Masters Theses

Date of Award

12-1981

Degree Type

Thesis

Degree Name

Master of Science

Major

Psychology

Major Professor

Schuyler W. Huck

Committee Members

Kathlenn Lawler, Donald Wheeler

Abstract

This study was conducted to examine the effects of item format upon the variables of reliability and student performance while holding constant the factor of item content. The four item formats compared were completion, multiple choice, true-false, and a focused true-false version which utilized underlining of a word or phrase in an attempt to reduce true-false ambiguity.

The test items were developed with the objective of assessing knowledge of research design and internal validity issues. Each of the 28 items was expressed in four formats with stem wording held constant. Four test forms were created, each of which contained four seven-item sections of differing formats. The content order was identical throughout the four forms, but the order of presentation of the item format was varied with a 4 x 4 balanced Latin square layout. The tests were randomly distributed to a total of 80 graduate students from four different classes during a regular classroom period.

Statistical analysis consisted of performing a Latin square analysis of variance for each of the dependent variables, reliability and performance. Significant Fs at the .05 level of significance for both variables were followed by mean analysis for each variable with Tukey's procedure for pairwise comparisons.

The reliability results showed that the reliability of the completion format was higher than that of the true-false format, while the multiple choice, focused true-false, and true-false reliabilities were not found to be significantly different from each other.

With regard to the performance variable, completion items were found to be more difficult than those items expressed in true-false or focused true-false format. The study did not find significant differences in student performance between the true-false, focused true-false, and multiple choice formats. The underlining technique did not result in any apparent improvement of the true-false items since the focused true-false items were not found to be significantly different from the true-false items for either the reliability or the performance variable.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS