Psychometric Approach to Error Analysis on Response Patterns of Achievement Tests.

Abstract

This study is an attempt to improve the quality of computerized adaptive testing as an integral part of instruction. An adaptive achievement test for teaching signed-numbers operations was implemented along with a computerized routing system on the PLATO system in 1977. A close investigation of the student's performance-scores on the posttest of this program led us to believe that a deeper level of considerations, not just the data scored right or wrong, in measuring students performance would be needed in future computerized tests. Diagnosing the misconceptions possessed by students is important not only for increasing efficiency of learning activities but also to score test-items properly. Some problems could be corrected by a wrong rule of operation. These may be called 'false corrects'. It is shown in this work that in some cases as many as 27 of the 64 test items need to be adjusted from right to wrong if we are to discredit 'false corrects'. Finding all types of wrong rules of operation associated with a particular teaching method for integer operations was tried by performing error analysis on some paper-and-pencil as well as on-line conventional tests.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 1980
Accession Number
ADA091713

Entities

People

  • Kikumi K. Tatsuoka
  • Maurice M. Tatsuoka
  • Menucha Birenbaum
  • Robert Baillie

Organizations

  • University of Illinois Urbana–Champaign

Tags

Communities of Interest

  • Biomedical
  • Human Systems

DTIC Thesaurus Topics

  • Achievement Tests
  • Air Force
  • Applied Psychology
  • California
  • Cognition
  • Cognitive Science
  • Computer Science
  • Educational Psychology
  • Illinois
  • Military Research
  • Personnel Management
  • Plastic Explosives
  • Psychology
  • Social Sciences
  • Students
  • Teaching Methods
  • United States

Readers

  • Artificial Intelligence
  • Psychometric Testing or Psychological Assessment.
  • Regression Analysis.