| The purpose of this study was to determine whether test takers use different cognitive processes when they solve multiple-choice versus constructed-response items. I conducted this study in an era when school accountability and high-stakes, large-scale assessments were seemingly as important as student learning itself. The high-stakes nature of testing created an environment in which both testing advocates and challengers scrutinized tests even more than normal. In particular, some challengers asserted that multiple-choice items were ill-suited for assessing certain types of cognitive processes or for providing useful information about student achievement.; Using information as the central idea, I approached the question from an information-value (or value-added) perspective and attempted to determine whether there was a difference in the type of cognitive processes elicited by each item format. The question was narrowly contextualized in one area of mathematics, namely 8th grade, algebraic pattern items. I selected 34 students who were enrolled in 8th grade mathematics courses in the spring of 1998. I examined the question using a think-aloud procedure—an analysis tool seldom used in the field of measurement. The overall results suggested that students used similar cognitive processes to solve both multiple-choice and constructed-response pattern items. However, the results were likely related to the characteristics of the items—that is, many constructed-response items allowed for one solution path. I refer to these items as “multiple-choice items in disguise”. Recommendations were offered to test users, developers, and other researchers. |
