1. Between February 20 and 24, 2018, a Suffolk University/USA Today Poll asked a random sample of 1,000 registered voters nationwide, "How much trust do you have in President Trump's denial that there was any collusion in his campaign with Russian meddling: a lot of trust, some trust, or little or no trust?"

The table below summarizes their responses:

Response %

A lot 24%

Some 12%

A little or none 57%

Unsure 7%

(a)Calculate a 95% confidence interval for the percentage of respondents who have a lot of trust in Trump's denial of collusion. Show your work. (Hint: Imagine every response is coded 1 or 0 those that responded "a lot" are coded 1, and all other responses are coded 0. See pages 139-140 of the Kellstedt & Whitten book for a similar example.)

(b) Calculate a 95% confidence interval for the percentage of respondents who have little or no trust in Trump's denial of collusion. Show your work. (Hint: This time, consider "a little or none" responses to be 1's and all other responses to be 0's.)

(c) In plain English, interpret each of these confidence intervals.

2.Imagine this survey was repeated with 5,000 respondents, and it produced the exact same results.

(a) Re-calculate a 95% confidence interval for the percentage of respondents who have a lot of trust in Trump's denial of collusion. Show your work. (b) Re-calculate a 95% confidence interval for the percentage of respondents who who have little or no trust in Trump's denial of collusion. Show your work.

(c) Use this new sample to explain the effect of sample size on confidence intervals.

3.President Donald Trump thinks the U.S. Court of Appeals for the Ninth Circuit "has a terrible record of being overturned" by the Supreme Court.1 Let's hypothesize that the court whose decision is reviewed by the Supreme Court affects the disposition of the case (i.e., whether they reverse the lower court, affirming the lower court, etc.). The Supreme Court Database includes both of these variables.

(a)Which is most appropriate to test our hypothesis: tabular analysis, a difference of means, or correlation? Why?

(b)Perform the test you identified in part (a) using Stata. Do the data support the hypothesis? Use a significance level of p < .05. Explain.

(c)Include the Stata commands you entered and the outputs Stata produced

4.Let's hypothesize that a state's unemployment rate affects its alcohol consumption (as measured in gallons per capita). The States dataset (available on Blackboard) includes both these variables (unemploy and alcohol).

(a) Which is most appropriate to test our hypothesis: tabular analysis, difference of means, or correlation? Why?

(b) Perform the test you identified in part (a) using Stata. Do the data support the hypothesis? Use a significance level of p < .05. Explain.

(c) Include the Stata commands you entered and the outputs Stata produced.

5.Let's hypothesize that whether or not an individual believes that abortion should be legal for any reason a woman wants one affects his/her partner's socioeconomic index (measured on a scale where higher values indicate higher levels of prestige). The 2016 General Social Survey (data available on Blackboard) collected data on both these variables (abany and cosei10).

(a) Which is most appropriate to test our hypothesis: tabular analysis, difference of means, or correlation? Why?

(b) Perform the test you identified in part (a) using Stata. Do the data support the hypothesis? Use a significance level of p < .05. Explain.

(c) Include the Stata commands you entered and the outputs Stata produced..