A Microcomputer-Based Evaluation Subsystem for the Textual Component of an Instructional System

Publisher Website
Google Scholar
Abstract
The purpose of this study was to extend the notion of evaluation to determine and improve the reliability and validity of an evaluation instrument. This process was facilitated by a microcomputer testing program designed to collect performance data of both students and the evaluation instrument. The study was conducted in two cycles. In the first cycle, potentially flawed items were identified. These items were then pulled from the test pool and examined by the instructional staff. Thirty-four items out of the 120 test items were flagged as being potentially faulty items. These flagged items fell into five categories: 1) items containing more than one correct answer, 2) items which were incorrectly keyed, 3) items with no correct answers, 4) items with “give-away” answer foils, and 5) items in which no flaws were identified. The faulty items were revised by the instructional staff and reinserted in the test pool. In cycle two of the study, the entire test pool, including the revised items, was readministered. Of the thirty-four flagged items in cycle one, twenty items fell within the acceptable range during cycle two. An added benefit to these revisions was that the mean score for each of the six behavioral objectives reflected by the test pool increased, while the standard deviation for each objective decreased. In addition to the discussion of this process in terms of reliability and validity, a discussion of the use of microcomputers as a means of facilitating this evaluation and redesign process is also given.

This publication has 2 references indexed in Scilit: