After the common support and overlap conditions were met, the balancing of covariates was done
between participants and non-participants. Caliendo and Kopeinig (2005) provide several procedures
for testing the balancing condition, including standard bias, t-test, joint significance, pseudo-
R2 and stratification.We used the t-test and found that before matching therewere significant differences.
However after matching, there were no significant differences implying that the covariates were balanced
(Table 2).