Journal of School Health · 2014 · 380 citations · 50 references
EducationEducational MeasurementEducational PsychologyAdolescent HealthPsychometricsStudent EngagementClassroom ClimateParent EngagementConfirmatory Factor Analysis InvarianceFull Measurement InvarianceInvariant Factor StructureMeasurement Model InvarianceMeasurement Parameter InvariancePerceived SafetySupportive Classroom ClimateSupportive Climate
School climate, encompassing safety, engagement, and environment, is linked to student behavioral, academic, health, and social‑emotional outcomes, and the USDOE has formalized this as a 3‑factor model. This study evaluates the factor structure and measurement invariance of the USDOE 3‑factor school climate model. Using two consecutive waves of data from over 25,000 high‑school students, the authors performed exploratory and confirmatory factor analyses of the Maryland Safe and Supportive Schools Climate Survey to test fit with the USDOE model. The analyses confirmed adequate fit of the 3‑factor model with 13 subdomains, demonstrated measurement invariance across sex, grade, and ethnicity, yielded low interclass correlations (0.04–0.10), and supported the 56‑item measure as internally consistent and efficient.
ABSTRACT BACKGROUND School climate has been linked to multiple student behavioral, academic, health, and social‐emotional outcomes. The US Department of Education ( USDOE ) developed a 3‐factor model of school climate comprised of safety, engagement, and environment. This article examines the factor structure and measurement invariance of the USDOE model. METHODS Drawing upon 2 consecutive waves of data from over 25,000 high school students (46% minority), a series of exploratory and confirmatory factor analyses examined the fit of the Maryland Safe and Supportive Schools Climate Survey with the USDOE model. RESULTS The results indicated adequate model fit with the theorized 3‐factor model of school climate, which included 13 subdomains: safety (perceived safety, bullying and aggression, and drug use); engagement (connection to teachers, student connectedness, academic engagement, school connectedness, equity, and parent engagement); environment (rules and consequences, physical comfort, and support, disorder). We also found consistent measurement invariance with regard to student sex, grade level, and ethnicity. School‐level interclass correlation coefficients ranged from 0.04 to .10 for the scales. CONCLUSIONS Findings supported the USDOE 3‐factor model of school climate and suggest measurement invariance and high internal consistency of the 3 scales and 13 subdomains. These results suggest the 56‐item measure may be a potentially efficient, yet comprehensive measure of school climate.
50