I like to do a post-hoc test to determine which group differs from the others. Is it khi 2 test with contingency tables followed by correction of bonferroni advised ?
Should I use another post-hoc test? Any suggestion would be welcome.
Yes, a common approach is to apply a chi-square analysis on the smaller component tables and apply a p-value correction. How you define the smaller tables will vary upon your purposes, but you might make all row-wise comparisons or column-wise comparisons. The bonferroni correction is usually too conservative, so you may want to choose a different correction method.