Statistical Analysis of Multiple Choice Testing
Abstract
The multiple-choice question exam is a very popular method of evaluation used by educators everywhere, The Air Command and Staff College Distance Learning Department uses multiple-choice exams for testing non-residence students, ACSC currently uses the Test Analysis and Development (TAD) software program's two quantitative measurements, Ease Index and Differentiation Index, to flag possible problem questions for qualitative review, They also use student feedback to flag questions for review, ACSC uses the Maxwell Academic Instructor School's Test Item Analysis Handout to examine qualitatively the flagged questions to determine which need revision, The purpose of this paper is to determine if the ACSC Distance Learning Department is doing a good job at test evaluation and whether there are better ways to determine the quality, effectiveness and fairness of multiple-choice questions, This paper compares the TAD program to the ITEMAN program for quantitative analysis, For qualitative analysis, prior studies and guidelines are compared to those used by AC SC, This paper found that ACSC is doing a good job at test evaluation by using the best threshold values for the Ease Index and Differentiation Index to flag items, This paper recommends that ACSC use the ITEMAN software program because of its ease, speed and superior output, This paper recommends that ACSC use the Biserial Correlation Coefficient as well to flag questions, Lastly, this paper recommends that ACSC use Hansen and Dexter's Item-writing Guidelines for qualitative review of flagged questions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2001
- Accession Number
- ADA407446
Entities
People
- Mark A. Colbert
Organizations
- Air Command and Staff College