RESUMO
The construct of a school collective efficacy to prevent bullying has attracted attention as a way to increase a positive, school-wide climate. The current study tested the fit of several first-order models of school collective efficacy to prevent (uni-dimensional, two-dimensional, and three-dimensional) bullying using a sample of 579 male (Mage = 14.31, SD = 1.78 years old) and 589 female (Mage = 14.56, SD = 1.83 years old) Mexican adolescents. The models were validated by the extent to which the model was invariant by gender and by educational level (secondary vs. high school). Moreover, the discriminant and concurrent validity of model dimensions were examined through their relationships with other constructs. The results suggest that school collective efficacy is a three-dimensional construct, with supporting evidence for cohesion, students social control, and teachers social control dimensions. Measurement invariance was found in this three-dimensional measurement model by gender and educational level. The latent means difference analysis showed some differences by gender and educational level on factors of school collective efficacy. Finally, results support our hypotheses related to discriminant and concurrent validity in relation to external variables. Overall, findings indicate this three-dimensional model is useful to measure adolescents perceptions of school collective efficacy. (AU)
El constructo eficacia escolar colectiva para la prevención del acoso escolar ha recibido una creciente atención como una forma de mejorar el clima escolar. En el presente estudio se evaluó el ajuste de varios modelos de primer orden para medir la autoeficacia escolar colectiva para la prevención del acoso escolar (unidimensional, bidimensional y tridimensional), utilizando una muestra de 579 chicos (Medad = 14.31, DS = 1.78 años) y 589 chicas (Medad = 14.56, DS = 1.83 años) mexicanos. Para validar el modelo en distintos grupos se analizó la invarianza de medida por sexo y nivel educativo (secundaria vs. bachillerato). Además, se examinó la validez discriminante y concurrente analizando la relación de la eficacia escolar colectiva con otros constructos. Los resultados indican que la eficacia escolar colectiva es un constructo tridimensional conformado por las dimensiones de cohesión y confianza, control social de los estudiantes y control social de los docentes. La evidencia constata la invarianza de medida del modelo por sexo y nivel educativo. El análisis de las diferencias de medias latentes muestra diferencias por sexo y nivel educativo en las dimensiones de la eficacia escolar colectiva. Finalmente, los resultados confirman las hipótesis relativas a las relaciones de la eficacia escolar colectiva con otros constructos. Los hallazgos sugieren que el modelo tridimensional es útil para medir la percepción adolescente de la eficacia escolar colectiva. (AU)
