ESS6 Errors in post-stratification weights

An error has been detected in the ESS Round 6 post-stratification weights (PSWEIGHT) in the ESS data available on the Data and Documentation pages.

At the current time, we would advise you not to use the current Round 6 post-stratification weights.  Initial assessments of the impact of this problem on univariate estimates of ESS substantive variables have shown they are small in most cases, and we can therefore expect the impact on multivariate models to be quite small.  However, the revised post-stratification weights will be needed to fully assess this. The design (DWEIGHT) and population weights (PWEIGHT) can continue to be used.

The ESS post-stratification weights are constructed using country information on age-group, gender, education, and region, which is obtained from the Labour Force Survey (LFS).  The problem relates to a misinterpretation of the population age distribution provided by LFS, which leads to a small overrepresentation of respondents aged 55+ in the ESS sample when the post-stratification weights are used.  Therefore, the post-stratification weights (variable PSPWGHT) for Round 6 need to be revised and will become available by early April 2016. We will inform all our data users once the revised post-stratification weights are available for download from our website.

We would like to sincerely apologise for any inconvenience caused and thank you for your patience.

(15/02/16)

[ARCHIVE]