Using repeated measures data to analyse reciprocal effects: the case of Economic Perceptions and Economic Values
Patrick Sturgis, University of Surrey

Peter Smith, Ann Berrington and Yougjian Hu, University of Southampton

RMP Project page

This paper will focus on the importance of modelling error variance and covariance structures when examining (possibly) reciprocal effects.

Slides

Discussant: Steve Fisher, University of Oxford

Addressing the problem of endogeneity in estimating area effects
Simon Burgess, University of Bristol

RMP Project page

Slides

Discussant: Peter Dolton, University of Newcastle and and London School of Economics

Missing data in multilevel models
James Carpenter, London School of Hygiene and Tropical Medicine

RMP Project page

Slides

We begin by briefly reviewing the problems introduced by missing data and how they affect the validity of standard analyses.

We then discuss the analysis of multi-level data sets with missing responses and give an example of a sensitivity analysis.

Finally, motivated by a study of the effect of class size on educational achievement, we compare multiple imputation and efficient inverse
probability weighting methods for handling missing covariates.

Discussant: Andrew Pickles, University of Manchester