Measuring Explanatory Variables
Population size is interpreted as the number of people living in the municipality. There are two types of data that can be used to measure size perceived this way.
First, the T-STAR dataset contains data on municipalities’ permanent population, i.e. the number of people who registered their main address in the municipality. As most people vote where registered, this measure is closer to the voting body of municipalities. The second indicator includes not only permanent, but also temporarily registered inhabitants. This measure indicates the number of people who actually participate in the everyday life of the municipality. I chose the second option for reasons of data quality. The 2001 census found much more people in Hungary than the previous administrative data, based on citizens’ registration, had suggested. After the socially turbulent years of the 1990s, census data are widely considered to be more reliable than the incrementally updated data collected by the police and public administration. The census gathered data on the people found in a certain place on a certain data, disregarding their officially registered address. To keep consistency, certain other data, e. g. the occupational structure of local societies, has also been taken from the census database.
As I mentioned, the local government system is highly fragmented in Hungary. That means that the range between very small villages and relatively large cities is huge: the largest local government analyzed here (Debrecen, 211,034 inhabitants) has more than 16,000 times more inhabitants than the smallest one (Tornakapolna, 12 inhabitants). In other words, the standard deviation around the mean (2,686 inhabitants) is very large (9,468). Such an enormous difference may cause a problem in multivariate statistical analyses, as the independent variables must meet certain conditions. To overcome this problem of skewedness, I often break down the analysis into population categories (usually by electoral systems, which form two distinct population categories).