In a Swedish longitudinal twin study, teacher ratings of school adjustment were collected in grades 3 and 6 for approximately 80 pairs of MZ twins, 100 pairs of DZ like-sex twins, and 70 pairs of opposite-sex twins. These same groups of twins then rated their own school adjustment in grades 4 and 6 as seen by the home, the school, their classmates, and themselves.
A comparison of within-pair similarity for the different types of self-ratings tends to show more consistency and a higher concordance for MZ compared to DZ in grade 6 than in grade 4.
The teacher ratings tend to show a larger and more consistent difference between MZ and DZ than the twins' own ratings.
The results reported have certain implications for heritability estimates based on different types of ratings. Ratings by others thus seem to give the highest intraclass correlations, probably due to a certain halo effect. Self ratings, on the other hand, tend to fluctuate more over time at least for children before and at puberty. Also the construction of items seems to influence the magnitude of the correlations.
