A Study of Tests of Writing Ability
Abstract
A number of experimental tests designed to measure qualities important in report writing were administered to 506 students at a graduate school of business administration. Data used in the study included scores on the experimental tests, the Cooperative English Test, and the Admission Test for Graduate Study in Business, as well as first-year average grades and grades in a report writing course. For a sub group of 162 students, evaluations of paragraphs and sentences from three reports written by each student were also obtained. Separate "set" and "content" scores were obtained from two of the experimental tests. Reliabilities of the experimental tests ranged from .145 to .79. The reliability of the sentence-paragraph evaluations was estimated to be only .59. The set scores, which are thought to measure criticalness, were found to have low negative correlations with measures of ability, while the content scores, particularly the content score of Recognizing Ambiguities, behave like conventional measures of verbal ability. Two other experimental tests, the Interlinear Test and the Selection and Organization Test, are reasonably reliable yet have low correlations with conventional measures of verbal abilities. The best of the experimental tests for predicting writing ability are Recognizing Ambiguities (content score), Alternative Expressions (set score), and the Interlinear Test. The Interlinear Test was found predominantly to reflect accuracy. A negative correlation was found between subscores based on items measuring accuracy and organization.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1958
- Accession Number
- AD0203571
Entities
People
- Norman Fredericksen
Organizations
- Educational Testing Service