Introduction

As of August 2024, the pandemic caused by the novel coronavirus, SARS-CoV-2 caused more than 7 million deaths worldwide1 along more than three years since the detection of the first case in China2. However, its impact varied considerably depending on the socioeconomic status (SES) of the affected population. Several studies have reported associations between SES and COVID-19 incidence, mortality, and vaccination coverage3,4,5. Locally, a seminal study showed that during the first stages of the pandemic, mortality attributed to COVID-19 was higher in places with lower SES in the Metropolitan Area of Santiago, Chile3. Data exhibited by this study corroborated that COVID-19 is, in fact, a syndemic disease: An epidemiological condition whose burden among the population is synergistic with both non-communicable diseases and SES6,7. Thus, wealthier municipalities—usually more affluent and therefore healthier—were found to be less vulnerable to COVID-19. A similar correlation has been reported for other countries8. For instance, in the USA, localities with lower levels of education and a higher proportion of African American population–both factors usually associated but not equivalent to lower SES–are linked with a higher number of COVID-19 cases and fatalities, together with a higher proportion of long-term consequences of the infection9,10,11. Similarly, in Sweden, a higher number of COVID-19-related deaths, occurred in areas of lower SES12, while in India, population density and literacy have been found to be positively and negatively correlated with COVID-19 infection rates, respectively, highlighting the syndemic nature of this disease13.

In an unprecedented biotechnological response, vaccine manufacturers developed, and National Regulatory Authorities approved in record time, several vaccines that were proven safe and effective in diminishing infection, hospital admissions, and deaths rates14,15,16. However, unequal access to vaccines creates enormous risks, prolonging the pandemic, increasing the likelihood of new variants17, and exacerbating existing health, economic, and social inequalities18,19,20. Moreover17, due to the influence of anti-vaxxer movements driven by the abundance of fake news, anti-science, and the promotion of highly individualistic behaviors21,22, assessing and comprehending the willingness to receive COVID-19 vaccination is essential for the development of more effective public policies aimed at promoting vaccination23,24,25, while keeping a balance with strict restrictions and possible socioeconomic impact26.

In addition to the correlation between socioeconomic status (SES) and both the incidence and severity of COVID-19, vaccination rates have also been linked to SES at the country level. For example, some countries, such as Israel and Chile, accessed vaccines early in 2021, achieving coverage rates exceeding 90% of their populations27,28. In contrast, less affluent nations like Haiti struggled to vaccinate even 5% of their population29. Within countries, a lower proportion of the population receiving the COVID-19 vaccine has been associated with lower SES, where the main factors mediating this relation are the perception of vaccine safety, the influence of anti-vaccine movements, and the effectiveness of vaccination campaigns, as families with lower SES may have less access to accurate sources of information, greater exposure to misinformation, and face more logistical challenges in accessing vaccination services4,23,24,25,30.

In the case of Chile, in addition to its successful COVID-19 vaccination campaign aimed at the general population31, the Chilean government conducted an extensive vaccination program for primary and secondary students. By October 2022, over 90% of children and adolescents aged 6 to 17 were fully vaccinated, having received at least two doses of the vaccine32. Following the government’s authorization of COVID-19 vaccines for very young children, the Ministry of Education developed a ‘safe back-to-school plan’. This plan included the lifting of seating capacity limitations in classrooms where vaccination rates exceeded 80% 33. Notably, even before the completion of the vaccination campaign for children under 12 years old was achieved, class attendance restrictions were lifted, eroding the availability of incentives for children to be vaccinated. In contrast, all individuals above 12 years old were compelled by the Ministry of Health’s to have a Sanitary Pass certifying that the person was up to date with the vaccination timetable before attending cultural, shopping, and eating venues34.

While the level of vaccination coverage achieved in adults was very high across the whole population regardless of their SES according to the published data31, the case for children and adolescents has not been analyzed. Although the level of vaccination for each school was closely monitored by the Chilean government, to the best of our knowledge, the relationship between SES, vaccination coverage and the nature of the school, was not assessed. This study aims to evaluate the correlation between SES and COVID-19 vaccination levels in primary and secondary students, in Santiago, the capital of Chile through a longitudinal observational study using publicly available data. Understanding this relationship is crucial for developing effective public health strategies that address disparities in vaccine coverage.

Results

In this study, we determined the SES of the 32 municipalities located in the Metropolitan Area of Santiago, the capital of Chile. For each municipality we computed the median school vaccination coverage at three different dates distributed along the pandemic so to cover the spread of SARS-CoV-2 variants Delta and Omicron BA.4 and BA.5: November 15, 2021; March 1st, 2022; and May 26, 2022, respectively. Our results indicate a strong correlation between SES indices and the median school vaccination coverage. We further explored differences between school types, evaluating private, subsidized, and public schools’ data separately. As expected, vulnerable municipalities with low SES exhibit lower levels of vaccination coverage. Surprisingly, while a strong correlation between vaccination coverage and SES is present in both public and state-subsidized schools, the correlation is meaningless for private schools. Therefore, in the latter, vaccination coverage seems to be independent of the SES of the municipality.

Exploring the correlation between SES indices: All indices are correlated with social priority index

Since the SPI index used in previous studies is only available for some municipalities, we evaluated the correlation between other SES indices and the SPI to determine whether other SES indices may be used as proxies to access the socioeconomic status of all municipalities in the Metropolitan Area of Santiago, Chile. Our results indicate that all the evaluated SES indices do correlate with SPI (Supplementary Fig. 1), with CDI being the one with the highest correlation strength (|ρ| >= 0.90).

Vaccination coverage is correlated with SES in schools belonging to Santiago, Chile

In this study we evaluated the vaccination and SES data of students enrolled in the 1,667 schools belonging to the 34 municipalities of the Metropolitan Area of Santiago, Chile. We divided the population according to the school type: Public, state-subsidized, and private schools. Their enrollments as of March 2021 were: 297,928 (29%), 575,426 (55%), 169,898 (16%), respectively. The population estimate of Santiago is 6,075,403 (30.9% of the country’s population). As in any other large city around the world, SES indices vary widely. For instance, the municipality of Providencia (Fig. 1A; CDI rank 1) has the lowest rate of Multidimensional Poverty (0.034), whilst Lo Espejo (Fig. 1A; CDI rank 33), has the highest rate of Multidimensional Poverty (0.375): More than 10 times higher than that of Providencia. As expected, the three municipalities with the highest CDI are also the ones with the highest level of vaccination coverage in schools (Table 1; Fig. 1A). In contrast, the four municipalities with the lowest CDI are those with the lowest vaccination coverage, except for Pedro Aguirre Cerda, a municipality exhibiting lower vaccination coverage and ranking 26th in CDI (Table 1; Fig. 1A). Notably, our study also shows that there are no municipalities with high vaccination coverage and low or medium CDI. Conversely, no municipalities with low vaccination coverage and medium or high CDI can be identified (Table 1; Fig. 1A).

Table 1 SES indices and vaccination coverage for each municipality.
Full size table

When analyzing the data gathered on May 26, 2022, three months after the peak of the Omicron spreading (around February 14), the correlation between each SES index and the median school vaccination for each municipality ranged from a very strong correlation, with a |ρ| >= 0.90 for the case of Economy and Resources, Education, CDI, and SPI indices; to a strong correlation, with a |ρ| >= 0.70 for the case of Multidimensional Poverty, and SVI; and to a moderate correlation, with |ρ| > 0.40 for the case of Poverty, PHLBS and Health and Social Welfare. All correlations were statistically significant (p < 0.01) (Fig. 1B). Of note, similar trends with different correlation values between SES indices and vaccination coverage were found on November 15, 2021, three days after the peak of the Delta variant spreading, and on March 5, 2022, in the middle of the second wave of Omicron infection produced by the dispersion of the BA.4/BA.5 subvariant (Supplementary Figs. 2 and 3).

We further explored whether the correlation between the vulnerability of different municipalities and the vaccination coverage may be accounted for in the three main types of schools existing in Chile: Private, state-subsidized, and public schools. While parents and tutors enrolling their school-aged children in private schools must pay a full tuition fee, in the case of state-subsidized schools, the tuition fee is importantly reduced by the state subsidy. In the case of public schools, the state covers the full tuition fee. Our analysis reveals that a similar correlation, as previously mentioned, exists within both state-subsidized and public schools, albeit with variations in the correlation values (Fig. 2A-D). Unexpectedly, our analysis reveals no observable correlation between the vulnerability of municipalities and vaccination coverage within the population attending private schools. (Fig. 2E-F).

Progression of inequity in school vaccination coverage over time

With the aim to assess the inequity of the vaccination coverage, we calculated the Gini coefficient in three dates, covering the spread of different SARS-CoV-2 variants. We found that overall, inequity decreases in all school types over time along the COVID-19 pandemic (Fig. 3, upper panels). Importantly, the largest reduction in the inequity of the COVID-19 vaccination coverage occurs between November 15, 2021, and March 01, 2022, after the peak of the Omicron variant dispersion. In all time periods analyzed, inequity in vaccination coverage persists at higher levels in public and state-subsidized schools compared to private schools. This occurs despite the existence of a significant correlation between the vulnerability of municipalities and vaccination coverage, indicating that lower SES is associated with reduced vaccination rates. Notably, this disparity in vaccination coverage is more pronounced in public and state-subsidized schools. In contrast, private schools demonstrate greater equity in vaccination coverage, regardless of the SES of the municipality in which they are situated. To analyze the data from another perspective, we separated the data according to grade and found that inequity in vaccination is greater in lower grades (Fig. 3, lower panels), and this same pattern is found in the whole country, but with even more marked differences in the younger groups. Again, inequity decreases over time, but the age-dependent pattern remains.

Fig. 1
figure 1

Correlation between school vaccination coverage and SES Indices in Santiago, Chile as of May 26, 2022. Panel A), municipalities are colored according to their school vaccination coverage and their Community Development Index (CDI) in a Bivariate Map. Municipalities are numbered according to their rank in CDI from higher to lower values. The color coding is shown adjacent to the map using a 3 by 3 matrix where the CDI increases from left to right and the vaccination coverage increases from bottom to top. The values are grouped into three levels. The number labeled in each municipality is sorted by decreasing ranking according to the CDI and can be seen in Table 1. (B) The correlation value between six SES Indices and the median school vaccination coverage was evaluated in the 34 municipalities of the Metropolitan Area of Santiago, Chile. ρ = Spearman’s rank correlation coefficient; p = p-value.

Full size image
Fig. 2
figure 2

Correlation between school vaccination coverage and SES Indices by school category in the Metropolitan Area of Santiago as of May 26, 2022. All maps represent Municipalities from the Metropolitan Area of Santiago. They appear colored according to their school vaccination coverage and their Community Development Index (CDI) in a Bivariate Map. Municipalities are numbered according to their rank in CDI from higher to lower values. The color coding is shown adjacent to the map using a 3 by 3 matrix, where the CDI increases from left to right and the vaccination coverage increases from bottom to top. The values are grouped into three levels. The number labeled in each municipality is sorted following a decreasing ranking according to the CDI, as can be seen in Table 1. Next to each map, the correlation of six SES indices and the median school vaccination coverage is show. ρ = Spearman’s rank correlation coefficient; p = p-value. (AB) Public schools’ data. (CD) Private subsidized schools’ data. (EF) Private schools’ data.

Full size image
Fig. 3
figure 3

Inequity assessment of COVID-19 vaccination coverage in Santiago, Chile. The Gini coefficient is used to evaluate the inequity of the vaccination coverage for each school type at three time points along the COVID-19 pandemic: on November 15, 2021; March 1st, 2022; and May 26, 2022, covering the spread of different SARS-CoV-2 variants: Delta, Omicron and Omicron BA.4-BA.5, respectively. While inequity in vaccination coverage within schools’ type tend to diminish over time, it is always lower in private schools. The same pattern but with even higher inequity is found when considering the whole country.

Full size image

Discussion

This study revealed a correlation between 9 SES indices and the COVID-19 vaccination coverage in schools belonging to 34 municipalities of the Metropolitan Area of Santiago, Chile; one of the countries with the highest rate of vaccination in the world31. After analyzing the Spearman’s rank correlation between socioeconomic indices and COVID-19 vaccination coverage, we found a statistically significant correlation between lower SES and lower vaccination coverage. Even though inequity in vaccination coverage diminished over time along the pandemic, this inequity remains higher in public and state-subsidized schools compared to that of private schools, for all municipalities (Fig. 3); this inequity pattern is even more pronounced when considering data from the whole country. In general terms, our results are consistent with previous literature, indicating that vaccine coverage is lower in vulnerable populations living in poverty, having lower literacy and with a lesser educational level4,30,35. A potential rationale for these observations may be rooted in the varied personal beliefs across different SES groups. Specifically, one study noted a discernible negative correlation between SES and skepticism towards the COVID-19 vaccine’s secondary effects and conspiracy theories, offering an insightful perspective into our findings36.

In addition to the statistically significant correlation between the vulnerability of different municipalities and the COVID-19 vaccination coverage–i.e. municipalities with lower SES exhibit lower coverage–, different school types exhibit distinct vaccination coverage. Thus, both public and state-subsidized schools show lower vaccination coverage than that of private schools. Although the strength of these correlations varies depending on the municipality (Fig. 1), this trend is maintained along the pandemic (Fig. 2). However, an important difference in vaccination coverage between school types, remains. Furthermore, analysis of vaccination coverage, in terms of percentages as reported by the Chilean government, uncovers significant differences across school types. However, a more alarming situation emerges when examining the raw numbers of unvaccinated children, particularly those attending schools with the lowest vaccination rates. The total number of children enrolled in public and state-subsidized schools, which demonstrate significantly lower vaccination coverage, is more than five times greater than that in private schools, with the respective figures being 873,354 and 169,898. This discrepancy highlights the critical issue of lower vaccination rates primarily among students in public and state-subsidized schools. Collectively, these two sectors account for approximately 84% of all student enrollments in the Metropolitan Area of Santiago, Chile. Considering that municipalities with lower SES are the ones having the highest population density37, a lower vaccination coverage together with higher population density, are both key factors to consider when dealing with the higher burden of COVID-19 on these populations. Unfortunately, we were unable to extend this detailed analysis to the entire country, which is particularly concerning given the higher levels of inequity in vaccination observed. This limitation arises because the SES indices are only available at the municipal level. All the rest of Chilean cities are smaller, and generally consist of a single municipality, thus restricting our ability to conduct a detailed analysis of the SES composition and its heterogeneity. Although the disparity in vaccination coverage diminishes over time, research has demonstrated that delays in vaccination schedules are associated with heightened adverse effects38.

Available literature points to a combination of factors that may produce this uneven vaccination coverage. Among others, the lack of proper information about the importance and safety of vaccination, insufficient incentives for vaccination, religious beliefs, the negative view of pharmaceutical companies, and low trust in the government, may all have played a role in this uneven vaccination rate4,23,24,25,30,36. Research indicates that the impact of incentives on vaccination uptake varies across populations. Notably, monetary incentives have been found to be more effective than either encouraging messages or the promise of increased freedoms39,40. A critical consideration is that children typically do not independently decide to receive vaccinations; instead, these decisions are made by their parents or caregivers. A key factor influencing vaccination decisions is the perceived risk associated with the vaccine41. Although the correlations discussed herein do not show significance when considering the adult population, it is plausible that concerns regarding vaccine risk may be amplified in the context of vaccinating children. Specifically, adults might be willing to accept the perceived risk of vaccination for themselves but hesitate to subject their children to the same perceived risks, and this hesitance apparently increases for younger children. The fact that our results show that inequity in vaccination coverage increases in younger children gives strength to this factor.

Beyond the willingness to vaccinate, logistical barriers play a crucial role in determining vaccination uptake differences. These include the availability of vaccine doses and the accessibility of vaccination centers, which is influenced by their location, transportation options, queue lengths, and operating hours. In the case of the Chilean vaccination campaign, it was based on an heterogenous network of vaccination centers, with some municipalities having single, large, centralized sites, while others had multiple small venues, so ease of access was very different to the population, although there was no assignment of vaccination centers, so people were free to attend the center of their choosing. Unfortunately, we could not access detailed data on this distribution network, so a more detailed analysis of these factors was not feasible. In any case, such factors are particularly relevant for children, who require accompaniment by an adult to attend vaccination sites. This necessity poses additional challenges for households where both parents work and lack childcare support, potentially impacting vaccination rates. To enhance vaccination campaigns and public policies, a more intricate examination of these contributing factors is necessary. The reliance on broad data may risk simplifying the complexities of pandemics, underlining the importance of detailed, granular data in formulating responses to not only COVID-19 but future public health crises as well.

One of the limitations of our study is that the SES indices correspond to the municipal averages, so we cannot perform analyses nor extract conclusions at a granular level. In other words, assuming that an averaging SES index of the municipality would be an adequate descriptor for the whole spectrum of wealth exhibited on the municipality could be considered as a naive approach. This limitation becomes evident when trying to explain the differences in vaccination coverage between school types; we can only assume that the sub-populations that are enrolled in private schools (16%) are those that can afford the high tuition costs corresponding to the least vulnerable families in each municipality. Conversely, students enrolled in state-subsidized or in public schools are considered most vulnerable to the burden of COVID-19 (84%). Anyhow, without further socioeconomic details at an individual student level, we can only speculate about the causal roots of this inequity in vaccination coverage. In more general terms, other limitations are the lack of adjustment for important confounders, such as individual health conditions, access to healthcare facilities, and government trust, among many others. These factors could potentially influence both SES and vaccination rates, thereby introducing bias into our results. Future research should aim to control for these variables to provide a more accurate assessment of the relationship between SES and vaccination coverage. Another limitation is relying on secondary records, which may have other limitations such as inaccuracies in data reporting, variations in data collection methods across different municipalities, and the potential for missing data. While these secondary records are essential for large-scale studies like ours, they may not capture all nuances or errors in data entry, which could affect the outcomes of the analysis.

Considering that COVID-19 vaccination in Chile is free of charge and governmentally orchestrated, determining the causal roots behind the lower vaccination levels noted here is crucial to develop better public policies. For example, survey studies should include questions about willingness to vaccinate children and elderly dependents of the respondent and detailed socioeconomic and educational background, vaccination throughput of each vaccination center should be analyzed in detail and correlation analysis should be performed with more detailed, granular data.

Even though we are unable to identify the causes of lower vaccination ratios among school-aged children in municipalities of lower SES, it is imperative that public policies counteract this behavior. Otherwise, the inequity in vaccination coverage will be perpetuated in future pandemics, exacerbating their impact on the more vulnerable population.

Methods

Data sources

Socioeconomic status

To generate a socioeconomic profile of the municipalities in Santiago, Chile, we used nine different SES indices so to obtain a broader multi-factorial socioeconomic description. The criteria for including indices was high reliability, originating from trusted sources and being last updated no more than three years before the vaccination campaign. While these indices may not represent the most current data publicly available at the time of publication, they should reflect the reality of the different municipalities at the time of the parents’ decision to vaccinate their children. Still, it is important to acknowledge the potential for changes in the socioeconomic indices subsequent to the time of data analysis, especially for decision making and policy design for future pandemics.

The Social Priority Index (SPI) is published by the Regional Ministerial Secretariat of Social Development and Family. It is a synthetic index that integrates relevant aspects of communal social development, including income, education, and health42. It has already been used to evaluate the correlation between SES and the impact of COVID-193, however, it is only available for some municipalities of the Metropolitan Region. This index is not updated in a regular schedule, the current version is from 2022, and the previous one is from 2019, which is the one used in our analysis and in the work above.

The School Vulnerability Index (SVI), published yearly by the National Board of School Aid and Scholarships (JUNAEB), is the ratio of the sum of the students in the first, second, and third priority ranking compared to the total school enrollment. Students are classified in priorities 1, 2, and 3 according to poverty conditions and risk of school failure. Thus, students living in lower socioeconomic conditions (Priority 1) tend to exhibit a higher risk of school failure, and thus receive the highest priority from the Board43. For this study we considered the 2021 data.

The Community Development Index (CDI), published in 2020, is a comprehensive index that integrates three socioeconomic dimensions: Health and Social Welfare; Economy and Resources; and Education. The development of the CDI involves a meticulous process of selecting a variety of standardized and comparable indicators, aims to provide a reliable instrument for monitoring and quantifying the impact of development processes from a territorial perspective. The CDI, as well as its individual dimensions, is available for each municipality in the country44, offering a tool for local management and decision-making. In this work, we have chosen to consider the CDI and its individual components separately. This decision stems from our aim to leverage the distinct yet complementary insights that the aggregate index and its underlying dimensions offer. The CDI provides a holistic view of community development across municipalities, facilitating broad comparisons and trend analysis. At the same time, analyzing the individual dimensions enables a deeper dive into specific areas of socioeconomic development, allowing for a nuanced understanding of strengths and challenges within each municipality. This dual approach enriches our analysis, contributing to a more comprehensive socioeconomic profile that can better inform targeted strategies for community development. This is the only index that does not come directly from a government agency, it was published by an academic institution. Even though the publication itself was not peer reviewed, the institution and team behind the study are reputable and the data has been used in different studies45,46.

The following indices are calculated as the percentage of people belonging to each category respectively for each municipality. Poverty is determined by a threshold defined as a function of income level and household size. Multidimensional Poverty is defined as being part of a household that cannot achieve adequate living conditions in a set of five relevant dimensions of well-being: (1) Education; (2) Health; (3) Labor and Social Security; (4) Housing and Environment; and (5) Networks and Social Cohesion. These conditions are measured through a weighted set of 15 indicators (three for each dimension). Households that accumulate deficiencies of 22.5% or more are classified as being in a situation of multidimensional poverty. The Poverty and Multidimensional Poverty indices were extracted from the National Socioeconomic Characterization Survey (CASEN)47, while the PHLBS, which is the percentage of People in Households Lacking Basic Services, was extracted from the Social Registry of Households 2020 44. Both instruments are used by the Ministry of Social Development to focus social assistance among the population .

The decision to evaluate these SES indices individually, rather than consolidating them into a single composite SES index, was driven by two primary considerations. First, analyzing the indices separately facilitates a nuanced understanding of the various dimensions of SES and their distinct impacts on vaccination coverage among school-aged children. Second, due to the complex and multifaceted nature of SES, a composite index could potentially obscure significant disparities and trends that are only discernible when these dimensions are examined on their own. In the same line of reasoning, the results show all the publicly available indices related to SES for completeness and to minimize biases in the analysis. This methodological approach aligns with our primary objective: to investigate the relationship between different aspects of SES and vaccination coverage in children, ensuring a more accurate and comprehensive analysis.

Vaccinations

Even though in Chile the COVID-19 vaccination campaign is orchestrated and conducted by the Ministry of Health, the Ministry of Education publishes vaccination coverage per grade and school on its website32. The Ministry of Health administers the National Immunization Program (Programa Nacional de Inmunización, PNI), a government initiative that has successfully eradicated important contagious diseases through mandatory vaccination and educational campaigns48. Through this program, the Ministry administered the vaccination centers, where every person had to show their official identification card in order to receive the vaccine, thus generating a reliable record by registering the national identity number (RUT), the date and the specific vaccine received. In the case of primary students, the vaccine rollout is detailed in Table 2. To accurately assess the extent of vaccine coverage in each school and grade, the Ministry of Education combines its own database of school enrollments with the vaccination records from the Ministry of Health. This integration is accomplished by using the RUT of each student, which is the unique personal identifier in Chile. By using the RUT, they can precisely link each student’s school records with their vaccination status, ensuring a detailed and accurate understanding of the vaccination coverage across different educational institutions. To study whether the spread of different SARS-CoV-2 variants among the population might influence the vaccination process among school-age students in different municipalities in Santiago, Chile, we downloaded data published on November 15, 2021; March 1st, 2022; and May 26, 2022. By selecting these dates, we covered the spread of different SARS-CoV-2 variants: Delta in November, Omicron in March and Omicron subvariants BA.4-BA.5 in May17. The data shows the proportion of students in each school and grade who are fully vaccinated, partially vaccinated, or not vaccinated at all. ‘Fully vaccinated’ refers to students who have received all required doses of the COVID-19 vaccine, ‘partially vaccinated’ indicates those who have only received some but not all of the necessary doses, and ‘not vaccinated’ denotes students who have not received any doses of the vaccine.

Table 2 Vaccine approval schedule.
Full size table

School metadata

The Chilean educational system has three main types of schools according to its funding source: Public schools are fully state-funded and under the administration of each municipality, these schools are free of tuition costs so they are attended mostly by low income families; private schools are The Chilean Ministry of Education provides easily accessible metadata for each educational institution on their public repository49. This metadata was used to identify public, state-subsidized and private schools in every municipality.

Statistical analysis

All statistical analyses were performed on the downloaded data by programming in Python v3.1 using the SciPy v1.7.1 package50. To measure inequity, we computed the Gini Coefficient as half of the relative mean absolute difference:

$$\:G=\frac{{\sum\:}_{i=1}^{N}{\sum\:}_{j=1}^{N}|{x}_{i}-{x}_{j}|}{2{n}^{2}\overline{x}}\:$$

Where N is the number of reported vaccination levels and \(\:{x}_{j}\)are individual percentages of fully vaccinated students and \(\:\overline{x}\) is the mean of the values.

Spearman’s rank-order correlation was computed using Statsmodels v0.13.1. Plots were made with Seaborn v0.11.2.

Ethical considerations

All data used in this research was sourced in an anonymous form.