Use of Generalizability Theory for Estimating the Dependability of a Scoring System for Sample Essays

Abstract
This article illustrates the use of generalizability theory for evaluating the dependability of an experimental system for scoring essays. The proposed system was originally planned to be transportable, that is, usable by educators who were trained only by reading and using the materials included in the self-instructional package accompanying the prompt and prescored examples. Because an intended use of the system involved national norms for interpreting performance of students writing essays, both absolute and relative error variances were analyzed.

This publication has 5 references indexed in Scilit: