Squared differences between observed and expected frequencies are used in chi-square tests because
a) variances always use squared quantities.
b) differences between observed and expected frequencies sum to zero.
c) the squaring operation weights each cell appropriately.
d) nominal data categories require it.