Unravelling the complexity of school effects: a bivariate multilevel analysis of academic performance in Chile

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

Abstract

Educational research has largely demonstrated the effects of socio-economic characteristics of schools and pupils on academic outcomes. In the case of Chile, the long-standing problem of socio-economic inequalities has taken a huge toll on the education system, producing segregation to such degree that children attend schools whose quality depends on what their parents can afford, which has had a major impact on the equality of opportunities for decades.Traditionally, contextualised value-added models (CVA) have been employed to assess variation in learning outcomes arising from schools and pupils. However, these models rather unrealistically assume, amongst other things, that only the last school attended affects pupils' performance, while there are no differences between classes within schools and that school performance does not vary across areas.Using data from the 2004-2006 SIMCE database (Chilean National Pupil Database), a 5-level bivariate cross-classified CVA model was implemented to estimate simultaneously the variation in Mathematics and Language between pupils, secondary schools, primary schools, classes (within secondary schools) and local authorities. This extended CVA model for progress in Mathematics and Language proved effective in reducing bias present in the traditional univariate value-added models, while controlling for the correlation between the two subjects at all levels. The model also controls for the effects of prior attainment, income, gender, year repetition, school type and school SES.School performance tables are derived from this model and compared to school rankings derived from less complex models and key differences lead to the conclusion that further adjustments to the traditional CVA models are not trivial, especially in a highly unequal country such as Chile. In sum, the choice of statistical model is a sensible matter when using these indicators for informing stakeholders and policymakers. A fair policy of school accountability should take into account the complexity of the school performance phenomenon.
Original languageEnglish
Title of host publication VI European Congress of Methodology
Subtitle of host publicationUniversity of Utrecht, the Netherlands. 24/07/2014.
PublisherEuropean Association of Methodology
Publication statusPublished - 24 Jul 2014
EventVI European Congress of Methodology - University of Utrecht, Utrecht, The Netherlands
Duration: 23 Jul 201425 Jul 2014

Conference

ConferenceVI European Congress of Methodology
CityUniversity of Utrecht, Utrecht, The Netherlands
Period23/07/1425/07/14

Keywords

  • School value-added
  • Multilevel Modelling
  • Chilean education system

Fingerprint

Dive into the research topics of 'Unravelling the complexity of school effects: a bivariate multilevel analysis of academic performance in Chile'. Together they form a unique fingerprint.

Cite this