Teacher ratings and proficiency scores : relationship with student academic self-concepts

Students' academic self-concept is known to be multidimensional and very domain specific. Self-concept in reading is expected to be correlated with achievement in reading whereas self-concept in maths is expected to be correlated with achievement in maths. Other self-concepts such as art and physical self-concepts may not have any direct association with reading and maths achievement scores. As students' academic selfconcept is primarily developed through their experiences in classroom environments, teachers may have significant influence on students' self-concept development. As such, we may speculate that the respective correlations between self-concept and achievement would be stronger for teacher ratings than standardised scores of achievement. A sample of Year 3 to Year 6 Australian students (N = 1790) responded to 5 self-concept scales in a survey (i.e., school, reading, maths, art, physical abilities). Their standardised reading and numeracy test scores, as well as their teachers' ratings in reading and maths were also obtained. Confirmatory factor analysis found positive correlations between reading self-concept and standardised reading test scores and between maths self-concept and standardised maths test scores, but the correlations between the respective self-concepts and teacher ratings were close to zero. The results cast doubt on the accuracy of teacher ratings in reflecting students' self-concepts in respective learning domains. Standardised tests may have merits in providing a realistic assessment of students' proficiency despite their limitations and unpopularity in primary levels of education.