ERIC Number: ED177199
Record Type: RIE
Publication Date: 1979-Apr-9
Reference Count: 0
A Comparison of Methods of Vertical Equating.
Loyd, Brenda H.; Hoover, H. D.
Rasch model vertical equating procedures were applied to three mathematics computation tests for grades six, seven, and eight. Each level of the test was composed of 45 items in three sets of 15 items, arranged in such a way that tests for adjacent grades had two sets (30 items) in common, and the sixth and eighth grades had 15 items in common. In the experiment sixth grade children took either the sixth or seventh grade tests, while children in seventh grade took sixth, seventh, or eighth grade tests and eighth grade children took seventh or eighth grade tests except for a few (91) who took the sixth grade test. The total group for all three grades was 1,956 pupils. The LOGIST program was used with a one parameter option. Several sets of estimations were obtained for each level for separate groups of children, and the obtained ability estimates were placed on a common scale through the use of overlapping items. Inconsistencies among equatings suggested that the test data did not fit the model of the one parameter logistic curve. Analysis of the content showed differences between tests for each grade level. Finally, equipercentile equating methods were applied to the data with a result that differed from the result of the Rasch method. (CTM)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: One Parameter Model
Note: Paper presented at the Annual Meeting of the National Council on Measurement in Education (San Francisco, California, April 9-11, 1979)