Using Generalizability Theory to Examine the Dependability of Scores from the Learning Target Rating Scale

The Learning Target Rating Scale (LTRS) is a measure designed to evaluate the quality of teacher-developed learning targets for embedded instruction for early learning. In the present study, we examined the measurement dependability of LTRS scores by conducting a generalizability study (G-study). We used a partially nested, three-facet model to estimate the variance of LTRS scores due to teachers, children, learning targets, and raters. We used these variance components to conduct a decision study (D-study) to investigate how data collection and scoring design affected the dependability of scores and to help inform future use. Findings supported the dependability of LTRS scores when used with one rater and when at least six learning targets are obtained from teachers. We discuss potential refinements for the LTRS based on the G- and D-studies and implications for using it in practice. [This article was published in "Topics in Early Childhood Special Education" (EJ1157940).]