Assessment Excerpt

This is a section from a recent PUBL0055 final assessment.  As you can see, the questions ask you to read a data set (provided on Moodle) into RStudio, run the relevant analyses, and interpret your results. Other questions ask students to explain some of the broader concepts from the module. The midterm will consist

of 2-3 such sections, each referring to a different topic and data set.

Here is some additional guidance to help with your preparations:

• Unless otherwise stated, answers should be written in complete sentences. Be sure to answer all parts of the questions posed as succinctly as possible. You are not required to provide detailed interpretations unless specifically stated in the question.

• The word count does not include the code, your output, or any words (or numbers) contained within tables or figures.

• Please submit your type-written (numbered) answers in a single document (a word file or a pdf file). You can create an appendix section at the end which contains all the R code needed to reproduce your results.  Your code has to work when we run it.  You do not need to include the code that failed to run, but just the well-annotated, cleaned-up version. If you do not provide the code to a question, any written answer to that question will be disregarded.

• Do not screenshot or copy and paste any brute R output (e.g. lm(y  ∼ x)) into your answers. Create a formatted table that is easy to read.

Section 2: Political Parties and Policy Outcomes

Does which political party is in power matter for policy outcomes? This is an important question for political scientists to answer, not least because many theories of voting assume that voters hold governing parties to account on the basis of their performance in office.  If such “retrospective voting” is to occur, it must be the case that different political coalitions have clear and consistent effects on policy outcomes in the time between elections.

To determine whether this is the case, in this section you will use data from the US to analyse the effects of the party in power in US state governments (specifically, which party holds the governorship of the state) on a number of different policy outcomes. The data comes from 864 elections across 50 states in the US, plus

the District of Columbia. The variables included in the data are:

The data is stored in governors .csv.  Once you have downloaded this file and placed it in the relevant folder, it can be loaded into R as follows:

governors  <-  read .csv ("data/governors .csv")

Question 6 (6 marks)

a. For each of the 4 outcome variables, estimate a linear regression where dem_governor is the only explana- tory variable. Present the results in a table.  (2 marks)

b . Interpret the regression coefficients in each model.  (4 marks)

Question 7 (15 marks)

a.   Adapt the regression models that you estimated above to include two additional control variables: dem_majority_house and dem_majority_senate.  Estimate these regression models and present the re- sults in a table. (2 marks)