ERIC Number: ED052206
Record Type: RIE
Publication Date: 1971-Apr
Reference Count: 0
Assessment of Language Ability by Computer.
Whalen, Thomas E.
A study was made of the capability of machine scoring of essays to determine a student's proficiency in English mechanics. A sample of 71 seventh grade essays were entered into a computer via punched cards and were processed using a modified version of the PEGFOR program. Prediction models, subtest models, and restricted models were tested for predictive efficiency. The most reliable model constructed was composed of seven variables: (1) number of capitalization errors, (2) standard deviation of sentence length, (3) standard deviation of word length, (4) number of connectives, (5) occurrence of "then", (6) average sentence length, and (7) number of usage errors. When used to predict overall essay grades, using regression weights derived from the mechanics criterion, the correlation between actual and predicted scores for 24 essays was .60, indicating a strong relationship between mechanical proficiency and overall writing ability at the seventh grade level. It is believed that computer scoring of essays may prove to be an important tool in the individualization of composition teaching. (DB)
Publication Type: N/A
Education Level: N/A
Authoring Institution: N/A
Identifiers: Project Essay Grade
Note: Paper presented to the California Educational Research Association (April 1971)