Work

Comparing school experience across Brazilian, Swiss, and Uruguayan adolescents: a measurement invariance study

Year: 2025

Type: preprint

Abstract: Background School experiences significantly impact adolescent development, yet measurement validity remains underexplored in diverse cultural contexts. Without establishing measurement invariance (MI), valid cross-cultural comparisons and educational interventions are limited. Methods This cross-sectional study evaluated MI of a school experience scale across three culturally and socioeconomically distinct urban contexts: Zurich, Switzerland ( n = 1,447), Montevideo, Uruguay ( n = 2,148), and São Paulo, Brazil ( n = 2,680). Participants were approximately 15 years old, attending secondary schools randomly selected to form representative samples in each city. Using multigroup confirmatory factor analysis with WLSMV estimation for ordinal indicators, we tested two distinct second-order factor models of school experience (with four and five dimensions). Latent means were compared using Empirical Bayes Modal estimation with Cohen’s d effect sizes calculated between cities. Results Our four-dimension hierarchical model (bond to class, bond to teacher, future orientation, and school difficulties) demonstrated excellent fit across all MI levels: configural, metric, and scalar (CFI = .953, RMSEA = .068, SRMR = .057). Model comparisons showed minimal changes in fit (ΔCFI ≤ − .01), confirming measurement equivalence. A five-dimension model including school commitment failed to achieve scalar invariance. Latent mean comparisons revealed that Zurich students reported more positive overall school experience than São Paulo ( d = 0.81) and Montevideo students ( d = 0.72), with particularly large differences in classroom bonds ( d = 0.72–1.16) and academic difficulties ( d = 0.56-.73), but negligible variation in future orientation ( d = 0.02-.12). Conclusions Despite substantial socioeconomical and cultural differences, adolescents across these contexts experience school similarly, with equivalent relationships between observed indicators and latent constructs. The established scalar invariance validates meaningful cross-cultural comparisons of latent means, while substantive differences highlight how educational systems and cultural contexts differentially shape specific dimensions of school experience. This framework provides a foundation for future comparative research and culturally-sensitive intervention development.

Source: Research Square

DOI: 10.21203/rs.3.rs-7160482/v1