Publications & Resources

Methodological Approaches to Online Scoring of Essays

Dec 1997

Gregory K. W. K. Chung and Harold F. O'Neil Jr.

In this report, we examine the feasibility of scoring essays using computer-based techniques. We review two approaches that have been used to classify documents, surface- and word-based analysis techniques. We review two existing approaches: Project Essay Grade (PEG) (Daigon, 1966; Page, 1966, 1968, 1994; Page & Peterson, 1995) and latent semantic analysis (LSA) (Foltz, 1996; Landauer, Laham, Rehder, & Schreiner, 1997). We evaluate strengths and weaknesses of each approach and consider both near- and long-term issues of availability, application to text processing in general, and computational complexity. We believe that LSA and PEG represent first attempts to specifically address the essay scoring problem. Given their infancy, we are impressed by their current performance, and we view the direction of automated scoring techniques as extraordinarily promising.

Chung, G. K. W. K., & O’Neil, H. F. Jr. (1997). Methodological approaches to online scoring of essays (CSE Report 461). Los Angeles: University of California, Los Angeles, National Center for Research on Evaluation, Standards, and Student Testing (CRESST).