If the rows and columns of a contingency table are independent, we can find the expected cell counts for all cells in the table, based only on the table's margins.
Repeat with a few different questions.