Evaluating Alternative Item Response Theory Approaches Accounting for Not-Reached Items in Word Dictation Responses of Students With Significant Writing Difficulties

The purpose of this study was to evaluate the Word Dictation performance of elementary students with significant writing difficulties using a classical item response theory (IRT) approach that considers writing accuracy and alternative IRT approaches that consider both accuracy and productivity. We used responses to Word Dictation Curriculum-Based Measures (CBM) in the writing of 523 elementary students from a larger efficacy trial of a professional development program designed to support teachers in implementing data-based instruction for students with significant writing difficulties. We found that classical approaches treating not-reached responses as either incorrect or missing may lead to under- and overestimation of writing accuracy of students with low and high productivity levels, respectively. Handling missing responses with alternative approaches can lead to more accurately estimating writing proficiency, and improved understanding of both productivity and accuracy could ultimately support teachers’ use of data to individualize instruction.