A method for choosing the number of groups for item characteristic charts using an information criterion

Hideki Toyoda, Takashi Akiyama, Norikazu Iwama

Research output: Contribution to journalArticle

Abstract

Item characteristic charts are important for analyzing test items. These charts reveal characteristics of the items, including item difficulty and whether the items reflect an essential concept of the test. To make an item characteristic chart, examinees must be divided equally into an arbitrary number of groups. Because there is no established criterion for choosing the number of groups, the number has to be selected arbitrarily, or else decided on empirically. Therefore, when an item characteristic chart is to be prepared, it would be useful if the number of groups could be determined on the basis of a statistical criterion. In the present article, a method for choosing the number of groups for item characteristic charts is proposed, using an information criterion, and the applicability and advantages of this method are suggested through a simulation and examples of application to factual test data.

Original languageEnglish
Pages (from-to)209-225
Number of pages17
JournalJapanese Journal of Educational Psychology
Volume62
Issue number3
Publication statusPublished - 2014

Keywords

  • Akaike's information criterion (AIC)
  • Bayesian information criterion (BIC)
  • Information criterion
  • Item characteristic chart (ICC)

ASJC Scopus subject areas

  • Developmental and Educational Psychology
  • Education

Fingerprint Dive into the research topics of 'A method for choosing the number of groups for item characteristic charts using an information criterion'. Together they form a unique fingerprint.

  • Cite this