DOES THE POPULATION SIZE IMPACT THE Sample Clauses

DOES THE POPULATION SIZE IMPACT THE. CONDITIONAL GRANTS PER CAPITA? To finish the statistical analysis, analyses were conducted to assess the impact of the population size on the conditional grants per capita received by a municipality since literature shows that larger cities tend to receive more conditional grants per capita. As can be seen in Attachment 5.1, the conditional grants per capita is positively associated with the total population of a municipality. In fact, the association appears to be significant for both 2012 as 2013 (p <.001). Furthermore, the linear regression that was applied in section 4.2.1 was executed on solely the municipalities with 30000 inhabitants or more. This was done to check if the home bias is stronger in the larger cities. Initially, the goal was to apply the linear regression on the central cities in Flanders. However, this only consists of thirteen cases so the decision was made to enlarge the definition of a ‘city’ to 30000 inhabitants or more. This way, the linear regression could be applied on 47 of the 308 Flemish municipalities. Five of the 47 municipalities were home towns of Flemish ministers. The linear regression thus was executed with the central independent variable and the control variables. As can be seen in Table 15, in 2012, the regression model seems to be stronger than the linear regression applied in section 4.2.1. Model 1, with only the independent variable explained 26,2% of the variance in the dependent variable while Model 2, after adding the control variables, explained a total 38,9% of the variance. The actual SPSS output tables can be found in Attachment 5.1. Model R R² Adj. R² R² Change Sign F. Change 1 (predictors) .543 .294 .262 .294 .001 2 (control variables) .734 .539 .407 .244 .042 Table 15 section 4.2.3 Linear regression 2012 Furthermore, the estimated model coefficients were analysed. In Table 16, the standardized Beta value and significance values (* p <0.1, ** p <0.05, *** p <0.01) are represented for every variable. In Model 1, the dummy variable home town sp.a. minister appeared to be significant (p = .000). Even after adding the control variables, the variable remains significant (p= .032). Consequently, the assumption can be made that the home bias is more present in the larger cities. Moreover, the tariff additional personal income tax (p = .013), unemployment (p = .050), sp.a. in the local government (p = .069) and VLD in the local government appeared to be predictors of the dependent variable. Finally, the data wa...