skip to main content skip to footer

Proposed Factorial Study of English Composition

Committee of Examiners in English
Publication Year:
Report Number:
ETS Research Memorandum
Document Type:
Page Count:
Subject/Key Words:
English Composition Tests, Essay Tests, Factor Structure, Research Needs, Test Construction, Writing Evaluation


The Committee of Examiners proposes that an experiment be conducted to determine the reliability of and the degree of correlation among various types of tests, both objective and non-objective. The goal of the experiment would be the determination of the measurable abilities and skills that make up proficiency in writing, the relationships among these skills, and the most effective methods of measuring them. Essay variables for this study are enumerated and discussed. Eleven essay types which should be used are listed and it is proposed that each be graded separately for: 1) material and organization, 2) spelling, 3) grammar, 4) punctuation, 5) diction, 6) sentence structure, and 7) general quality. It is further proposed that the list of these 77 (11 x 7) possible variables be reduced to 10 or 15 in the preliminary studies. The 11 essay types are designed to measure: 1) the effect of speededness on writing, 2) whether an essay written on one of several choices of topics results in measuring a different ability from one written on a single question, and 3) what abilities are measured by different types of essay. A supplement to the report sets forth the steps which should be taken as an initial phase of the study. A factor analysis of the rather large matrix of intercorrelations of test scores, essay scores, English grades and ratings, will constitute the main body of the study. This factor analysis should reveal whether the essays or new types of test measure only the established factors measured by present tests or whether they measure some other writing factors. Specific suggestions for preliminary studies are included in an appendix.

Read More