Exploring Person-Fit Patterns in High-Stakes Alternative Educational Testing

• In Event: AERA Roundtable Session Friday 8:00 am Four Seasons Ballroom 4
  In Roundtable Session: Advancing Psychometric Modeling: Innovative Techniques for Educational Assessment (Table 8)

Abstract

Person-fit statistics hold practical significance because both score interpretations and educational policy decisions heavily rely on the validity of individual scores. However, misfitting statistics from students with significant cognitive disabilities in K-12 assessments have not been adequately investigated. This study conducts person-fit analysis utilizing data from a high-stakes alternative assessment to explore diverse response patterns among student groups across subjects and grades and to examine the potential impact of misfitting responses on item parameter estimates. Results show that 1) English Language Arts exhibits a higher percentage of misfitting responses than math and science, for both the entire student sample and subgroups; 2) only 11 of 561 items (1.9%) have their b-parameter estimates affected by misfitting responses across subjects and grades.

Authors

• Jie Chen, Measurement Incorporated

• Ning Jiang, Measurement Incorporated