A multiple choice test (59 items) was given to 141 sophomore males as a final examination in General Psychology. The students were required to indicate their degree of confidence on each item by writing to the left of their answer a 1 (pure guess), a 2 (some guessing-some certainty), or a 3 (complete certainty). The scoring involved making the confidence values positive or negative according to whether the selected answers were correct or incorrect. Three scoring procedures were then compared by correlation techniques to determine if there were any major shifts in student ranking. It was found that by scoring the number right or by the algebraic total of confidence scores there was very little change in rankings. It was recommended that the multiple choice test using discrimination-in-weighting scores had advantages over either traditional number-right scoring or formula scoring. The discrimination-in-weighting score took into account the students' degree of confidence on each item, thus allowing him to determine his own weighting without causing a major change in his class rank.
Proceedings of the Iowa Academy of Science
©1968 Iowa Academy of Science, Inc.
Gannon, Timothy J. and Sannito, Thomas
"Discrimination Weighting on a Multiple Choice Exam,"
Proceedings of the Iowa Academy of Science, 75(1), 318-323.
Available at: https://scholarworks.uni.edu/pias/vol75/iss1/44