Weighting in the ESS

By Dr. Matthias Ganninger

Before we begin, we will first provide a brief overview of the organisational structure and the basic functioning of the European Social Survey (ESS). This will help us to better understand the differences and similarities in the countries' implementation of important aspects of the survey, such as sample design and fieldwork.

The ESS is managed by a Central Coordinating Team (CCT), an international group of experts in various fields of methodological research. Members of the CCT contribute their knowledge to work packages aimed at supporting central routine tasks facing the ESS in each round. The sampling expert panel is a group of five experts in the field of sampling who, in cooperation with the National Coordinator, develop a sample design to be applied in a given country and round. This development is guided by a principle that has become widely accepted in cross-national sample survey research. It states that 'workable and equivalent sampling strategies in all participating countries' have to be developed [Häd07]. This goal should be seen in conjunction with [Kis94]:

‘Sample designs may be chosen flexibly and there is no need for similarity of sample designs. Flexibility of choice is particularly advisable for multinational comparisons, because the sampling resources differ greatly between countries. All this flexibility assumes probability selection methods: known probabilities of selection for all population elements.’

Sample designs will therefore vary between countries and may even vary within the same country between rounds. This is not a shortcoming as long as the sample designs applied have some well defined properties that enable the researcher to plan and implement equivalent samples across countries. These properties include, but are not limited to, the availability of so-called inclusion probabilities.

This ESS Training Module will discuss aspects of sample designs that have an influence on design weighting as well as the construction and use of more complex types of weights. It is structured as follows: after briefly answering the fundamental question ‘why weighting?’ in Chapter 1, we will discuss some important and frequently used sample designs and their implications for design weighting in Chapter 2. Then, in Chapter 3, we will take a look at some often-used estimators. In Chapter 4, we will learn about the topic of weighting in the ESS over multiple rounds and/or countries. The last chapter, Chapter 5 goes more into detail about the effect of sample design.

The examples used in this Training Module use publicly available ESS data that are analysed using the online statistical software package Nesstar and R (also see [Rde09]).

Go to next chapter >>

References