Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

Independent Analysis Project Guidelines

For your independent analysis project, you will answer a question using data and regression tools learned this

quarter. You will produce a very brief high-level policy memo detailing your fndings, and provide between

2-5 additional items (at least one fgure and one table +3 optional items) to present your fndings in more

detail.  Please choose your topic from the three categories listed below, and use the data provided for that

topic (data fles are available on Canvas).

You are expected to work independently on this project and not collaborate with any classmates. Please read these guidelines carefully and in their entirety, as it is your responsibility to understand all of the project requirements. me last section of these guidelines is a page outlining all of the assignment rules. Submission of this assignment indicates that you understand and have adhered to all of the assignment rules, and have worked entirely independently.

In all cases, you must do cross-sectional analysis (i.e., no time series or panel), and part of your job will be to fgure out how to best address your topic using a cross-sectional framework. Across topics, you may only use the provided versions of the data. You are not required to do any outside research, but you may. You must cite any sources that you use in accordance with normal academic integrity standards.

1.    Topic Summaries

(A)  Environmental Policy:

Environmental and Sociodemographic Brazilian Municipal-Level Data

Global warming is one of the most pressing issues today. Yet, any global strategy that seeks to address this crisis requires a deep understanding of local cases. Due to its massiveness, Brazil’s Amazon rain- forest is one crucial local case.  me Legal Amazon encompasses 775 Brazilian municipalities, and it covers 59% of its territory. It represents 67% of the world’s tropical forests. A third of the world’s trees are in the region, in addition to 20% of global freshwater reserves. However, this vastness is a double- edged sword. If preserved, the Amazon rainforest could contribute to cooling down the planet; if not, a sizeable stock of carbon may go to the atmosphere as trees are burned to open space for croplands.

In fact, 44% of all Brazilian greenhouse gas (GHG) emissions were attributedto changes in land use in

2019.

But what are the main drivers of deforestation in the Legal Amazon?  What could account for the increasing numbers of emissions in this crucial region?  me diversity of human culture, ethnicities, as well as economic activities, and political variables may all play a role. You may consider questions related to how municipal characteristics impact the environment:

–  Do municipalities with better educational standards generate less emissions?

–  What are the impacts of agriculture production on deforestation? Do diferent crops have coun- tervailing impacts? What about livestock activities?

–  Is state capacity a relevant dimension for fghting environment degeneration?

–  Are poverty and inequality associated with emissions or deforestation?

me available cross-sectional dataset compiles a wide variety of information at the municipality level from all the 775 municipalities of the Legal Amazon (administrative region). me data includesgreen- house gas (GHG) emissions and deloration rates, as well as recent data on demographicseconomic activity, and state capacity.

(B)  International Development:

Country-Level World Development Indicators Data

me World Bank’s World Development Indicators is a database containing 1,400 indicators for 217 Countries, over 50 years.  You will be working with a fraction of the dataset only focusing on one

year: 2015. A selected number of indicators have been provided; extended information is available for each of the indicators in the “Codebook.xlsx” fle.

Use the dataset provided to answer one of the following questions or come up with your own research questions:

–  How does education inluence unemployment rates?

–  How do any of these indicators afect GDP?

–  Does higher alcohol consumption lead to more deaths?

–  Are there gender gaps in any of these relations?

You can fnd more information about World Development Indicators here:

https://datatopics.worldbank.org/world-development-indicators/

(C)  Political Behavior and Opinion:

Cross-National Individual-Level Survey Afro-Barometer 2019 Data

Political scientists and policy makers care about people’s beliefs and opinions. In the context of democ- racy promotion, we are ofen interested in what determines how people view the government and its institutions. We also care about citizens’ views on issues from gender to climate change to create efec-

tive policies. Use the dataset to explore an important dimension of public opinion or political behavior in Africa.

me dataset is an individual-level dataset from the most recent round of the Afro-Barometer in 2019. It contains responses to questions about political opinions and behaviors across 34 African countries. It includes information on 45,838 respondents (49.9% male, 50.1% female & 42.8% Urban, 55.5% Ru- ral).  To generate ideas for your own research, think about what determines people’s public opinion and political behavior—e.g., their educational background, their socio-economic situation, or even the experiences they had with political institutions. me dataset ofers questions on a variety of interesting topics for you to examine including, for example, questions about beliefs in democracy, corruption, women’s rights, trust in institutions, or the government’s handling of important issues. Make sure to check for missing values/non-responses in the data. men perform a statistical analysis that could help us understand people’s opinions on these matters in these African states.

Here are some ideas to get you started, but many alternative questions are possible:

  Governance performance:

–  Are people with better economic situations more likely to believe that their government is doing well? How does the experience of crime relate to evaluations of government perfor- mance? What if a person fears violence?

–  Does exposure and access to government institutions improve people’s evaluation of govern- ment performance?

  Trust in Institutions:

–  What inluences citizens trust in electoral institutions?  What is the efect of ethnicity or gender on trust in institutions? Do people who experience particular hardships show more skepticism in institutions?

–  Does the exposure to political institutions decrease the belief in corruption? Are people who experienced coercion or crime more likely to view their government as corrupt?

  Media Freedom:

–  Some people think the media should be able to report freely and without government con- trol. Others believe that the government has the right to control the media. Are people who believe that their government is performing well more likely to support media restriction? How about people who view their government as undemocratic?

You can fnd more information about the Afrobarometer survey here:

https://afrobarometer.org/publication-round/round-7

(D)  Social Policy and Global Health:

Demographic and Health Survey (DHS) 2011 in Uganda

As a policy maker, it is important to analyze the available resources in order to maximize health bene- fts. me Demographic and Health Surveys (DHS) program has collected representative data on health and nutrition in over 90 countries. mis dataset provides you selected cross sectional information on households in Uganda in 2011.

Possible research questions you may explore include:

–  How does Human Development Index (HDI) impact the probability of getting vaccinated?

–  Does owning a television improve health outcomes in children in households in Uganda?

–  Does a child’s vaccination status depend on the mother’s educational level?

–  Do households that have a female head spend diferently on, say, education?

You can fnd more information about this survey here:

https://dhsprogram.com/methodology/survey/survey-display-399.cfm

2.    Guidelines

You are responsible for three deliverables (all of which will be submitted electronically via Canvas/Turnitin):

•  MEMO: A 1000-word maximum, single-spaced policy memo presenting your research and conclu- sions.  me font should be 12 point Times New Roman for text, including footnotes and references. You will fnd it hard to summarize everything you have done in this amount of space.  But you must concisely describe how you used data to answer the question at hand, your main fndings, and the fun- damental limitations to your analysis in clear language for a reader that may not have a comprehensive understanding of regression analysis.  (Any references you include will not count against your word total.)

•  SUPPORTING INFORMATION: A combination of up to 5 additional items—Figures or Tables—that help present your research. You must have at least one fgure and one table; you can add up to three more items (graphs or tables). mese fgures and tables should include descriptive captions that let them stand on their own, and should ft on 3 pages maximum (you may have small margins for your fgures page). mese items should be clearly labeled, included in a separate section following your memo, and should be referenced from within your memo.

•  SCRIPT: A .do fle that replicates all of the analysis for your memo and supporting information. Please comment your .do fle so that we can easily navigate your code (e.g., /* Generate Figure 1: Interaction Efects */).

Your analysis (all three components considered together) should have the following general structure:

1.  Motivation and meoretical Underpinnings: You should begin by presenting the motivating ques- tion, or set of questions, and a clear explanation of the theory guiding your analysis.  WHY are you doing what you are doing? Are there intellectual schools of thought that guide your intuition? What hypothesis (or hypotheses) are you testing and what do you expect to fnd?

2.  Data Selection: You must explain how you use the data to test your hypothesis and answer the question at hand. What are the data you are using and why can they help you answer the question of interest?

What is the dependent (outcome) variable? (Or if you opt for more than one dependent variable, what are these?)  What is (are) the independent variable(s) of interest?  mis should include a discussion of case selection in light of your theory: explain why you are using the subset of data you are using (both in terms of observations and variables).  You should also include a concise description of any data manipulations/variables you have generated (how and why). Anyone who reads your paper and looks at your do fle should be able to easily replicate your analysis.

3.  Methodology / Explanation of Model(s): You should present your model(s) with clear justifcations for your variable selection and the functional form of your variables, including any interaction terms. What are you controlling for, and why?

4.  Regression Analysis and Results: Your main analysis should be a series of regressions testing the im- pact of your independent variable(s) of interest on your outcome variable.  All models should be re- ported in a clearly-labeled regression table on your supporting information. Explain the progression of your analysis clearly (e.g., adding other independent variables; testing interactions, etc.). Use graphics and simulations where appropriate. Discuss which model(s) have the strongest statistical and practical

signifcance. Interpret the meaning of your coefcients in a useful manner and discuss the goodness of ft of your model.

5.  mreats to Validity, Regression Diagnostics: Your analysis should include discussion of potential vi- olations of the Gauss-Markov Assumptions. If you exclude variables because of high multicollinearity,

please explain why, and present the appropriate diagnostics.  You should discuss potential problems with the Zero Conditional Mean and Homoskedasticity assumptions. If such problems exist, discuss the implications for your analysis. Deal with these problems as you are able; ifyou are unable to address them sufciently in your analysis, discuss the impact on your ability to estimate regression parameters and conduct hypothesis testing.

6.  Discussion and Conclusion: You should conclude with a thoughtful summary of your results, and a clear set of policy-relevant conclusions. You should also discuss the limits of your analysis, including problems with the data (e.g., selection bias and measurement error).  How would you improve this research design? What would be the next steps in your research?

A fnal note: You will do much more analysis than you can present in the memo and graphics. A huge part of the work here will be in compressing what you have done into your fndings. You will need to spend time on the writing and presentation, so make sure to leave yourself time to do that. We recommend spending Week 8 getting to know your data, reading, and planning your analysis; Week 9 should be devoted to running all of your analysis; and Week 10 should be devoted to writing, editing, and putting together your fnal deliverables.

3.    Grading Rubric

me project is out of 100 points total:

45pts: Memo

•  Have you clearly articulated the motivation for the analysis, your theory, and how you used the data to test your hypotheses?  What outcomes are you examining, and what is/are the main independent variables of interest? Are there other observable implications of your theory/hypothesis? How did you examine those? (10pts)

•  What are alternative explanations for what you fnd and how did you account for them? What kinds of controls did you use and why? How did you balance the various goals of model building (thorough- ness v. simplicity, etc.)? Have you interpreted your models clearly and correctly? Does your analysis progression make sense? You should be telling a story here. (13pts)

•  Have you addressed potential issues with data (outliers, measurement error) and violations of the Gauss-Markov assumptions, and dealt with them as you are able? (10pts)

•  Have you summarized your fndings with appropriate policy-relevant conclusions and discussed limi- tations to your analysis? (7pts)

•  Presentation:  Is your memo clearly-written, easy to follow, and compelling?   Have you eliminated spelling and grammatical errors? (5pts)

45pts: Supporting Information

•  Motivating fgure - you should have one fgure that motivates your analysis clearly by providing a visual display of the question. Remember that the caption of your fgure should explain what is meaningful, so that readers can understand even if they do not read the memo. (13pts)

•  It is said that a picture is worth a thousand words.  Have you used additional graphics/fgures to tell your story (motivation, illuminating results, diagnosing problems, etc.)? Remember, you can have up to 5 items including your regression table. (10pts)

•  Have you included a clearly-labeled regression table with all of your results? Have you summarized the table clearly with a few sentences-long caption, so that a casual reader can understand what the table shows without having to read the memo?  Remember that variable names are ofen meaningless to a reader, and you can ofen summarize grouping variables and controls for clarity. (22pts)

Remember that presentation matters for each of these items: Are your graphics presented professionally; is your table easy to read? Would we put it on the wall here at GPS?

10pts: Do File

•  Does your .do fle run without error? (5pts)

•  Does your .do fle run and replicate everything in your report?  Have you commented it so we can navigate it easily? (5pts)

4.    Rules

Submission of your project indicates that you understand and attest to the following:

1.  I have worked independently on this project and have not discussed or shared information about this assignment with anyone else.

2.  I understand that my fnal report will be screened for plagiarism and unfair collaboration via tur- nitin.com.

3.  I have not used anyone else’s materials as guidelines, including past and present students at GPS.

4.  I understand that I am not allowed to ask the TAs/professor for any help on this project beyond clari- fcation of data set structure and/or language translation questions.

5.  I have not asked other faculty for guidance on this analysis.

6.  I have fully and correctly cited all external sources that have contributed to my thoughts, development, and presentation of this analysis.

7.  I understand that failure to abide by any of these rules will result in automatic failure of both the project and the course.