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.
An Application of the Continuous Response Model for Subtest Data
Item response theory (IRT) has become a popular and wide spread approach to measurement across many different fields in recent years due to its many benefits over classical test theory. However, traditional IRT models rely on item-level data and large sample sizes, which can be limiting factors in applied research. Researchers may find themselves limited to subtest or booklet scores due to issues in test security, storage of information, or due to relying on secondary data. The continuous response model (CRM) offers a way to obtain IRT estimates of ability when subtest or booklet data are available. The current study investigated the utility of the continuous response model when applied to subtest data. The current study hypothesized that the CRM applied to subtest data would result in estimates of ability on par with those obtained using a more traditional IRT model applied to item-level data. Additionally, it was hypothesized that the CRM would perform better at smaller sample sizes due to its parsimonious nature. The current study also investigated factors that influence the performance of the CRM to determine when it can best be utilized. Results from the simulation study showed that the continuous response model did not perform as well as the more traditional IRT model. However, item fit indices from the empirical study suggest that the CRM may perform better when items are especially difficult or discriminating. In both the simulation and empirical study there was a high degree of concordance between the CRM and the traditional IRT model. The CRM did not perform as well at smaller sample sizes, but overall performed quite well across all conditions. While the continuous response model may not perform exactly as well as the more traditional approach, it offers researchers a new avenue for IRT ability estimation when item-level data are unavailable but subtest or booklet scores are.
Educational psychology|Psychology|Quantitative psychology
Smith, Weldon Zane, "An Application of the Continuous Response Model for Subtest Data" (2018). ETD collection for University of Nebraska - Lincoln. AAI10843855.