Presentation (10 marks)
Points will be deducted for bad presentation, which includes (but is not limited to):
- Failure to write the answers in full sentences
- Failure to clearly indicate which question is answered where and which code pertains to which question
- Including screenshots from R output
- Including long print outs of data sets and objects (e.g., using View(), show())
- Reporting unrounded numbers
- Including unnecessary code or output
- Presenting figures with no or unclear axis labels (or labels that are the unedited variable name in the dataset)
- Presenting tables that are hard to read/not well formatted
- Presenting unnumbered and/or untitled tables and figures
- Referring to the variables in-text by their unedited name from the dataset
Corruption Incidence and Local Health Councils in Brazil
The relationship between well-functioning local institutions and corruption is an important issue across the globe.
To explore the relationship between local governance in the health sector and incidence of corruption, we will conduct some analyses loosely based on the research conducted in Brazil by Avelino, Barberia, and Bilderman (2014).
To measure how established a health council is, Avelino et al. recorded how long a local health council had been established at the time of the audit. Health councils are responsible for overseeing the local provision of health services in the municipality, as well as approving the municipal health budget and monitoring expenditures.
The authors collected data from a set of Brazilian municipalities that had been randomly selected to be audited by the federal government. Detailed memos are produced for each source of federal funding in the selected municipalities, including for health grants. Auditors include “evidence reports” to these memos,depending on the number of irregularities identified.
To measure corruption, the authors looked through the evidence reports for federal grants, and counted the number of such reports that mention irregularities. The municipal-level percentage of these evidence reports that mentioned irregularities was treated as the corruption index score, ranging between 0 and 100.
The author’s primary hypothesis is that the more well-established a health council is in a municipality, the more likely corrupt practices will be uncovered. This is because the authors believe that “local governments acquire expertise to manage the health system over time and each additional year represents a marginal gain in local capacity”, including in the control of corruption. Thus, we would expect more established health councils (that is, those that are older) to have a lower incidence of corruption.
You can download the data set as trading.csv from the PUBL0055 Moodle page. Once you have downloaded this file and placed it in the relevant folder, it can be loaded into R as follows:
brazil <- read.csv(“data/brazil.csv”)
Question 1 (6 marks)
We are first interested in exploring the data set and conducting some descriptive analyses.
a.For how many of the municipalities do the authors have no data on the age of the health council? (2 marks)
b.Plot and interpret a boxplot of the health council age (council.age). (2 marks)
c.Interpret the median and mean of the variable corruption (2 marks)
Question 2 (8 marks)
We then proceed with a simple linear regression analysis.
a.Fit and present a simple linear regression with the corruption index as the outcome and age of council as the explanatory variable. (1 mark)
b.Discuss the statistical and substantive significance for the intercept and the estimated regression coef ficient for council.age. Is the intercept meaningful in this model? (4 marks)
c.Under which assumptions can we interpret the regression coefficient as the average effect of council age on corruption? (3 marks)
Question 3 (10 marks)
As the authors did in the original study, we now add a number of other municipal-level explanatory variables to our regression model: margin of victory for the Mayor in the last election; whether the Mayor is re-elected; and the poverty level.
a.Fit a multiple linear regression model, adding margin, reelected, and poverty to the model in the previous question. Present this model alongside the simple linear regression model. (1 mark)
b.How has the estimated coefficient for council.age changed? What does that tell us about the variables we have added to the model? (3 marks)
c.Discuss and compare the model fit for the multiple and the simple linear regression models. (3 marks)
d.What is the predicted corruption index score for a municipality health council that is 10 years old, that has a re-elected Mayor, where the Mayor won the last election by 12 percentage points, and where the poverty level is 50? (3 marks)
Question 4 (14 marks)
Although it was not explored in the original paper, we are interested in whether the relationship between the age of the health council and incidence of corruption differs between municipalities with and without a reelected Mayor.
a.Fit a multiple linear regression model, adding an interaction between reelected and council.age to the multivariate model from the previous question. Present this model alongside the model without the interaction. (1 mark)
b.Interpret the estimated coefficient for margin. You do not need to discuss statistical significance. (2 marks)
c.Calculate and interpret the 95% confidence interval for the estimated coefficient of poverty. (3 marks)
d.Interpret the relationship between council.age and corruption. (3 marks)
e.Using the model you estimated in 4.a, calculate the fitted values for health councils with ages between 0 and 20 years, separately for municipalities with and without reelected mayors. Set the electoral margin to 10 percent and poverty score to 50 percent. Present the fitted values visually and describe what the graph shows. (5 marks)
Asset trading and attitudes to peace
What are factors that determine the extent to which people support peace processes? Research suggests that (ethnic) violence tends to harden ethnic identities and increase out-group discrimination (see, e.g.,Shayo & Zussman 2017). While one approach to decrease exclusionary attitudes has been to encourage individuals to put themselves in the shoes of members of the out-group (see, e.g., Adida, Lo & Platas 2018),Jha & Shayo (2019) argue that exposure to financial markets might be another factor that can help promote more inclusionary, pro-peace attitudes in protracted conflicts. The basic idea is simple: conflict tends to be financially costly. Financial markets “demonstrate the shared risks from conflict and the returns from peace” (p.1561-1562). Therefore, individuals who invest in financial assets that are negatively affected by conflict (e.g., stocks of companies located in conflict areas) might have a better understanding of the financial risks of conflict and, in addition, be (financially) negatively affected by conflict. Accordingly, they hypothesize that individuals exposed to financial markets will become more pro-peace. To test their theory, they conduct a field experiment in Israel, in which they randomly assigned a sample of Israeli voters to a financial asset treatment group or a control group. Individuals in the treatment group received vouchers to invest in specific stocks or indices from Israel and the Palestinian Authority. Participants were surveyed before and after the experiment.
本网站支持 Alipay WeChatPay PayPal等支付方式
E-mail: email@example.com 微信号:vipnxx