ERIC Number: ED187749
Record Type: RIE
Publication Date: 1980-Apr
Reference Count: 0
The Application of Rasch Model Equating Techniques to the Problem of Interpreting Longitudinal Performance on Minimum Competency Tests.
This study presents an application of the Rasch equating methodology on a minimum competency testing program in reading and mathematics. Common item equating was performed in two stages to link the 1978-1979 form of the New Jersey Minimum Basic Skills tests to the 1977-1978 form. Both fit to the Rasch model and stability of the common item pools were investigated prior to the actual equating. Equivalent raw scores derived from the Rasch methodology were compared with a prior linear equating. Special attention was paid to those raw score points around the state's cut-off score. Results from the study indicated moderate to good fit of the tests to the Rasch model. Several "unstable" equating items were found and the equating was therefore carried out in two ways: 1) using all twenty-five equating items, and 2) using just the "stable" equating items. Raw scores derived from both the Rasch and linear methods showed close though not perfect agreement. The equating using only the "stable" equating items changed six out of the seven equating tables. In only two of these six tables, did this change move the scores closer to that given by the linear method. (Author/GSK)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Linear Models; New Jersey; New Jersey Minimum Basic Skills Program; Rasch Model; Test Equating
Note: Paper presented at the Annual Meeting of the American Educational Research Association (64th, Boston, MA, April 7-11, 1980).