top of page

Our Findings

& Conclusion

Finding 1a:  5-Number Summary Analysis

First, we found out that the number of confirmed cases at the time of closure is associated with the economic level of the nation. The higher the economic level of the nation, the larger the number of cases at times of closure. This suggests that the richer countries waited for more cases to appear when closing schools. This answers our first research question--if and how has the economic status of the country been affected when its schools were closed.

Five-Number Summary

This conclusion is found through the analysis on the 5-Number Summary. In the 5-Number Summary, the variables from our dataset we chose to examine were: (a) the country-wise data for the number of confirmed COVID-19 cases at the time of school closure, and (b) the respective income groups - lower-income, lower-middle-income, upper-middle-income and high income (based on the definition given by the World Bank) -  of these countries. In examining these two variables, we hoped to uncover the relationship between a country’s economic status and its coronavirus-related school closure dates, specifically through calculating the sample minimum, the lower quartile, the median, the upper quartile, and the sample maximum. 

While we hypothesized that the high-income countries, with their better public health knowledge, systems, and resources relative to lower-income countries, would have been quicker to identify the threat of the virus and shut down schools, the five-number summary analysis disproved this hypothesis, revealing that schools in lower-income countries took the COVID-19 pandemic more seriously in this regard. We found that the mean number of confirmed cases at the time of school closure and quartile numbers of confirmed cases are positively correlated with the economic status of the nations. From the Five-Number Summary table, we can see that the mean number of confirmed cases at the time of school closure was just 1 in low-income countries, increasing to 8 in upper-middle-income countries, and going all the way up to 17.5 in high countries. We also created a bar chart by using the medium number of cases to compare between different income groups. From the bar chart, it directly shows that the richer countries closed school when there were more cases compared to poorer countries. Both result indicate that lower-income countries closed schools when there were only a few confirmed cases, and higher-income countries waited until more confirmed cases emerged.

Hence, earlier and more severe public health policy is more economic and efficient for them. For richer countries, this scenario may be also due to the fact that richer countries take more consideration into the economics impact of the lock-down, worrying the potential harm to the economy (Gopinath). In addition, democracy is more common in wealthy countries (Londregan and Poole, 49). hence richer countries tend to take longer time to create and implement policies responding to the pandemic, which explains why richer countries closed school later.

​Bar Chart：Medium Number of Confirmed Cases at the Time of Schools Closure Between Different Income Group

392.3

145.3

13.5

4.5

Overall, based on our five-number summary, we can conclude that richer countries, that is, upper-middle and high-income nations, closed schools when there were more confirmed cases while lower-income countries reacted much earlier. The possible explanation for this scenario is that lower-income countries have less medical resources and capacity to combat COVID-19.

Note: the unit is confirmed cases

This finding is further supported by the results of an ANOVA Test. We picked our outcome variable “Confirmed Cases at Time of Closure of Schools” and the categorical variable variable “Income Group” to test if the number of confirmed cases at the time of closure is affected by the economic level of the nation. Since we need to compare more than two groups, we chose to use the ANOVA Test instead of the T-Test. Our null hypothesis is that the economic level is not associated with the number of cases at the time of school closures

The result shows that the F value is much larger than 1 and the P-value is much smaller than the Alpha (0.05). Therefore, the hypothesis that the economic level does not associate with the number of cases at the time closure of schools is rejected, and the result is also statistically significant. In addition, the F value is significantly greater than 1. This implies that the mean squares between groups is greater than the mean squares within groups, which provides additional evidence for a correlation between the number of confirmed cases at the time of closure and the economic level of the nation.

ANOVA Test

Finding 1b: Exploring the results of Finding 1 with an ANOVA test

​Finding 2 

Our second find is the positive correlation between the income group and the proportion of countries in that income group that delivered education online. We chose to analyze the relation between online education delivery and income group as it relates to one of our research questions regarding the relationship between countries’ income status and the extent to which they used different modes of education following COVID-19 triggered school closures. This correlation is shown through the contingency table in the right and it suggests that online education, which requires relatively fast and reliable internet access and internet-connected devices by a majority of the educator and student populations, is a mode of education that is more feasible and achievable for countries with higher incomes, which is likely facilitated by their relatively more developed internet and technology infrastructures. Lower-income countries, meanwhile, may not have as reliable or widespread internet connectivity and/or distribution of internet-connected devices, making the diversion of educational resources to delivering education online less practical and thus less suitable an option. 

Contingency Table：Percentages of countries that delivered online education in different economic groups.

Source: The Center for Global Development Dataset

Note: Numbers mean percentage, e.g., 95.7 is 95.7%.

Finding 3

Third, we use Chi-Squared Test to determine whether the categorical variables (number of confirmed cases at the time of closure and the economic level of the nation) are independent or related. Through Chi-Squared Test, we found out that the observed outcomes are not likely due to chance and the variables are associated with each other. This shows that our findings are meaningful as they are not coincidental. It exemplifies the reliability of our data. 

Chi-Squared Test

In conclusion, based on the dataset from the Center for Global Development and through using statistical methods such as the Five-Number Summary analysis, Chi-Squared Test, and ANOVA Test, our research shows that the higher the economic level of the nation, the larger the number of cases at times of closure. Moreover, there is a positive correlation between the income group and the proportion of countries in that income group that delivered education online. We also found out that the observed outcomes are not likely due to chance and the variables are associated with each other. Our research is significant as it demonstrates the fact that higher-income countries have more resistance against the global pandemic as shown by the delay in the closing of schools and the more diversified online education methods. For further research, it would be meaningful to investigate the change of forms of education across schools. We would want to investigate whether covid has increased the practice of online education. It would be interesting to investigate the impact that covid has on education methods.

Conclusion

​Dataset Download

References

​Dataset

Center for Global Development. “COVID-19 Education Policy Tracker.” Washington, DC, 25 Mar. 2020.

External Sources​

Acknowledgements

We would like to thank Professor Sanders for guiding and supporting us throughout this project, and for providing helpful feedback and tools necessary to complete it.

bottom of page