An Interrogative Model of Computer-Aided Adaptive Testing: Some Experimental Evidence

Abstract

This thesis proposes and validates an interrogative diagnostic model of computer-aided, adaptive testing. The model uses an algorithm based on Bayesian techniques to determine both the number and the difficulty level of questions to present. The power of the model lies in its ability to automatically adapt a testing session to the individual's level of knowledge by choosing the domain, type and number of questions. The model was implemented as a computer program and was tested in an experiment with 34 military officers who were masters degree students. Analysis of the results indicated that the IDM can evaluate student knowledge as effectively as traditional written examinations but with a significantly less number of questions and shorter test duration. Keywords: Interrogative diagnostic model.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1988
Accession Number
ADA201675

Entities

People

  • Patrick A. O'donnell

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Algorithms
  • Application Software
  • Artificial Intelligence
  • Charts
  • Computer Programs
  • Computers
  • Databases
  • Factor Analysis
  • Information Systems
  • Instructors
  • Probability
  • Reasoning
  • Schools
  • Security
  • Students
  • United States
  • United States Naval Academy

Readers

  • Instructional Design and Training Evaluation.
  • Statistical inference.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference