The Effect of Removing Examinees with Low Motivation on Item Response Data Calibration

摘要

The purpose of this study was to evaluate the effect of removing examinees with low motivation on the estimates of test-item parameters when using an item response model (IRM) for large-scale assessment (LSA) data. This study was conducted using a Grade-9 LSA of mathematics. Current IRMs do not flag or filter the effect of low motivation on the estimates of test item parameters data calibrations used to assess examinee abilities and design exams for LSA. The effect of low motivation may pose a threat to the validity of test data interpretations. Motivation, as defined by expectancy-value and self-efficacy theory, was identified from self report data using a principal component analysis (PCA). The PCA scores were used to create two groups of examinees with high and low motivation to examine the effect of removing examinees with low motivation on the estimates of test item parameters when comparing a standard 3-parameter logistic (3PL) IRM to a 3PL low motivation filter IRM. The results suggested that test item parameters seemed to be overestimated under the 3PL IRM when examinees with low motivation were not removed from the test data calibration. The outcome of this study supports the literature and may provide an avenue to flag the effect of low motivation on LSA data analyses.