Surveys and aggregate data analysis - A discussion of the usability of different approaches in a comparative analysis of political behaviour

Researchers who make comparative analyses of political behaviour in sub-national territorial units must often choose between the use of aggregate data and surveys. The use of surveys has long been considered the superior option, as it allows researchers to avoid the danger of ecological fallacy connected with the use of aggregate data, but it is also an extraordinarily expensive option. The article elaborates the pros and cons of both methodological approaches, and suggests the use of a method that seeks to combine the advantages of both. The method is based on combining the information from surveys on the national level with the aggregate data describing the sub-national territorial unit which are available from other sources, like electoral statistics or censuses. The method using the program LOCCONTINGENCY is tested on the data from Czech regions and its usability is verified by comparing model results with the results of surveys that were conducted in four model regions.