2026-10-23

Data

  • The data used can be found here on the Our World in Data website.
  • The data covers global vaccination coverage for one-year-olds from 1980-2025.
  • To test this hypothesis, the DTP vaccine will be studied for each country that has at least 5 years of data after the breakout of the pandemic.
    • The report states DTP as the benchmark for determining how well countries deliver routine vaccinations.

Visualizing The Data Set


- To balance the advances in medicine over the years, an equivalent period of time before and after the pandemic will be studied for comparison.

Hypothesis

  • Has child DTP3 vaccine coverage changed following the COVID-19 Pandemic?
  • Null hypothesis: \(H_0: \mu_D = 0\)
  • Alternative hypothesis: \(H_1: \mu_D \neq 0\)
    • \(\mu_D\) is the mean of the difference of each country’s mean coverage after and before 2020
  • Significance level: \(\alpha = 0.01\)

Setup

  • To test this hypothesis, a paired T-test is being performed to compare the mean coverage before and after 2020
    • Only countries with 5 years of data before and after 2020 will be tested to ensure there is enough data to test and compare.
    • The year 2020 is excluded from both samples as that is the condition for each country.
  • To perform the T-test, the country’s DTP3 coverage will be computed for both sets of years.
    • This allows for the mean difference \(D\) to be computed.
  • Next, the confidence interval will be calculated.
    • CI = [\(\bar{d} - t_{\frac{\alpha}{2},n-1}\frac{S_D}{\sqrt{n}}\), \(\bar{d} + t_{\frac{\alpha}{2},n-1}\frac{S_D}{\sqrt{n}}\)]
    • If the interval contains 0, then there is not enough evidence that the mean DTP3 coverage before and after 2020 are different. If it does not contain 0, there is sufficient evidence that they are different.

Computing Values

difference <- mutate(before, 
                     After = after$After, 
                     difference = after$After - Before)

t_test <- t.test(
  difference$After, difference$Before, 
  paired = T, conf.level = 0.99)
  • Test statistic (\(t_0\)): \(\frac{\bar{d} - \Delta_0}{\frac{S_D}{\sqrt{n}}}\) = -4.34
  • Confidence interval: \(\bar{d} \pm t_{\frac{\alpha}{2},n-1}\frac{S_D}{\sqrt{n}}\) = [-3.34, -0.84]
  • p-value: 0.00002218319
  • Sample mean difference (\(D\)) = -2.0882

Test Results

  • This graph shows the avg for both ranges before and after 2020.
    • The linear line has a slope of 1 and intercepts with the origin.
    • The values above the line have a positive difference while those below have a negative one.

Differences For Each Country


- The distribution is roughly bell-shaped with some outliers.

Conclusion

  • Reject the null hypothesis \(H_0: \mu_D = 0\).
    • Confidence interval [-3.34, -0.84] does not contain 0.
    • p-value 0.00002218319 < \(\alpha\) (0.01)
  • There is sufficient evidence that the mean coverage for the DTP3 vaccine has changed between 2015-2019 and 2021-2025.
    • Since the CI contains the true mean with 99% confidence, the results show that the true mean difference is negative, indicating the coverage of the vaccine has fallen following the pandemic in 2020.

Citations

  • WHO & UNICEF (2026); UN, World Population Prospects (2024) – processed by Our World in Data. “Share of one-year-olds who have had three doses of the hepatitis B vaccine – WHO/UNICEF” [dataset]. WHO & UNICEF, “WHO Immunization Data – Vaccination coverage”; United Nations, “World Population Prospects”; United Nations, “World Population Prospects – Interim Update” [original data].