ERIC Number: ED171751
Record Type: Non-Journal
Publication Date: 1979-Apr
Reference Count: N/A
A Comparison of Linear and Rasch Equating Results for Basic Skills Assessment Tests.
Beard, Jacob G.; Pettie, Allan L.
Test results from the Florida Educational Assessment of third and fifth grade communications and mathematics skills were used to compare linear and Rasch equating results. The samples consisted of over 5,000 cases for each grade and content area. The tests contained some items common to both the 1976 and 1977 test forms, but no fewer than 20 percent of the total number of items in each test. Item discrimination and shifts in calibration were examined. Analysis indicated that there was little instability in each set of common items; that each set of common items contained a low proportion of misfitting items based on the discrimination index; and that the total tests contained a low proportion of misfitting items compared to Anchor Test Study data. Results indicated close similarity in reference T scores and high comparability between linear and Rasch equating results. Rasch procedures appeared to be applicable to negatively skewed mastery tests of basic skills. The common items demonstrated stability in difficulty parameter estimates, and very few items did not fit the model using the discrimination criterion. It was concluded that Rasch equating was comparable to linear equating in the analysis of longitudinal trends in basic skills achievement. (MH)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A