To Choose or not to Choose Multiple-Choice

Conference contribution, 2019

In this study we analyze the outcome of introducing Multiple-Choice (MC) questions to assess engineering students’ ability to apply basic concepts in previously unfamiliar problems (in the field of semiconductor devices). Using Item Response Theory (IRT) it is observed that MC questions of the design we have employed can work very well, while at the same time presenting a risk of resulting in test items that are worse than any of the Constructed Response (CR) questions they are compared to. We argue that MC test items intended for assessment of engineering skills have a higher need of proper validation before use than CR problems.