Written Assignment 4
Submit your answer through Canvas.
Researchers as well as students are often interested in the relationship between two variables. Are economically unequal countries less likely to be democratic? Do people with higher incomes or education participate more in politics? Do women experience more discrimination in the labor market than do men? In this homework you will engage with such questions; questions of the general form “is X related to Y?” The idea is that you should start getting accustomed to interpreting results from bivariate regression analysis and learn to perform and interpret such analyses using Stata.
1. The relationship between democracy and the provision of public goods has long been of interest to social scientists. One common hypothesis is that we should expect nondemocratic rule to be accompanied by low public good provision. We are going to use some data from various international databases (The World Bank and the WHO) to explore this issue using bivariate regression. The independent variable Democracy is measured on a scale that ranges from 0-10, where 0 is least democratic and 10 most democratic. There are four indicators of public goods provision that are each used as dependent variables. Education expenditures measures total government expenditure on education expressed as a percentage of total general government expenditure on all sectors. Access to electricity measures the percentage of the population that has access to electricity. Basic sanitation measures the percentage of the population using safely managed sanitation services (toilets and latrines). Finally, CO2 emissions is measured in metric tons per capita. Table 1 contains descriptive statistics for these variables.
Table 1: Descriptive statistics
| Variable | Observations | Mean | Sd | Min | Max |
|---|---|---|---|---|---|
| Democracy | 194 | 6.742 | 3.072 | 0 | 10 |
| Education expenditures | 138 | 4.639 | 1.723 | 1.371 | 12.459 |
| Access to electricity | 192 | 81.530 | 28.501 | 7.252 | 100 |
| Basic sanitation | 192 | 73.708 | 29.284 | 7 | 100 |
| CO2 emissions | 189 | 4.507 | 6.099 | 0.044 | 45.423 |
Table 2 shows the results from four bivariate regression analyses. Each column shows the results from one of these. As explained above, they all have the same independent variable (Democracy) but the dependent variable varies. So, for example, in the first column the dependent variable is Education expenditures.
Question: Is the hypothesis about democracy and public goods supported by the analysis? Specifically: (1) How should the regression coefficients (b coefficients) for democracy be interpreted? (2) Are the relationships between democracy and the various indicators of public goods provision statistically different from 0? Motivate! (3) What conclusion can be drawn about the relationship between democracy and public goods provision based on the answers to (1) and (2)? (4) Comment on the measures of goodness of fit!
Table 2: Democracy and Public Goods Provision
| (1) Education expenditures |
(2) Access to electricity |
(3) Basic sanitation |
(4) CO2 emissions |
|
|---|---|---|---|---|
| Democracy | 0.237 | 2.825 | 2.624 | -0.123 |
| (0.0469) | (0.648) | (0.665) | (0.147) | |
| Constant | 2.961 | 62.43 | 56.10 | 5.342 |
| (0.359) | (4.802) | (4.903) | (1.091) | |
| Observations | 138 | 192 | 192 | 189 |
| R-squared | 0.158 | 0.091 | 0.076 | 0.004 |
| Root MSE | 1.587 | 27.24 | 28.23 | 6.104 |
Note: Standard errors are in parentheses.
2. Lately, there has been considerable efforts by social scientists to explain the rise of support for populist radical right parties in Europe and North America. In the beginning, the literature focused on explanatory factors such as immigration, but also on political institutions such as the electoral system. Recently, many scholars have claimed that such factors are insufficient to understand the success of radical right parties. Instead, radical right populism should, according to these scholars, be seen as a backlash against the globalization and urbanization that has characterized recent decades in Western Europe and North America.
Scholars in this more recent literature have therefore hypothesized that voter support for radical right populists will be higher in geographic areas that are experiencing economic decline and outmigration. Relatedly, some have argued that globalization and urbanization has led certain areas to become gender imbalanced. Since women are less likely to vote for populist radical right, it can be hypothesized that these types of parties will have lower support in areas with a larger share of women.
In this question, you are going to explore some of these more recent explanations for populist radical right party success by applying regression analysis to the data on Swedish municipalities (data_wa4.dta). The dependent variable is going to be the Sweden Democrats’ vote share (in percent) in the national election for each municipality (share_sd). You are then going to study the impact of the following independent variables on the Sweden Democrats’ vote share: 1. the total number of outmigrants as a share (in percent) of the total population (share_outmigration), 2. the share (in percent) of women in the total population (share_women) and 3. the share (in percent) of employed in the total population (share_employed).
As a first step, you need to create share_outmigration, share_women and share_employed. In the data, you only have the total number of outmigrants, women and employed per municipality. However, municipalities are of very different sizes. Stockholm has a population of nearly one million but some other municipalities have just a few thousand inhabitants. So, to be able to compare municipalities in your analysis you have to standardize the variables by population size.
To create the shares, you can use the following commands in Stata:
generate share_outmigration = 100*outmigrants/population
generate share_women = 100*women/population
generate share_employed = 100*employed/populationa) As a first step, make a table containing the mean, standard deviation and range for the four variables share_sd, share_outmigration, share_women and share_employed.
b) Create three scatterplots with regression lines, one for each independent variable, to illustrate their relationship with share_sd. Remember to always have the dependent variable (share_sd) on the y-axis. Save them and include them in your submission.
c) Run three bivariate regressions with share_sd as the dependent variable, one for each of the three independent variables. Present the results in one (1) table.
d) Interpret the regression coefficients and perform statistical significance testing. Also comment on goodness of fit. How do you think the hypotheses developed by scholars in this more recent literature on populist radical right voting fare when confronted with your Swedish data? Discuss!