NotesFAQContact Us
Search Tips
ERIC Number: ED455246
Record Type: Non-Journal
Publication Date: 1997-Mar
Pages: 67
Abstractor: N/A
Reference Count: N/A
Estimating Minimum Sample Sizes in Random Groups Equating.
Tsai, Tsung-Hsun
The primary objective of this study was to find the smallest sample size for which equating based on a random groups design could be expected to result in less overall equating error than had no equating been conducted. Mean, linear, and equipercentile equating methods were considered. Some of the analyses presented in this paper assumed that the test scores were normally distributed. Other analyses were not based on this assumption. Real test data were used to check whether the theoretical methods provide reasonably accurate results for use in estimating sample size requirements. The science subtest of the ACT assessment provided the basic data for investigating the standard errors of equating and the minimum sample sizes needed to obtain less equating error than the identity equating. In general, as the sample size increased, the magnitude of the standard errors decreased for both forms of the test considered. In linear equating, the standard error becomes less as the raw score value approaches the mean score. In equipercentile equating, with nonnormality assumptions, raw scores less than or equal to 10 are associated with greater standard errors but the standard errors become smaller as the raw score approaches the middle score. Based on these results, it is reasonable to conclude that standard errors become less as sample size increases, and that they tend to be less for middle scores than the extreme scores for both the linear and equipercentile methods. (Contains 14 tables, 11 figures, and 12 references.) (Author/SLD)
Publication Type: Numerical/Quantitative Data; Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Identifiers - Assessments and Surveys: ACT Assessment