Issues surrounding accurate assessment of depression in children have received much attention. However, the stability of scores from depression measures has generally been estimated using only classical test score theory, rather than the more powerful generalizability theory. The dependability of scores from the Children's Depression Inventory (CDI) was investigated using both generalizability and classical test score analyses. Data from 164 children aged 11 to 16 years from small Texas communities were the basis for the analysis. Results suggest that the sources of error variance interact to decrease the dependability of CDI scores. Several sample measurement protocols were also investigated. The results indicate that depression in children might be better assessed using planned multiple testing sessions to reflect depression more accurately. Two tables present study findings, and there is an appendix of selected study results. (Author/SLD)
Paper presented at the Annual Meeting of the American Educational Research Association (Atlanta, GA, April 11-16, 1993).