American Review of Mathematics and Statistics

Option Probability Theory: A Quest for Better Measures
John J Barnard

Abstract
Most test takers do sometimes guess when responding to multiple-choice questions. Different theories and models attempt to address this issue in different ways. It is argued that test takers seldom randomly guess; they rather eliminate one or more options based on partial knowledge. Option Probability Theory (OPT) deals with guessing as a personal interaction parameter through requiring test takers to indicate how sure they are about each option as being the correct answer. Through calculating a realism index an indication can be obtained about how realistic a test taker has assigned these probabilities and as such serves as an indicator of the amount of guessing and certainty of answers. A small study suggests that OPT results are comparable to traditional results, but with significantly more information about test takers’ knowledge and understanding. Practitioners can use this additional information to identify misconceptions, where there is uncertainty about answers and where guessing can be suspected. Using the realism index, a diagnostic analysis of response patterns can be made and detailed feedback can be given to test takers.

Full Text: PDF     DOI: 10.15640/arms.v3n1a7