Construct validation of the attitudes towards science scale: measurement invariance analyses on gender and grade level

BackgroundAlthough extensive research was done on gender and grade-level attitudes towards science differences, few studies have examined the measurement invariance of these scales before making cross-group comparisons. This gap raises concerns regarding the validity of the conclusions drawn from such comparisons.PurposeThis study addressed this issue by investigating the measurement and structural invariance of an attitude towards science scale across gender and grade level.Sample and methodsCross-sectional data were collected from 584 Turkish middle school students (55% female, 45% male). Multi-group confirmatory factor analysis was utilized for the examination of measurement invariance.ResultsResults of the multi-group confirmatory factor analysis supported partial measurement invariance, characterized by the invariance of all factor loadings, invariance of all but two intercepts and invariance of all but six-item uniquenesses for gender. On the other hand, there was a reasonable level of measurement invariance and partial structural invariance for 6th and 8th grades.ConclusionsBased on these results, future use of the attitudes towards science scale is recommended with paying particular attention to several items for gender and the construct of self-concept in science for grade level. The psychometric implications of the findings, as well as limitations of the study, are discussed.