Achievement data in IEA studies and Simpson's Paradox

Ruth Zuzovsky*, David M. Steinberg, Zipi Libman

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


This paper is meant to highlight the occurrence of Simpson's Paradox when using aggregated data obtained from two IEA studies in Israel, while ignoring the effect of a powerful intervening variable in the local context - the ethnicity factor. It will demonstrate faulty conclusions regarding either the absence of relationships between a contextual variable and achievement - when such relationships do exist, or conclusions on the existence of such relationships - when in reality they do not. Our intention in writing this paper is to draw the attention of our fellow researchers to similar faulty inferences they might come across when analyzing their local database within the scope of international comparative studies.

Original languageEnglish
Pages (from-to)144-151
Number of pages8
JournalStudies in Educational Evaluation
Issue number2-3
StatePublished - Jun 2011


  • Achievement data
  • Evaluation methods
  • Simpson's Paradox


Dive into the research topics of 'Achievement data in IEA studies and Simpson's Paradox'. Together they form a unique fingerprint.

Cite this