Project Guidelines

Statistical Methods for Economics

2022/2023

1. Introduction and Submission Instructions

2. Stata

3. Data

4. Project Instructions

1.  Introduction and Submission Instructions

The purpose of this project is for students to apply some of the statistical techniques introduced in the course, to acquire a basic understanding of the statistical software Stata, and to learn how to search for appropriate academic economics articles.

This project will require you to learn some Stata techniques on your own, and it might be challenging to get started.   For this reason, you should attempt the project reasonably well in advance so you can seek out extra help if needed.  You can post your questions to Piazza.

Some of the required techniques to finish this project will be introduced in the last few weeks of the course.  You can finish Section A now, but to finish Section B, you will need to be familiar with the material covered in sections 11.1 to 11.5 of the required textbook. Section C requires you to summarise an academic article, and it might be useful to refer to Chapters 11 and 12.

The project can be completed individually, which is the default option.  However, to give you a bit more flexibility, we have decided to allow you to complete the project in groups of two or three if you would instead prefer.

•   If you would like to complete your project in a group of two or three, then please join a Project Group using the sign-up portal on Learn > “Assessment” section > “Projects” folder > “Main Project” folder (i.e., the same location as this instructions file). You must join a group by 12:00 noon on Tuesday 14th March.

•   If you do not join a group by the above deadline, then you must complete the project  individually.  If you  submit  your  project  individually,  then  it  must  be original and it must be only your work. Please refer to the University’s guidelines onplagiarism.

There  will  be  two  submission  portals,  one  for  students  who  complete  the  project individually and one for students who complete the project in a group of two or three.

•   For students completing the project as a group,  ONLY ONE MEMBER needs to submit  the  project  on  Learn.  Please  state  the  Group  Number  and  the  Exam Numbers (the B-numbers) of ALL members on the first page of the submission.

The projects are to be graded anonymously, so please do not put your name on the actual assignment.

The  project  is  due  Tuesday  28th  March  at  10AM.    The  project  will  be  submitted electronically through Learn.  Refer to the programme handbook for information on late submissions and penalties.  The project should be typed and must be between 1000 and 1400 words.   This word count excludes the bibliography (which you should include), tables, figures, and equations.  A 5 percentage point penalty will be applied if the world limit is exceeded.  This word limit might be binding for some of you, in which case you need to be  concise and prioritise content.   A bibliography should be included which contains  references  to  all  sources  used   (including  data  sources).     Any   standard referencing system, such as Chicago or MLA, is appropriate.

Projects will be graded according to the University’s Common Marking Scheme.  This can be found in the Programme Handbook.   Each project will be awarded a single grade. Markers will ask the following questions when assigning a mark  (roughly in order of importance):

•   Have all elements of the assignment been adequately and correctly addressed?

•   Does it appear that students have put forward a reasonable amount of time and effort?

•   Are the statistical techniques appropriately carried out?

•   Is the chosen article appropriate, and is a suitable understanding demonstrated given  the   statistical/economics   knowledge  that   is   expected   of  a   2nd    year economics student?

•   Are all figures clear, appropriately labelled, and informative?

•   Are explanations convincing, well-informed, and supported by evidence where appropriate?

•   Has independent learning and insight been demonstrated?

•   Are sources clearly cited?

In line with  University  policy,  feedback will be  provided within  15 working days  of submission.  Feedback will be provided to each student’s University of Edinburgh email address.

2.  Stata

There are three sections of this project.  Section B must be completed through the use of the statistical software Stata.  Please refer to the separate document “Stata for Statistical Methods for Economics.” This document provides instructions on how to install Stata, as well as a brief overview of how you might interact with Stata.

The Stata techniques that you will need to use for this project are as follows:

1.   Opening Excel Files with Stata

•   There  are  two  ways  to  get  data  from  Excel  into  Stata.    You  can copy/paste from Excel  (Copy/Paste Instructions) or you can import your data directly (Import from Excel Instructions).

•   To transfer data from Excel to Stata, it is important to have your data laid out properly.   The columns should represent different variables and the rows should represent different countries.  The first row can include the names of the variables.

•   Note that to view your data, you can simply type “browse” (or “edit” if you want to make changes) in the Command window.

•   Type “help import” in the command window for more information.

2.   Calculate descriptive statistics:

3.  Data

For this project, you will be downloading data from the World Development Indicators Database.  You can access it through this link:WDI Databank.     This  resource   contains data on about 1,500 different “development indicators,” where a development indicator is some measure of a country’s overall development and performance.

On the WDI webpage, the page is split into two panels.  The left-hand side panel is where you can select your data or change the layout of your data. The right-hand side panel is where you can preview the data that you have selected.

Before  doing anything,  click “clear  selection”  as a set of countries or  indicators  may already be chosen.  On the left panel, you can tick the countries, the series, and the year that you are interested in.  Be sure to click the “Apply Changes” button after making your selection.

As is discussed in the next section, you can choose any indicator from this database that you  are  interested  in.    Note,  however,  that  for  many  indicators  and/or  years,  data availability is poor and information is only available for a handful of countries or years. Very recent data (e.g. for 2022) is less likely to be available.  It is not a problem for you to use data from earlier years as availability dictates.

For this project, you will be considering all of the countries listed in the database. You need to make sure that you are selecting countries, and not aggregates of countries (e.g. you want to include Tunisia rather than “Arab World”). To do this, under the country tab:

It is important that your dataset is laid out properly before downloading it.  Otherwise, you won’t be able to carry out the necessary tasks in Stata.  There is a “Layout” tab on the left side of the page.  For the purposes of getting the data into Stata, it is easiest to select the exact layout indicated here:

After you have made your selection, you can download your data in Excel format.  To do this, click the “Download Options” tab on the top right-hand side of the page. You should click “Advanced Options,” and then select “Blank” for “NA preference.”  If you do not do this, then Stata will not treat the data as numerical.

4.  Project Instructions

This project has three sections.  It is up to you how you want to divide your word count between the three sections, withinreason.

Using the World Development Indicators website, construct a dataset which contains information on GDP per capita as well as information on one additional indicator of your choosing.  This dataset should include every possible country, and it should be for one year (any year that you choose). For example, your dataset might contain information on GDP per capita and on infant mortality rate in the year 2016.  There are a few different ways to measure GDP per capita (PPP, local currency units, etc…)  You are free to choose any of the GDP per capita measures that you would like.

Note that for nearly every indicator, the relevant data will be missing for some countries. Make sure you pick an indicator and a year where there are at least 30 countries in your dataset with all of the necessary data.

Section A

For section A, describe the indicator that you have chosen. In this section, include the following information:

•   Provide a clear definition of this development indicator and what it measures.

•   Who is typically responsible for collecting this data and how is this data collected? Are there any challenges or measurement issues associated with collecting this data?

•   Why is this indicator of interest to policymakers?

•   What factors might determine how countries perform according to this indicator?

Note:thissectionisveryopen-ended, andhowyouapproachitdependsontheparticular indicatorthatyouhavechosen. Thebeststartingpointistoclickthesymbolnexttoeach indicatorthatlookslikethis: .   Thispullsupthemetadataforthatindicator. However, youshouldnotjustcopyfromthissourceandyoushouldtrytoincorporateothersources. Youcaneitherrefertotheoriginaldatasources,  orstartwithafewkeywordsearchesin Google.

Section B

You must use Stata to complete this section.  Note that Stata will automatically ignore all countries with missing values.  This section should include the following:

1.   What  is  the  mean,  median,  standard  deviation,  minimum,  maximum,  and interquartile range of the indicator that you have chosen?

2.   Which country (or countries) take on the minimum value and which country (or countries) take on the maximum value for your chosen indicator?  Provide a brief explanation as to why these countries have such different outcomes.

Hint:type “sortvariable_name”  (wherevariable_name is the name ofyour variable) in the Command window and then type “browse” to identify the minimumandmaximum.

3.   Provide a histogram showing the distribution of the indicator that you have chosen, and briefly comment the shape of the distribution.

4.   Provide a scatterplot which shows the relationship between the indicator you have  chosen  (on  the vertical  axis)  and  the  log  of GDP  per  capita  (on  the horizontal axis).  Comment on the relationship, and calculate the correlation between your chosen indicator and log GDP per capita.  Briefly speculate as to why this relationship exists.  If the two variables appear to be independent, is that expected?

5.   Run a regression where your chosen indicator is the dependent variable (the Y variable) and log GDP per capita is the independent variable (the X variable). What is the estimated intercept and what is the estimated slope?  What is the r-squared  from  this   regression?     Briefly  interpret  the   estimated   slope coefficient and r-squared from this regression.  Paste in a screenshot of Stata’s regression output.

Note:Asnotedpreviouslyinthisdocument, regressionisintroducedinlecture inWeek 8.  IfyouareworkingonthisprojectinadvanceofWeek 8 (whichisa godidea),youshouldreadsections 11.1– 11.5ofthetextbook.

6.   From the regression output, manually verify (a) the t statistic for the slope, (b) the confidence interval for the slope, and (c) the regression r-squared. That is, from the information given in the regression output you need to calculate these

3 items yourself using the techniques described in the book.  Clearly show all steps in your calculations.

Note:The critical t values that are reported in the book andelsewhereare typicallyroundedtotwo-decimalplaces. Therefore, yourconfidenceintervals willbeveryslightlydifferentthanwhatisreportedbyStata. Don’tworryabout this. Notethatyoucanusethe “invttail(df,p)”functionin Stata,  alongsidethe “display”commandtofindappropriatecriticaltvalue.

Section C

For Section C, you must find any one academic article of your choosing that somehow relates (even quite loosely) to the indicator you have chosen.  For example, if you choose the Gini coefficient as your indicator, then you should select an article that addresses some aspect  of income inequality.   This article should  have  an  element of empirical work/data analysis in it.  This article must be an economics article.  To find an article, it is easiest to make use of Google Scholar. Some of you may wish to make use of the advanced search feature, which allows you to search by journal.

It is sometimes unclear whether an article is an economics article or not, but as a rough rule  of thumb you  can  check whether  the  author  is  an  academic  economist  (just by googling their name) or you can check if the article is published in an economics journal. Here is a list of (mostly) economics journals, crudely ranked: Repec Journal Rankings. Take some time to find an article that you find interesting.   If you are having difficulty

finding an appropriate article, you can email me atsteven.dieterle@ed.ac.uk. In your summary of the article, you should touch on the following points:

1.   What question does the study seek to address and why is it important?

2.   How does this study fit into the broader literature and what is the contribution?