Investigation of possible item bias on PISA 2015 science items across Chile, Costa Rica and Mexico (Estudio de los posibles sesgos entre los items de ciencias de la prueba PISA 2015 de Chile, Costa Rica y Mexico)


Dzul-Garcia C., ATAR B.

CULTURE EDUCATION, vol.32, no.3, pp.470-505, 2020 (SSCI) identifier identifier

  • Publication Type: Article / Article
  • Volume: 32 Issue: 3
  • Publication Date: 2020
  • Doi Number: 10.1080/11356405.2020.1785158
  • Journal Name: CULTURE EDUCATION
  • Journal Indexes: Social Sciences Citation Index (SSCI), Scopus, Academic Search Premier, EBSCO Education Source, Psicodoc, DIALNET
  • Page Numbers: pp.470-505
  • Keywords: item bias, PISA 2015 science items, cultural differences, Latin American countries, LOGISTIC-REGRESSION, EFFECT SIZE, COMPARABILITY, VERSIONS
  • Hacettepe University Affiliated: Yes

Abstract

Programme for International Student Assessment (PISA) provides valuable information to participating countries by allowing them to compare the performance of their students with the performances of students in other countries. When making such comparisons, the construct being measured should be the same across groups at the item and test levels. The purpose of this study was to examine the responses of students to 16 dichotomously scored science items in PISA 2015 for item bias using statistical procedures at the item level. Student data in three Latin American countries (Chile, Costa Rica and Mexico) were used to run the differential item function (DIF) analyses across cultural groups. About 25% of the items for each country pair were detected as DIF items with large effect sizes. Based on the findings of the study, it can be concluded that the possible item bias effect on the validity of the interpretation and the use of PISA 2015 science test results should be considered while comparing the test results across the cultures investigated in this study.