אלה אנגל

For Massive Open Online Courses to have trustworthy credentials, assessments in these courses must be valid, reliable, and fair. Item Response Theory provides a robust approach to evaluating these properties. However, for this theory to be applicable, certain properties of the assessment items should be met, among them that item difficulties are stable over time. The present study evaluates whether this property applies to the assessments of these courses using a case study – an AP Physics course provided by MITx. To do that, we estimated the item parameters in three administrations of the course and compared them across the administrations. We found that while many items did not meet certain quality criteria, more than a third of the items had stable characteristics over time. Our results demonstrate that Item Response Theory can help evaluate assessment quality in Massive Open Online Courses and suggest that these high-quality items can be used as the core of the course assessment as their properties are consistent across runs. We recommend that creators and instructors of online courses apply similar procedures to evaluate their courses' assessment and instructional design. Making sure the items have similar properties over time can support the assessment validity, reliability, and fairness of online courses.