Two different election polls are held in the US:
In the first 100 US voters were sampled (at random) and asked whether they would vote Republican or Democrat (poll 1).
In the second 1000 US voters were sampled (at random) and asked the same question (poll 2).
Some researcher has a crystal ball and knows that of all US voters (the population) 55% will vote Democrat and 45% will vote Republican. Based on this result, the Democrats would thus win the election.
1. Given all of this information, estimate how likely it is that poll 1 and poll 2 will predict the opposite (Republicans win in stead of Democrats, meaning that Republicans should have at least 51% of the votes in the sample).
2. Imagine that only New Yorkers would have been sampled. Would that make any difference?