This study examines the accuracy with which the computer could simulate
human judgments of the overall quality of formal letters written by adults aged
26-35. The initial correlation between machine and human scoring was .85, and
a subsequent test of the generalizability of the findings yielded a correlation
of .78. Measures of fluency, diction, and spelling were important predictors of
overall quality.
Get full access to this article
View all access options for this article.
References
1.
National Assessment of Educational Progress, Report 3, NAEP, Denver, 1971. _______, Report 8, Denver, 1972. _______, Report 10, Denver, 1973.
2.
Page, E.B., and D. Paulus, The Analysis of Essays by Computer, Storrs, University of Connecticut, 1968.
3.
Slotnick, H.B., An Examination of the Computer Grading of Essays, Urbana , University of Illinois, 1971 (unpublished doctoral dissertation). _______, "Toward a Theory of Computer Essay Grading," Journal of Educational Measurement, 1972, pp. 253-63.
4.
, J.V. Knapp, and R.L. Bussell, "Bits, Nybbles, and Bytes: A View of an Electronic Clerk," Journal of Business Communication, 1971, pp. 35-52.
5.
Whelan, T., The Analysis of Essays by Computer: A Simulation of Teachers' Ratings, presented at AERA, New York, 1970.
