Testing the Invariance of the Metacognitive Monitoring Test

Abstract Metacognition is predominantly measured by the self-report and think-aloud methods. This is problematic since they produce considerable both respondent and confirmatory biases, which implies damage to the measurement. The Metacognitive Monitoring Test (MMT) was created to evaluate metacognition through performance and eliminate the aforementioned biases. There is evidence of MMT convergent, divergent, structural, predictive and incremental validity. This article expands the validity studies about the MMT by analyzing the configural, metric and scalar invariance of MMT across sex, nationality, and educational level variables. The sample is composed of Brazilian and Honduran subjects, as well as 6st to 12st grades and higher education students. Results indicate configural, metric and scalar invariance for the sex variable, as well as configural invariance and metric and scalar partial invariance for nationality and educational level. It is concluded that the MMT allows comparing means of the latent variable measured in the analyzed groups.