Analysis of complex survey data

The measurement of change over time is a central problem for many users of social, economic and demographic data and is of interest in many important areas of economics and the social sciences.

Common problems are:

• How to compare two cross-sectional estimates for the same study variable that are taken on two different occasions or waves.
• How to judge whether the observed change is statistically significant
  In this ESRC-funded project, Yves Berger and Pedro Silva are investigating new approaches to estimating standard errors for estimates of changes over time. The methods that are under development allow for both the rotating structure of the repeated survey and the effect of nonresponse. The properties of the proposed estimators are being evaluated and compared using theory, simulation and empirical comparison based on their application to two major national social surveys: the UK Labour Force Survey and the Brazilian Labour Force Survey.