Normally, exit polling is done on election day, to compare official results with how people say they voted immediately after having done so. Was that not done? (It's voluntary of course, but usually provides a good sample group of the voting population over all.)