![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
Notes
1Information on the project and sample can be found in the introductory paper of this collection. No weighting has been applied to the data.
2The actual share was 2.54% of all EU-27 tax revenues in 2006. Calculation is based on EU-27 tax revenue in 2006 (Lupi Citation2008) and total commitment appropriation in 2006 (European Commission Citation2006).
3The questions used were asked on a 1–4 ordinal scale and this compelled us to make a methodological compromise as principal component analysis would require data measured on an interval scale. However, results were straightforward to interpret and all the statistical measures of reliability (Bartlett's test of sphericity and the Kaiser–Meyer–Olin statistic) of the analysis were suitable enough. Compared to other methods we could have chosen (such as cluster analysis) principal component analysis has the advantage that it provides us with individual factor scores which can be used for further analysis. Other scholars have used this methodology for the same purposes with similar conditions, see Bruter (Citation2004), Haller (Citation2003) and Carey (Citation2002).
4According to its very low correspondence with the factor structure of European identification, ‘to be a Christian’ could be left out of the analysis. However, in order to maintain comparability with factors of national identification we have decided to keep it.
5Logistic regression, contrary to linear regression, does not suppose a structure behind the observations and does not suppose the constant variation of the error term. However, the interpretation of the results is more complicated—the exponential of the regression coefficients are to be interpreted as odds of the probability of occurrence as opposed to the probability of non-occurrence of the dependent variable. Logistic regression estimates the regression coefficients using maximum likelihood estimation.
6It should be noted that the logistic regression models presented here do not have a very high explanatory force in explaining between 8% and 19% of the total variance (adjusted R-square), the proportion of explained variance being lowest in the case of the tax redistribution. This implies that although the explanatory variables are significant, other exogenous factors might play a role.