路嘉锋/Jiafeng Lu
Relationship between
Education and Economic
Growth
Data Visualization to Display Government Spending on Different Levels of Education's impact to Economic Growth
Data Visualization INFO 658
Under the Guidance of Professor John Lauermann
There’s a saying: Education is the most powerful weapon which you can use to change the world. Nowadays, it takes a great deal of time and effort for a mass of countries to improve their national education level while they realize the importance of education. In this project, I conduct research on how government spending on education affects the population’s general education level, and how a citizen's education level is related with the nation’s economic growth.
In the first part of the study, I choose several developing and developed countries which are most located in Europe to gather the data and see the relationship between each country’s government spending on education and school enrollment data. Then I use data from government spending and economic growth to run the regression and conclude the relationship between them.
Government Spending Affect Education Level
I want to conduct research on how government spending on education affects the population's general education level, and how citizens' education level is related with the nation's economic growth. I would compare the government's expenditure on education based on students per capita and institutions from three perspectives: primary, secondary and tertiary education.
I believe that a stated correlation between education level and economic growth implies the importance of the quality of labor, as well as how more developed countries could have higher labor education levels. There could be motivations for the government to contribute more on education if the correlation between government spending and people's education level is proven to be positive, and vice versa.
I plan to compare trends between different developed and developing countries, including traditionally high education level countries from Nordic, fair level like the United kingdom, and developed countries such as Mexico.
Questions to be Addressed
Data Preparation Lists
Diversified Data from The WorldBank and OECD organization
1.https://data.worldbank.org/indicator/SE.XPD.TOTL.GB.ZS (Government expenditure on education (% government expenditure)
2.https://data.worldbank.org/indicator/SE.XPD.TOTL.GD.ZS (Government expenditure on education (% of GDP)
3.https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG (GDP growth (annual %)
4. https://data.oecd.org/eduresource/public-spending-on-education.htm (Public spending on education)
Missing data and outliers
There is some missing data, and some data is not up to date, thus causing trouble when we are trying to generate a trend over years. Here I plan to move on with data from 1980s to 2010s as that period is most well recorded within the database. I would have to narrow down to specific countries to focus on before I look for more data.
Shortcomings
of the Data
Softwares
& Methods and Rationsales of Desiging Choices
Software
Tableau Public, Python packages including Pandas& Matplotlib, Stata, Excel ANOVA
Selecting and Refining Data
I chose the dataset because it provides concrete and comprehensive information about government expenditure on education and school enrollment rate that I wanted to dive into. I downloaded this file as a CSV and used Excel to refine it, filtering out to expenditures for all levels (primary, secondary and tertiary) to compare them together.
At the primary stage of data cleaning, I choose the net percentage of school enrollment rate instead of the gross percentage, however, flaws are realized that the gross percentage does a better representation of combined accumulative data for representing the whole picture.
Using Tableau for the initial round
I run Tableau to gather information for changes in government expenditure on education and school enrollment rate for the three countries. And government expenditure on different levels of education(Appendix). After running the UK and Mexico, I find the visualization does quite match my purpose of comparing together at the same time, and combining them takes a longer time than Matplotlib. After comparing that with using python Matplotlib, which gives a more text-based and clear visualization, I decided to move forward with Matplotlib instead.
Furthermore, three multiple linear regressions for three countries are made to find the correlation between spending on education and each level of school enrollment rate. It was primarily implemented on Excel ANOVA but did not achieve the desired result, thus it has been remade on Python.
Moreover, a scatter plot trendline is visualized using single linear regression models to represent the above correlation for primary and secondary education perspectives for the three countries.
For rationales of choosing colors and styles of charts. Primarily I choose bar charts with line visualization because it gives a better view if the two measures match each other, or follow the same trend. While comparing each enrollment rate part, I use all trend lines because they are the same measurement. (Tableau).
While on Python, I choose strong diversified colors to represent each trendline, and for the scatter plot, I choose red and blue, the two counter colors to give the audience a clearer view of deciphering dots to lines.
Visualizting Using Tableau First
and Furthersteps
Analysis
Three different countries with different levels of Education Background and Economic standing
I choose three countries: the UK, Sweden, and Mexico. Two of them are developed countries and located in Europe, with the UK being well-known for its private education system and Sweden being well-known for its public education system. For the third one, located in North America, Mexico is a developing country. Such selection is made to represent three different education ecosystems in order to eliminate biases when comparing data. I extract data from these countries from 1970 to 2020 to find out the trend in government spending on education and their school enrollment. I have my graph plotted with three enrollment rate data, from primary education, secondary education, and tertiary education. And generally, the secondary school enrollment rate follows the increase in government expenditure most closely, while there is a weak positive correlation between government expenditure on education and primary enrollment rate, as well as tertiary enrollment rate.
Changes in Government spending and Different School Enrollment Levels for Three countries
The first graph is data from the United Kingdom, with the second from Sweden and the third from Mexico, with blue lines representing government expenditure percentage and red lines representing secondary school enrollment percentage.
From the plotting, I conclude that developed countries have been having a gradual increase of share on education from government expenditure following a higher rate of secondary school enrollment, though there is a downturn around 2010.
While for developing countries like Mexico, there has been a downfall of government expenditure on education while enrollment rate still keeps climbing, that is probably due to the citizens' rising awareness of the importance of education. Hence I could temporarily draw a conclusion that education level is positively correlated with government expenditure in developed countries and people’s education level are generally higher in these countries. While puzzles still remain in developing countries as government expenditure on education may not necessarily be a major indicator for their population's education level.
Visualizations
Using Matplotlib
To Differentiate
Government Spending on different levels of School Enrollment
MLR to determine
Enrollment
Coefficient
Using Multiple Linear Regression and Stata to compare
Spending on Education and School Enrollment for three countries
Multiple Linear Regression on Spending on Education and School Enrollment for UK
Multiple Linear Regression on Spending on Education and School Enrollment for Sweden
Multiple Linear Regression on Spending on Education and School Enrollment for Mexico
I also run a multiple linear regression on spending on education and school enrollment for all three countries. Choosing government spending on education as dependent variable and school enrollment as independent variable and finding the coefficients of them, I found that only the United Kingdom have all three enrollment coefficients positive, and there’s a negative coefficient in secondary school enrollment for Sweden and tertiary enrollment for Mexico, which illustrates that school enrollment rate does not necessary impact government expenditure on education in a positive direction, thus it may not be legit to determine the relationship between the two factors and I could state that only primary school enrollment have a positive relationship with government expenditure on education for the three countries.
Scattered Plot
UK
Sweden
Mexico
While on the reverse side, when I take government spending as the independent variable and school enrollment rate as the dependent variable and run a simple linear regression, I find that, especially for secondary school enrollment (as a student must finish his/her primary education to get to secondary while tertiary education may not be an option for everyone), there is a strong positive relationship between them for all three countries, and for developed countries, the UK and Sweden, most dots are on the top right-hand side corner and that illustrates both developed countries’ spending on secondary education leads higher secondary school enrollment rates. For Mexico, though more dots dropped on the left side of the graph, the graph trend still tells me a weak positive relationship. Furthermore, as all three plots for primary enrollment rate illustrate a weak positive relationship between the two factors. At this stage, I could conclude that at least an increase in government spendingon education will result in a climb in primary and secondary school enrollment.
However, the puzzle still remains as I am not able to determine whether government spending on education influences the school enrollment rate or whether the school enrollment rate influences government spending.
For school enrollment rate and economic growth. I have taken a close look between Gross GDP growth and school enrollment rate from the World Bank database. There is no obvious relationship between them as the three countries generally follow the trend of Global GDP growth while school enrollment rates of primary and secondary for the three countries diversify. There may be several other reasons. As economic growth has plenty of determining factors. It’s undeniable that school enrollment rate, or education level, is an important factor. However, there may occur some other factors in the country which may offset the effort by the education sector. Thus, even if there is a high school enrollment rate, the economic growth still remains low or doesn’t go with the direction of the school enrollment rate. Moreover, though educated workers can be more efficient during work, they may go outside their own countries to work or choose to migrate to countries that would provide them with a better standard of living, which results in economic growth outside of their countries. When a large group of the labor force who have received the education benefits leave the country after a while, there is going to be a brain drain period occurring, which means economic growth will stay trapped as well. Furthermore, the benefits of education on economic growth take time to reveal. As the government spends on education may take one to two decades to see the benefit of this ‘education investment’. I am not able to draw a conclusion about the relationship between education and economic growth as there are many other factors impacting and constraining the result.
Reverse Side
Conclusions and Directions
Overall, I could conclude that increase in government spending could boost the primary school and secondary school enrollment rate while a climb in school enrollment rate does not necessarily mean the country’s economy will grow significantly regarding factors such as time and immigration, no matter in developing or developed countries.
For future research, I want to dig into how education could affect the country's economic growth by eliminating the effect of other factors, including time, immigration, global trend, broad environment, and so on. Besides, I could take brain drain into account. I can gather more detailed data including the labor force’s educational background so I can analyze the data more accurately.
Graph of Government expenditure on different education levels
