Abstract
Psychometric properties of recently developed measures of mathematics anxiety have not adequately been studied. Using a large sample (N = 815), the construct validity of the revised Math Anxiety Rating Scale (MARS-R) was assessed. Participants were randomly assigned to a validation or a replication sample. In the validation sample, confirmatory analyses revealed that the original model fits data poorly. The instrument was modified, resulting in the elimination of 12 items, and the fit of the two-factor model was improved considerably. The revised measure was applied to the replication sample, with comparable fit indices obtained and excellent generalization across male and female participants. Internal consistency reliability coefficients of the revised measure were strong, and good convergent and divergent validity was shown with the original MARS-R and other anxiety-related measures. The revised measure may represent a more parsimonious and valid measurement approach for assessing mathematics anxiety.
Get full access to this article
View all access options for this article.
