Calculating the random guess score of multiple-response and matching test items

Silvester Draaijer*, Sally Jordan, Helen Ogden

*Corresponding author for this work

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

For achievement tests, the guess score is often used as a baseline for the lowest possible grade for score to grade transformations and setting the cut scores. For test item types such as multiple-response, matching and drag-and-drop, determining the guess score requires more elaborate calculations than the more straightforward calculation of the guess score for True-False and multiple-choice test item formats. For various variants of multiple-response and matching types with respect to dichotomous and polytomous scoring, methods for determining the guess score are presented and illustrated with practical applications. The implications for theory and practice are discussed.

Original languageEnglish
Title of host publicationTechnology Enhanced Assessment - 20th International Conference, TEA 2017, Revised Selected Papers
EditorsEric Ras, Ana Elena Guerrero Roldán
PublisherSpringer/Verlag
Pages210-222
Number of pages13
ISBN (Electronic)9783319978079
ISBN (Print)9783319978062
DOIs
Publication statusPublished - 2018
Event20th International Conference on Technology Enhanced Assessment, TEA 2017 - Barcelona, Spain
Duration: 5 Oct 20176 Oct 2017

Publication series

NameCommunications in Computer and Information Science
Volume829
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference20th International Conference on Technology Enhanced Assessment, TEA 2017
CountrySpain
CityBarcelona
Period5/10/176/10/17

Keywords

  • Cut score
  • Item-writing
  • MCQs
  • Multiple-choice questions
  • Question development
  • Selected response questions
  • Selected response test items
  • Standard setting
  • Test development

Fingerprint

Dive into the research topics of 'Calculating the random guess score of multiple-response and matching test items'. Together they form a unique fingerprint.

Cite this