent- and school-level reliability, and estimation of correlations of school climate measures over time and with student academic performance and suspensions. The study finds that student and staff surveys validly and reliably assess distinct school climate domains, such as safety and connectedness, meaningful participation, bullying and discrimination, and caring staff-student relationships. All school-level domain measures were associated in expected ways with school-level student academic performance and suspensions. The following are appended: (1) The link between school climate and student learning; (2) School climate items on the California School Climate, Health, and Learning Survey student and staff surveys; (3) Methodology; (4) Detailed results; (5) Sensitivity analyses; (6) Student survey exploratory factor analysis results: Factor loadings from preferred model; and (7) Staff survey exploratory factor analysis results: Factor loadings from selected model.
