Cluster sampling

Varying customer needs and adjustment to international trends made it necessary to elaborate a new, cost-effective sampling method to ensure reliable data quality. ‘Cluster sampling’ is the outcome of one and a half year preparation, and will be described in this guide.

1. Cluster sampling

1.1. Sampling

For sampling we use 174 subregions of Hungary. In addition, districts of Budapest, county seats and county towns are included, as they serve as a part of every nationally representative sample. In total 219 subregions are used to develop the sample.

Based on subregions, the country is divided in two parts, namely sample A and sample B. On a county level they have equal number of subregions with equal number of population.

CAPIBUS Example

Picture 1. Example of subregions selected for sample A and B::
In Csongrád county, subregions of sample A are highlighted with red and subregions of sample B are highlighted with green

Every half year when the respective samples are defined, towns and villages – chosen from different subregions - will be rotated. At the same time, districts of Budapest, county towns and cities will always be a part of the cluster sample.

1.2. Clustersk

In the settlements of the sample, the interviews have to be made in regionally concentrated clusters which are defined by street- and a housenumber-grids. In a particular survey, a cluster will be chosen in every selected settlement, except a few districts of Budapest and county seats. One cluster covers an area of cca. 800-1000 residents, where a defined number of questionnaires has to be conducted, considering the sampling aspects. The number of conducted interviews in the clusters will be at least 5 questionnaires. It is important that interviews can only be conducted within the clusters./p>