Cross-Cultural Assessment: Examining Academic Engagement Measurement Equivalence Among Adolescents in Colombia and the United States

• In Event: Poster Session 06
  In Poster Session: PS 06 Topic - Methods, History, Theory

Abstract

Academic engagement (i.e., school involvement school and learning persistence learning) is well established as promotive of positive academic outcomes in the United States (U.S.; Lam et al, 2014; Skinner et al., 2008). As attention to global education standards and student outcomes expands and draws comparisons to U.S. performance, research studies conducted outside the U.S. have increasingly utilized measures originally developed in the U.S. Given this growing practice, coupled with interest in the construct of academic engagement, there is a need for more research that tests the cross-cultural equivalence of existing measures of academic engagement. The present study tested the theorized 2-factor structure of the Academic Engagement Scale (i.e., behavioral and emotional engagement) in samples of Colombian and U.S. adolescents.

Sample 1 comprised adolescents from schools in Medellín, Colombia (N=462, Mage=15.90, SD=1.15, 47% female). Sample 2 comprised Latino adolescents from a school in the Southwest U.S. (N=370, Mage=16.14, SD=1.10; 54% female). All participants completed a 10-item self-report measure of academic engagement (Skinner et al., 2008; Sample 1 in Spanish: behavioral α=.78, emotional α=.75; Sample 2 in English: behavioral α=.81, emotional α=.80). Additional descriptive results in Table 1.

Results from confirmatory factor analytic (CFA) 2-factor models by sample suggested model fit below the acceptable threshold (Hu & Bentler, 1999) for Sample 1 (χ2(26)=222.57, p<.05; CFI=.88; RMSEA=.13; SRMR=.06) and Sample 2(χ2(26)=177.33, p<.05; CFI=.89; RMSEA=.13; SRMR=.08). A multi-group CFA model using sample as the grouping variable tested the factor structure equivalence across samples. Relatively poor fit indicated lack of configural invariance (χ2(76)=585.16, p<.05; CFI=.83; RMSEA=.12; SRMR=.13), suggesting the dimensions of academic engagement may be conceptually distinct across these two cultural contexts. The discussion will focus on future directions for developing and validating culturally relevant measures, as well as the implications of these results for assessing academic engagement across linguistically and culturally diverse contexts.

Authors

• Shandra M. Jones, Harvard University
  Presenting Author

• Michael R Sladek, University of Oklahoma
  Non-Presenting Author

• Adriana J. Umaña-Taylor, Harvard University
  Non-Presenting Author

• Liliana Uribe Tirado, Universidad de San Buenaventura
  Non-Presenting Author

• Luz Tilano Vega, Universidad de San Buenaventura
  Non-Presenting Author