One of the challenges in polling for elections is deciding who to include in your frame and who might actually turn out to vote.
(a) Suppose you were asked to conduct a poll for a senatorial election. Explain how you might design your sample. In your explanation include a discussion of the difference between "registered voters" and "likely voters." What role would stratification play in your sampling?
(b) Voter turnout is different for presidential election cycles (2012, 2016, 2020, and so on) versus non-presidential election cycles (2014, 2018, 2022, and so on). Explain the role election cycle plays in voter turnout and explain how this may affect your sampling methodology.
(c) During the 2014 election, Nate Silver of FiveThirtyEight said "the pre-election polling averages (not the FiveThirtyEight forecasts, which also account for other factors) in the 10 most competitive Senate races had a 6-percentage point Democratic bias as compared to the votes counted in each state so far." Explain what this means and explain how this would have impacted polling results compared with actual results.