Can Teachers' Global Ratings Identify Children with Academic Problems?

Abstract

Physicians often elicit ratings from teachers when making diagnostic, treatment, or referral decisions. The purpose of this study was to view the relationship between teachers' ratings and children's academic skills, assess the utility of teacher ratings in detecting academic problems, and thus determine whether physicians can depend on teacher ratings when making decisions about patients' needs. Subjects were a national sample of 80 teachers and 934 children between 6 and 13 years of age participating in a test standardization study. Families were representative of United States demographics in terms of parental level of education, income, and ethnicity, and sites were geographically diverse elementary schools. Children were administered the Comprehensive Inventory of Basic Skills--Revised (CIBS-R), a diagnostic academic achievement test. Teachers rated children's academic performance on a five-point scale ranging from far above average to far below average and were blinded to the results of the CIBS-R. Teacher ratings varied significantly with children's performance for all academic domains. Logistic regression revealed that teacher ratings were best predicted by children's performance in basic reading skills, followed by math skills, and were not influenced by race, parents' level of education, history of retention, or gender. Participation in Title I services, testing in winter or spring, and parents who spoke a language other than English produced significantly lower ratings. Nevertheless, teachers rated as average many students with mild to moderate academic difficulties. School system personnel and health care providers should avoid sole dependence on global teacher ratings when deciding which students need special education referrals or other services. Supplementing teacher ratings with standardized screening test results is needed to ensure accurate decision-making.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 27 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.