Abstract
The purpose of this study was to explore the benefits of using Item Response Theory (IRT) for the Board of State and Community Corrections (BSCC) Selection Exam. The BSCC Selection Exam is a key component in the selection process for entry level workers in the Adult Correctional Officer (ACO), Juvenile Corrections Officer (JCO), and Probation Officer (PO) classifications within the state of California. Historically, the BSCC has used Classical Test Theory (CTT) methods to develop and analyze test forms. CTT is known to have some limitations such as sample dependency and oversimplicity that can be overcome by integrating IRT methods into the test development process. In this study, archival data from two alternate forms of the BSCC Selection Exam (N = 20,635) were used to test IRT assumptions, choose and run an IRT model, analyze item and person fit, and develop various graphs and figures to analyze trait levels and item difficulty levels. The exam was separated into four exam subdimensions in order to meet the unidimensionality assumption, which was considered a limitation of the study. The Rasch model was then chosen due to its alignment with the scoring of the BSCC Selection Exam. It was concluded that the average person trait levels exceeded the item difficulty levels, suggesting that more difficult items should be added to the exam. Overall, the Rasch model proved to be a useful tool for future exam development and allowed for a detailed analysis of the BSCC Selection exam.