Secondary analysis of large-scale assessment data: an alternative to variable-centred analysis

Kui Foon CHOW, Kerry John KENNEDY

Research output: Contribution to journalArticlespeer-review

16 Citations (Scopus)

Abstract

International large-scale assessments are now part of the educational landscape in many countries and often feed into major policy decisions. Yet, such assessments also provide data sets for secondary analysis that can address key issues of concern to educators and policymakers alike. Traditionally, such secondary analyses have been based on a variable-centred approach that gives rise to league tables. In the study reported here, a person-centred analysis is used as an alternative to the traditional approach. Data from the International Civic and Citizenship Education Study (ICCS) were analysed to investigate Asian students' attitudes to their future civic participation. Cluster analysis with validity measures showed that 4 distinct groups of students were identified within the societies studied thus highlighting the diversity within the samples. These results cannot be achieved with a conventional variable approach to analysis, and they suggest the usefulness of exploring alternative approaches to secondary data analysis. Copyright © 2014 Routledge.
Original languageEnglish
Pages (from-to)469-493
JournalEducational Research and Evaluation: An International Journal on Theory and Practice
Volume20
Issue number6
Early online dateAug 2014
DOIs
Publication statusPublished - 2014

Citation

Chow, K. F., & Kennedy, K. J. (2014). Secondary analysis of large-scale assessment data: an alternative to variable-centred analysis. Educational Research and Evaluation: An International Journal on Theory and Practice, 20(6), 469-493.

Keywords

  • Variable-centred analysis
  • Active citizenship
  • Citizenship education
  • Cluster analysis
  • Large-scale assessments
  • Person-centred analysis

Fingerprint

Dive into the research topics of 'Secondary analysis of large-scale assessment data: an alternative to variable-centred analysis'. Together they form a unique fingerprint.