ERIC Number: ED368773
Record Type: Non-Journal
Publication Date: 1993-Dec
Reference Count: N/A
Some Measurement Characteristics of Aggregated versus Individual Scores.
Brennan, Robert L.
Not infrequently, investigators assume that reliability for groups is greater than reliability for persons, or that the error variance for groups is less than that for persons. Using generalizability theory, it is shown that this "conventional wisdom" is not necessarily true. Examples are provided from the course-evaluation and the performance-testing literature. In the cases considered in this paper, the conventional wisdom necessarily holds only for comparative statements about person versus group error variance when the universe of generalization has persons fixed and items random. In all other cases, the conventional wisdom may be false, in particular when the generalization is over both samples of persons and samples of items, which often represents the most sensible universe of generalization. An appendix elaborates on reliability. (Contains 8 references.) (Author/SLD)
Descriptors: Comparative Analysis, Course Evaluation, Generalizability Theory, Measurement Techniques, Performance Based Assessment, Reliability, Sampling, Scores, Statistical Studies
ACT Research Report Series, P.O. Box 168, Iowa City, IA 52243; American College Testing, 2201 North Dodge Street, Iowa City, IA 52243.
Publication Type: Reports - Evaluative; Numerical/Quantitative Data
Education Level: N/A
Authoring Institution: American Coll. Testing Program, Iowa City, IA.