Measuring Mindsets and Achievement Goal Motivation: A Validation Study of Three Instruments

To evaluate the validity of scores from three instruments measuring achievement goal motivation-related constructs: a shortened version of Dweck's Implicit Theories of Intelligence Scale (ITIS-S), measuring incremental and entity mindsets; Elliot's Achievement Goal Questionnaire-Revised (AGQ-R), measuring mastery-approach, mastery-avoidance, performance-approach, and performance-avoidance achievement goals; and Midgley's Patterns of Adaptive Learning Scales (PALS), measuring mastery, performance-approach, and performance-avoidance achievement goals.

High school students participating in a medical simulation training activity in May 2017 completed each instrument. The authors evaluated internal structure using reliability and factor analysis and relations with other variables using the multitrait-multimethod matrix.

There were 178 participants. Internal consistency reliability (Cronbach alpha) was > 0.70 for all subscores. Confirmatory factor analysis of ITIS-S scores demonstrated good model fit. Confirmatory factor analysis of AGQ-R scores demonstrated borderline fit; exploratory factor analysis suggested a three-domain model (approach, mastery-avoidance, performance-avoidance). Confirmatory factor analysis of PALS scores also demonstrated borderline fit; exploratory factor analyses suggested consistent distinction between mastery and performance goals but inconsistent distinction between performance-approach and performance-avoidance goals. Correlations among AGQ-R and PALS scores were large for mastery (r = 0.72) and moderate for performance (≥ 0.45) domains; correlations among incremental and mastery scores were moderate (≥ 0.34). Contrary to expectations, correlations between entity and performance scores were negligible. Correlations between conceptually unrelated domains were small or negligible.

All instrument scores had good internal consistency and generally appropriate relations with other variables, but empirically determined domain structures did not consistently match theory.