HYPOTHESIS 1: Countries with higher mean years of schooling have higher internet usage

We chose to check for a correlation between these two variables because we were interested in seeing how higher levels of education affect the daily lives of subjects. While we aren’t certain there is a connection between the two variables (as we would have been between variables like HDI and life expectancy, since HDI takes into consideration life expectancy), we expect there to be a positive relationship between the two.

Analysis: The scatter plot of countries with mean years of schooling on the x-axis and internet users per 100 people on the y-axis reveals a positive trend. The correlation coefficient r of this linear regression is 0.81, which is a relatively strong correlation. We infer that this relationship exists due to the modern, symbiotic relationship between education and technology. Naturally, a country whose citizens undergo more schooling would similarly have citizens who are more inclined to use the internet, the cornerstone of modern day technology.

Examining internet usage by region:

## # A tibble: 5 x 4
##     region     n avg_schooling avg_internet
##     <fctr> <int>         <dbl>        <dbl>
## 1   Africa    48      4.975000     17.11396
## 2 Americas    28      8.489286     46.57571
## 3     Asia    40      8.165000     44.53850
## 4   Europe    39     11.241026     75.56333
## 5  Oceania     9      8.888889     35.83222

When the countries are grouped by region, the averages for the two variables (mean years of schooling and internet usage per 100 people) seem to support the conclusion drawn from the trendline in Figure 1. Africa, the region with the lowest mean schooling also has the lowest mean internet usage. Europe, the region with the highest mean schooling has the highest mean internet usage. We are interested in exploring the relation between the two variables for countries within each region. We don’t expect to see much from the Oceania plot because the summary table above shows Oceania only includes nine countries.

The scatter plot for countries in Africa reveals a much weaker correlation between schooling and internet usage. There seems to be a concentration of countries in the area of both low internet usage and mean years of schooling. The correlation coefficient is about 0.60 for this scatter plot. Even as years of schooling go up, only five countries break 40 internet users per 100 people. We can see a large deviation from the purported trend in Morocco, which has a mean year of schooling of 4.4, but 56.80 Internet Users per 100 People.

Here, we can see another seemingly strong corrolation for the Americas, yet in reality, the r value of this relationship is merely 0.69, largely attributed to the presence of datapoints in the lower right quadrant. Something to note is that the United States was omitted from this data, as the data set does not contain any data for its internet users per 100 people. But at the top right of the plot we can see a point at around 13 mean years of schooling, and 87 internet users per 100 people. This data point is none other than Canada, a country that is demographically, societally, and culturally most similar to the United States in the entire data set. From this, we could infer that the U.S. would show similar or even greater numbers for internet users per 100 people (they report 12.9 mean years of schooling), and would thus increase the correlation of the relationship.

With Asia, we have an especially low r value of 0.59. This exists due to large deviations from the general trend, with Turkemenista at 9.9 mean years of schooling and 12.2

The correlation coefficient for the scatter plot of European countries is 0.39. The plot does not show strong evidence for a positive correlation between internet usage and schooling within Europe.

Above, we can see a visualization of Internet Users per 100 people on its own. Some things to note: 1. Europe and Asia seem to be the most uniform of the regions. This makes sense, as its scatter plot was all over the place. 2. Oceania seems to be skewed left, and this can be attributed to its nine data points, a relatively small number that does not adhere to the Law of Large Numbers. 3. Africa has five outliers, presumably due to the great economic disparities between all the African countries. These outliers are: Morocco (56.8), Seychelles (54.28), South Africa(49), Tunisia(46.18), and Kenya(43.4). 4. The entire distribution of Internet usage per 100 people in Europe appears to lie above the maximum in Africa’s distribution.

Conclusion Education and access to internet seem to go hand in hand. However, within individual regions of the world, the relationship between the two variables is not as clear. We cannot conclude a causative relationship from these factors alone, but we can further investigate the determinants of internet and education acessibility.