Understanding uncertainty in school league tables

Authors
Leckie, G. and Goldstein, H.
Year
2011
Journal
Fiscal Studies, Forthcoming,
Abstract

In England, contextual value added (CVA) school performance tables are published annually by the government. These tables present statistical model based estimates of the educational effectiveness of schools together with 95% confidence intervals to communicate their statistical uncertainty. However, this information, particularly the notion of statistical uncertainty, is hard for users to understand. There is a real need to make school performance tables clearer. The media attempt to do this for the public by ranking schools in so-called "school league tables", however they invariably discard the 95% confidence intervals and in doing so encourage the public to over interpret differences in schoolsâ?™ ranks. In this paper, we explore a simulation method to produce simple graphical summaries of schools' ranks which clearly communicates their associated uncertainty.

Number of levels
2
Model data structure
Response types
Multivariate response model?
Yes
Longitudinal data?
Yes
Substantive discipline
Substantive keywords
Paper submitted by
George Leckie, Graduate School of Education, University of Bristol, g.leckie@bristol.ac.uk
Edit this page