VIP panel composition
Every month we examine the opinions of Dutch citizens on numerous currently relevant topics through online panel research. For this purpose, we have put together a special VIP panel. This panel consists of about 16,000 Dutch citizens and the results are representative of the Dutch population (18+) in terms of gender, education, age, region (Nielsen region), migration background and voting behavior in the House of Representatives elections.
The VIP panel is composed on the basis of stratified random sampling from the opt-in non-probability Kieskompas panel, taking into account five characteristics: gender, age category, region (Nielsen region), education level, and migration -background. For this we use the Golden Standard of the Central Bureau of Statistics (CBS). Where necessary, the VIP panel is expanded by snowball sampling.
In general, between 35% and 50% of the invited respondents takes part in our online opinion polls.
Kieskompas has separate panels in a number of regions. That is the case in the province of Gelderland, for example. These panels are also composed on the basis of a stratified random sample from the opt-in non-probability Kieskompas panel, taking into account four characteristics: gender, age category, education level, and migration background. For this we use the Golden Standard of the Central Bureau of Statistics (CBS). Where necessary, the panels are expanded by applying snowball sampling.
In order to make the results generalizable for the Dutch population, aged 18 or older, we weight the collected data by means of an iterative post-stratification method based on various population characteristics.
The standard applied weighting method is explained below. Specific datasets may be slightly modified to better suit the research design.
Weighing is imlpemenented with two joint distributions. The first comes from the Golden Standard of the Central Bureau of Statistics (CBS):
- Age (4 categories),
- Gender (2 categories),
- Education (2 or 3 categories)
- Nielsen region (5 or 6 categories) or region (12 categories)
- Migration background (2 or 3 categories)
The second distribution concerns the election results of the House of Representatives elections 2017 by province (source: electionsuitslagen.nl or the Electoral Council). Each sitting party is weighted separately per region, just like all other parties together, blank voters and non-voters. Young people over the age of 18 but were not yet allowed to vote in the previous parliamentary elections are weighed separately.
Weights above the 99.5th percentile are normally trimmed.
For datasets collected outside of the Netherlands, we utilise census data from the respective national election studies, Eurostat, or other relevant data.
VIP panel composition
Looking for more information about our research? Click on the links below: