Equating for Long-Term Scale Maintenance of Mixed Format Tests Containing Multiple Choice and Constructed Response Items

An item response theory–based equating method is proposed for the long-term scale maintenance of a mixed format test consisting of constructed response items and multiple choice items. The proposed method is a modification of the traditional common-item nonequivalent groups design. Rather than collecting equated item parameters from previous tests, the methodwould require the collection of actual examinee responses from previous years and the rerating of those responses in a linking study conducted for the current equating. The resulting design controls for any year-to-year changes in the standards of the rating team, isolating the effect of any change in the examinee ability distribution in the computation of the equating transformation. Simulated data are used to illustrate the proposed method and demonstrate the possible consequences of an inappropriate use of traditional equating. The strengths and weaknesses of the proposed method relative to traditional common-item nonequivalent groups equating with...