I have the following problem where I have to test the hypothesis for difference in proportions in the voters for the Republicans party.
$$ \begin{array}{c|lcr} n & \text{East} & \text{Middle West} & \text{South} & \text{West} \\ \hline Republicans & 275 & 228 & 202 & 248 \\ Democrats & 225 & 272 & 298 & 252 \\ Total & 500 & 500 & 500 & 500 \end{array} $$
I understand that I have to use Chi-Square Test statistic and then see whether it overpasses the significance level,i.e.:
$T=\sum\frac{(Observed_i-Expected_i)^2}{Expected_i}$
I don't really get what is the expected value in this case. I assume the Observed is $\frac{275}{500}, \frac{228}{500}$, and so on, but what is the expected proportion considering the fact that we have two rows. If the rows was 1 I would just divide n by 4, but in this case I'm not entirely sure what to do.