Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Testing the robustness of item response theory equating to violation of the condition of equivalent item difficulty parameters in both the extant pool and infused items: A comparison of the effects on true -score and observed -score equating
This study, to assess the robustness of Item Response Theory (IRT) equating to violation of the condition of equivalent item difficulty parameters in both the extant pool and the infused items, was conducted in two parts. Two types of IRT equating methods, true-score equating and observed-score equating, were investigated. ^ The first part of the study included a comparison of the two equating methods across the raw score scale for baseline and experimental conditions. In the baseline condition, both the original item pool and the infused items shared item difficulty parameters in the easier range of the theta scale ( b < +1.5). In the experimental condition, the same original item pool was used although the infused items contained difficulty parameters concentrated in the more difficult range of the theta scale (b > 0.0). The second part of the study investigated the differences in the true-score and observed-score equating methods in intervals of ten across the raw score scale. ^ Part one yielded little difference in the results of the two equating methods. On average, the difference between true-score and observed-score equating for the baseline condition was 0.3 units with true-score equating yielding results more similar to the raw scores. For the experimental condition, on average, the difference between true-score and observed-score equating was 0.32 units with true-score equating again showing results more similar to the raw scores. ^ The results of the second part of the investigation indicated that in both the baseline and experimental conditions, both true-score and observed-score equating appeared to have produced results more similar to the raw scores in the lower extreme of the scale. ^ Further research is necessary to replicate and extend these results to ensure that the results will generalize from the simulated data used in the present study to actual data. ^
Education, Tests and Measurements
Roos, Linda Lee, "Testing the robustness of item response theory equating to violation of the condition of equivalent item difficulty parameters in both the extant pool and infused items: A comparison of the effects on true -score and observed -score equating" (2002). ETD collection for University of Nebraska - Lincoln. AAI3041360.