Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Mass gatherings for political expression had no discernible association with the local course of the COVID-19 pandemic in the USA in 2020 and 2021

Nature Human Behaviour volume 7, pages 1708–1728 (2023)Cite this article

Subjects

Abstract

Epidemic disease can spread during mass gatherings. We assessed the impact of a type of mass gathering about which comprehensive data were available on the local-area trajectory of the COVID-19 epidemic. Here we examined five types of political event in 2020 and 2021: the US primary elections, the US Senate special election in Georgia, the gubernatorial elections in New Jersey and Virginia, Donald Trump’s political rallies and the Black Lives Matter protests. Our study period encompassed over 700 such mass gatherings during multiple phases of the pandemic. We used data from the 48 contiguous states, representing 3,108 counties, and we implemented a novel extension of a recently developed non-parametric, generalized difference-in-difference estimator with a (high-quality) matching procedure for panel data to estimate the average effect of the gatherings on local mortality and other outcomes. There were no statistically significant increases in cases, deaths or a measure of epidemic transmissibility (Rt) in a 40-day period following large-scale political activities. We estimated small and statistically non-significant effects, corresponding to an average difference of −0.0567 deaths (95% CI = −0.319, 0.162) and 8.275 cases (95% CI = −1.383, 20.7) on each day for counties that held mass gatherings for political expression compared to matched control counties. In sum, there is no statistical evidence of a material increase in local COVID-19 deaths, cases or transmissibility after mass gatherings for political expression during the first 2 years of the pandemic in the USA. This may relate to the specific manner in which such activities are typically conducted.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Overview of matching and estimation.
Fig. 2: Impact of all mass gatherings for political expression on COVID-19 mortality and case rates in the USA.
Fig. 3: Impact of the primary elections on COVID-19 mortality and case rates.
Fig. 4: Impact of the GA special election and the NJ and VA gubernatorial elections combined on COVID-19 mortality and case rates.
Fig. 5: Impact of Donald Trump’s rallies on COVID-19 mortality and case rates.
Fig. 6: Impact of the BLM protests on COVID-19 mortality.
Fig. 7: Impact of large-scale political events on COVID-19 Rt.
Fig. 8: Impact of large-scale political events on COVID-19 mobility.

Data availability

The data on COVID-19 case and death counts are available from the Johns Hopkins Coronavirus Resource Center (https://coronavirus.jhu.edu/data). US Census data are also publicly available (https://www.census.gov/programs-surveys/acs). The mobility tracking data are available from SafeGraph, Inc. and are freely available to academic researchers (https://www.safegraph.com/products/places). The elections turnout data are available directly from state governmental election agencies. The protest event data are available as the US Crisis Monitor dataset from the Armed Conflict Location & Event Data Project (https://acleddata.com/special-projects/us-crisis-monitor/). County-level masking data are available from The New York Times GitHub repository (https://raw.githubusercontent.com/nytimes/covid-19-data/master/mask-use/mask-use-by-county.csv).

Code availability

Data cleaning and preparation was carried out with the R programming language106. All analysis was conducted in the Julia programming language105 with the ‘TSCSMethods’ Julia package104 (https://github.com/human-nature-lab/TSCSMethods.jl). In addition, paper replication materials are available on GitHub as the ‘COVIDPoliticalEvents’ Julia package (https://github.com/human-nature-lab/COVIDPoliticalEvents.jl).

References

  1. Christakis, N. A. Apollo’s Arrow: The Profound and Enduring Impact of Coronavirus on the Way We Live (Little, Brown Spark, 2020).

  2. Enten, H. Trump is creating an untraditional partisan divide on vote by mail. CNN https://www.cnn.com/2020/07/27/politics/vote-by-mail-partisan-divide-analysis/index.html (2020).

  3. Milosh, M., Painter, M., Sonin, K., Dijcke, D. V. & Wright, A. L. Unmasking partisanship: polarization undermines public response to collective risk. J. Public Econ. 204, 104538 (2021).

  4. Easley, J. & Miller, M. Pandemic sparks partisan brawl over voting by mail. The Hill https://thehill.com/homenews/campaign/491862-pandemic-sparks-partisan-brawl-over-voting-by-mail (2020).

  5. Perrett, C. Fauci says ‘there’s no reason’ in-person voting shouldn’t be safe with masks and proper social distancing. Business Insider https://www.businessinsider.com/fauci-says-in-person-election-with-distancing-masks-is-safe-2020-8 (2020).

  6. Reichmann, D. As virus surges, Trump rallies keep packing in thousands. AP NEWS https://apnews.com/article/donald-trump-rallies-virus-surges-50e79fabd46472c51ecc1444184082de (2021).

  7. Cillizza, C. Donald Trump is (still) totally obsessed with crowd size. CNN https://www.cnn.com/2020/10/27/politics/donald-trump-crowds-rallies/index.html (2020).

  8. Buchanan, L., Bui, Q. & Patel, J. K. Black Lives Matter may be the largest movement in U.S. history. The New York Times (2020).

  9. Madrigal, A. C. & Meyer, R. America is giving up on the pandemic. The Atlantic https://www.theatlantic.com/science/archive/2020/06/america-giving-up-on-pandemic/612796/ (2020).

  10. Palus, S. Public health experts say the pandemic is exactly why protests must continue. Slate https://slate.com/technology/2020/06/protests-coronavirus-pandemic-public-health-racism.html (2020).

  11. Yancy, C. W. COVID-19 and African Americans. JAMA 323, 1891–1892 (2020).

    Article  CAS  PubMed  Google Scholar 

  12. Agarwal, V, Pokharel, K. & Mandhana, N. As Covid-19 surged, Indian teachers died after working elections. Wall Street Journal https://www.wsj.com/articles/as-covid-19-surged-indian-teachers-died-after-working-elections-11620558475 (2021).

  13. A Guide to 2021 Latin American Elections https://www.as-coa.org/content/guide-2021-latin-american-elections (Americas Society/Council of the Americas, 2021).

  14. Keating, C. Connecticut House OKs absentee voting expansion despite Republican concerns. Hartford Courant https://www.courant.com/2022/03/16/connecticut-house-oks-absentee-voting-expansion-despite-republican-concerns/ (2022).

  15. Pasieka, C. How the pandemic might affect voter turnout, and Ford’s chances in June. CBC News https://www.cbc.ca/news/canada/toronto/ontario-election-turnout-2022-1.6388401 (2022).

  16. Kong, K. Seoul’s full cafes, Apple store lines show mass testing success. BNN Bloomberg https://www.bloomberg.com/news/articles/2020-04-18/seoul-s-full-cafes-apple-store-lines-show-mass-testing-success#xj4y7vzkg (2020).

  17. Spinelli, A. Managing Elections Under the COVID-19 Pandemic: The Republic of Korea’s Crucial Test (International Institute for Democracy and Electoral Assistance, 2020).

  18. Robles, F. Cubans denounce ‘misery’ in biggest protests in decades. The New York Times https://www.nytimes.com/2021/07/11/world/americas/cuba-crisis-protests.html (2021).

  19. COVID ‘superspreader’? India warns against mass farmer protests. Al Jazeera https://www.aljazeera.com/economy/2021/5/24/india-warns-farmer-protests-can-be-covid-super-spreader (2021).

  20. Balmforth, T. Russia cracks down on home front as its troops invade Ukraine. Reuters https://www.reuters.com/world/europe/russia-cracks-down-home-front-its-troops-invade-ukraine-2022-02-24/ (2022).

  21. Sonabend, R. et al. Non-pharmaceutical interventions, vaccination, and the SARS-CoV-2 Delta variant in England: a mathematical modelling study. Lancet 398, 1825–1835 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Sharma, M. et al. Understanding the effectiveness of government interventions against the resurgence of COVID-19 in Europe. Nat. Commun. 12, 5820 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Callaway, E. Delta coronavirus variant: scientists brace for impact. Nature 595, 17–18 (2021).

    Article  CAS  PubMed  Google Scholar 

  24. Earnest, R. et al. Comparative transmissibility of SARS-CoV-2 variants Delta and Alpha in New England, USA. Cell Rep. Med. 3, 100583 (2022).

  25. Kamp, J. & Abbott, B. Delta surge of Covid-19 recedes, leaving winter challenge ahead. Wall Street Journal https://www.wsj.com/articles/delta-surge-of-covid-19-recedes-leaving-winter-challenge-ahead-11635672600 (2021).

  26. Ferré-Sadurní, L. & Goldstein, J. 1st vaccination in U.S. is given in New York, hard hit in outbreak’s first days. The New York Times https://www.nytimes.com/2020/12/14/nyregion/coronavirus-vaccine-new-york.html (2020).

  27. COVID-19 Vaccinations in the United States (CDC, 2022).

  28. Spiegel, M. & Tookes, H. Business restrictions and COVID-19 fatalities. Rev. Financ. Stud. 34, 5266–5308 (2021).

    Article  Google Scholar 

  29. Sabin, S. 2 in 3 voters at least somewhat concerned about voting in person during coronavirus. Morning Consult https://morningconsult.com/2020/03/20/voting-options-poll-coronavirus/ (2020).

  30. Dave, D. M., Friedson, A. I., Matsuzawa, K., Sabia, J. J. & Safford, S. Black Lives Matter Protests and Risk Avoidance: The Case of Civil Unrest During a Pandemic (NBER, 2020); https://www.nber.org/papers/w27408

  31. Chitwood, M. H. et al. Reconstructing the course of the COVID-19 epidemic over 2020 for US states and counties: results of a Bayesian evidence synthesis model. PLoS Comput. Biol. 18, e1010465 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Petrone, M. E. et al. Combining genomic and epidemiological data to compare the transmissibility of SARS-CoV-2 variants Alpha and Iota. Commun. Biol. 5, 439 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Kishore, N. et al. Evaluating the reliability of mobility metrics from aggregated mobile phone data as proxies for SARS-CoV-2 transmission in the USA: a population-based study. Lancet Digit. Health 4, e27–e36 (2022).

    Article  CAS  PubMed  Google Scholar 

  34. Weill, J., Stigler, M., Deschenes, O. & Springborn, M. Researchers’ Degrees-of-Flexibility and the Credibility of Difference-in-Differences Estimates: Evidence From the Pandemic Policy Evaluations (NBER, 2021); http://www.nber.org/papers/w29550.pdf

  35. Imai, K., Kim, I. S. & Wang, E. H. Matching methods for causal inference with time-series cross-sectional data. Am. J. Polit. Sci. https://doi.org/10.1111/ajps.12685 (2021).

  36. Imbens, G. W. & Rubin, D. B. Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction (Cambridge Univ. Press, 2015).

  37. Fowler, J. H., Hill, S. J., Levin, R. & Obradovich, N. Stay-at-home orders associate with subsequent decreases in COVID-19 cases and fatalities in the United States. PLoS ONE 16, e0248849 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Hsiang, S. et al. The effect of large-scale anti-contagion policies on the COVID-19 pandemic. Nature 584, 262–267 (2020).

    Article  CAS  PubMed  Google Scholar 

  39. Jia, J. S. et al. Population flow drives spatio-temporal distribution of COVID-19 in China. Nature 582, 389–394 (2020).

    Article  CAS  PubMed  Google Scholar 

  40. Crawford, F. W. et al. Impact of close interpersonal contact on COVID-19 incidence: evidence from 1 year of mobile device data. Sci. Adv. https://doi.org/10.1126/sciadv.abi5499 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  41. Chang, S. et al. Mobility network models of COVID-19 explain inequities and inform reopening. Nature 589, 82–87 (2021).

    Article  CAS  PubMed  Google Scholar 

  42. Badr, H. S. et al. Association between mobility patterns and COVID-19 transmission in the USA: a mathematical modelling study. Lancet Infect. Dis. 20, 1247–1254 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Zhang, J. et al. Changes in contact patterns shape the dynamics of the COVID-19 outbreak in China. Science 368, 1481–1486 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Kang, Y. et al. Multiscale dynamic human mobility flow dataset in the U.S. during the COVID-19 epidemic. Sci. Data 7, 390 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Verity, R. et al. Estimates of the severity of coronavirus disease 2019: a model-based analysis. Lancet Infect. Dis. 20, 669–677 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Flaxman, S. et al. Estimating the effects of non-pharmaceutical interventions on COVID-19 in Europe. Nature 584, 257–261 (2020).

    Article  CAS  PubMed  Google Scholar 

  47. Phillips, A. What you need to know about the Georgia Senate runoff elections. Washington Post https://www.washingtonpost.com/politics/interactive/2020/georgia-senate-runoff-guide/ (2021).

  48. McCartney, A. Turnout hits historic highs in contentious Georgia Senate races. Bloomberg https://www.bloomberg.com/graphics/2021-georgia-senate-runoff/ (2021).

  49. Constantino, A. K. Virginia election sees highest turnout in recent history, fueling Glenn Youngkin’s victory. CNBC https://www.cnbc.com/2021/11/03/virginia-election-sees-highest-turnout-in-recent-history-fueling-glenn-youngkins-victory.html (2021).

  50. Enten, H. 2021 shows Republicans shouldn’t fear high voter turnout. CNN https://www.cnn.com/2021/11/07/politics/turnout-republicans-analysis/index.html (2021).

  51. Petherick, A. et al. A Worldwide Assessment of COVID-19 Pandemic-Policy Fatigue (SSRN, 2021); https://papers.ssrn.com/abstract=3774252

  52. Crabtree, S. Americans’ social distancing habits have tapered since July. Gallup https://news.gallup.com/poll/322064/americans-social-distancing-habits-tapered-july.aspx (2020).

  53. Gen Z and the Toll of the Pandemic (AP-NORC, 2021); https://apnorc.org/projects/gen-z-and-the-toll-of-the-pandemic/

  54. The Red/Blue Divide in COVID-19 Vaccination Rates (KFF, 2021); https://www.kff.org/policy-watch/the-red-blue-divide-in-covid-19-vaccination-rates/

  55. Hughes, M. M. et al. County-level COVID-19 vaccination coverage and social vulnerability — United States, December 14, 2020–March 1, 2021. MMWR Morb. Mortal. Wkly Rep. 70, 431–436 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. King, G., Rosen, O. & Tanner, M. A. Ecological Inference: New Methodological Strategies (Cambridge Univ. Press, 2004).

  57. Shridhar, S. V., Alexander, M. & Christakis, N. A. Characterizing super-spreaders using population-level weighted social networks in rural communities. Phil. Trans. R. Soc. Math. Phys. Eng. Sci. 380, 20210123 (2022).

    Google Scholar 

  58. Mahale, P. Multiple COVID-19 outbreaks linked to a wedding reception in rural Maine — August 7–September 14, 2020. MMWR Morb. Mortal. Wkly Rep. 69, 1686–1690 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Hamner, L. High SARS-CoV-2 attack rate following exposure at a choir practice — Skagit County, Washington, March 2020. MMWR Morb. Mortal. Wkly Rep. 69, 606–610 (2020).

    Article  CAS  PubMed  Google Scholar 

  60. Leung, K., Wu, J. T., Xu, K. & Wein, L. M. No detectable surge in SARS-CoV-2 transmission attributable to the April 7, 2020 Wisconsin election. Am. J. Public Health 110, 1169–1170 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  61. Cotti, C., Engelhardt, B., Foster, J., Nesson, E. & Niekamp, P. The relationship between in-person voting and COVID-19: evidence from the Wisconsin primary. Contemp. Econ. Policy 39, 760–777 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  62. Berry, A. C., Mulekar, M. S. & Berry, B. B. Increase in daily new COVID-19 cases not seen following the Wisconsin primary election April 2020. J. Infect. Dis. Epidemiol. https://doi.org/10.23937/2474-3658/1510148 (2020).

    Article  Google Scholar 

  63. Neyman, G. & Dalsey, W. Black Lives Matter protests and COVID-19 cases: relationship in two databases. J. Public Health https://doi.org/10.1093/pubmed/fdaa212 (2020).

    Article  Google Scholar 

  64. Cassan, G. & Sangnier, M. The impact of 2020 French municipal elections on the spread of COVID-19. J. Popul. Econ. 35, 963–988 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  65. Duchemin, L., Veber, P. & Boussau, B. Bayesian investigation of SARS-CoV-2-related mortality in France. Peer Community J. 2, e6 (2022).

  66. Zeitoun, J.-D. et al. Reciprocal association between participation to a national election and the epidemic spread of COVID-19 in France: nationwide observational and dynamic modeling study. Eur. J. Public Health https://doi.org/10.1093/eurpub/ckab140 (2021).

    Article  PubMed  Google Scholar 

  67. Palguta, J., Levínský, R. & Škoda, S. Do elections accelerate the COVID-19 pandemic? J. Popul. Econ. 35, 197–240 (2022).

    Article  PubMed  Google Scholar 

  68. Cipullo, D. & Le Moglie, M. To vote, or not to vote? Electoral campaigns and the spread of COVID-19. Eur. J. Polit. Econ. 72, 102118 (2022).

    Article  PubMed  Google Scholar 

  69. Tarouco, G. Covid-19 and the Brazilian 2020 Municipal Elections (International IDEA, 2020).

  70. Bach, L., Guillouzouic, A. & Malgouyres, C. Does holding elections during a Covid-19 pandemic put the lives of politicians at risk? J. Health Econ. 78, 102462 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  71. Bernheim, B. D., Buchmann, N., Freitas-Groff, Z. & Otero, S. The Effects of Large Group Meetings on the Spread of COVID-19: The Case of Trump Rallies (SSRN, 2020); https://doi.org/10.2139/ssrn.3722299

  72. Korolev, I. Identification and estimation of the SEIRD epidemic model for COVID-19. J. Econ. 220, 63–85 (2021).

    Article  Google Scholar 

  73. Weed, M. & Foad, A. Rapid scoping review of evidence of outdoor transmission of COVID-19. Preprint at medRxiv https://doi.org/10.1101/2020.09.04.20188417 (2020).

  74. Dave, D., McNichols, D. & Sabia, J. J. The contagion externality of a superspreading event: the Sturgis Motorcycle Rally and COVID-19. South. Econ. J. 87, 769–807 (2021).

    Article  PubMed  Google Scholar 

  75. Javid, B. et al. Should masks be worn outdoors? Brit. Med. J. 373, n1036 (2021).

    Article  PubMed  Google Scholar 

  76. Jia, J. S., Yuan, Y., Jia, J. & Christakis, N. A. Risk perception and behaviour change after personal vaccination for COVID-19 in the USA. Preprint at PsyArXiv https://doi.org/10.31234/osf.io/afyv8 (2022).

  77. Brauner, J. M. et al. Inferring the effectiveness of government interventions against COVID-19. Science https://doi.org/10.1126/science.abd9338 (2021).

    Article  PubMed  Google Scholar 

  78. Li, Y. et al. The temporal association of introducing and lifting non-pharmaceutical interventions with the time-varying reproduction number (R) of SARS-CoV-2: a modelling study across 131 countries. Lancet Infect. Dis. 21, 193–202 (2021).

    Article  PubMed  Google Scholar 

  79. Haug, N. et al. Ranking the effectiveness of worldwide COVID-19 government interventions. Nat. Hum. Behav. 4, 1303–1312 (2020).

    Article  PubMed  Google Scholar 

  80. Election demographics and voter turnout. Bloomberg Government https://about.bgov.com/brief/election-demographics-and-voter-turnout/ (2022).

  81. Dong, E. et al. The Johns Hopkins University Center for Systems Science and Engineering COVID-19 Dashboard: data collection process, challenges faced, and lessons learned. Lancet Infect. Dis. 22, e370–e376 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  82. Chin, T. et al. US-county level variation in intersecting individual, household and community characteristics relevant to COVID-19 and planning an equitable response: a cross-sectional analysis. BMJ Open 10, e039886 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  83. Miller, I. F., Becker, A. D., Grenfell, B. T. & Metcalf, C. J. E. Disease and healthcare burden of COVID-19 in the United States. Nat. Med. 26, 1212–1217 (2020).

    Article  CAS  PubMed  Google Scholar 

  84. Cottrell, D., Herron, M. C. & Smith, D. A. Vote-by-mail ballot rejection and experience with mail-in voting. Am. Polit. Res. 49, 577–590 (2021).

    Article  Google Scholar 

  85. Shino, E., Suttmann-Lea, M. & Smith, D. A. Determinants of rejected mail ballots in Georgia’s 2018 General Election. Polit. Res. Q. 75, 231–243 (2022).

    Article  Google Scholar 

  86. Merkley, E., Bergeron, T., Loewen, P. J., Elias, A. & Lapp, M. Communicating safety precautions can help maintain in-person voter turnout during a pandemic. Elect. Stud. 75, 102421 (2022).

    Article  PubMed  Google Scholar 

  87. Fernandez-Navia, T., Polo-Muro, E. & Tercero-Lucas, D. Too afraid to vote? The effects of COVID-19 on voting behaviour. Eur. J. Polit. Econ. 69, 102012 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  88. In praise of replication studies and null results. Nature 578, 489–490 (2020).

  89. Gardner, L., Ratcliff, J., Dong, E. & Katz, A. A need for open public data standards and sharing in light of COVID-19. Lancet Infect. Dis. 21, e80 (2021).

    Article  CAS  PubMed  Google Scholar 

  90. Chitwood, M. H. et al. covidestim: COVID-19 nowcasting (CDC, CSTE, 2021); https://covidestim.org/

  91. Unwin, H. J. T. et al. State-level tracking of COVID-19 in the United States. Nat. Commun. 11, 6189 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  92. Nouvellet, P. et al. Reduction in mobility and COVID-19 transmission. Nat. Commun. 12, 1090 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  93. Athey, S., Blei, D., Donnelly, R., Ruiz, F. & Schmidt, T. Estimating heterogeneous consumer preferences for restaurants and travel time using mobile location data. AEA Pap. Proc. 108, 64–67 (2018).

    Article  Google Scholar 

  94. Squire, R. F. Advanced methods for correlating safegraph patterns with other datasets across time. SafeGraph https://colab.research.google.com/drive/16BELpcum4TKoH-5wg8Xym_CGgIGgpu1I?usp=sharing#scrollTo=fZ0aE3PBEafp (2020).

  95. Demonstrations and Political Violence in America: New Data for Summer 2020 (ACLED, 2020); https://acleddata.com/2020/09/03/demonstrations-political-violence-in-america-new-data-for-summer-2020/

  96. Sobolev, A., Chen, M. K., Joo, J. & Steinert-Threlkeld, Z. C. News and geolocated social media accurately measure protest size variation. Am. Polit. Sci. Rev. 114, 1343–1351 (2020).

    Article  Google Scholar 

  97. Georgia Early Voting Statistics - 2021 Senate Run-off Election (United States Elections Project, 2021); https://electproject.github.io/Early-Vote-2020G/GA_RO.html

  98. Election Information and Results Archive (NJ Division of Elections, 2021); https://nj.gov/state/elections/election-information.shtml

  99. Election Results (Virginia Department of Elections, 2021); https://www.elections.virginia.gov/resultsreports/election-results/

  100. County Adjacency File (The United States Census Bureau, 2018); https://www.census.gov/geographies/reference-files/2010/geo/county-adjacency.html

  101. American Community Survey 5-Year Data (2009–2018) (The United States Census Bureau, 2018); https://www.census.gov/data/developers/data-sets/acs-5year.html

  102. The New York Times. Mask-Wearing Survey Data (GitHub, 2020); https://github.com/nytimes/covid-19-data

  103. Bezanson, J., Edelman, A., Karpinski, S. & Shah, V. B. Julia: a fresh approach to numerical computing. SIAM Rev. 59, 65–98 (2017).

    Article  Google Scholar 

  104. Feltham, E. M. TSCSMethods.jl. GitHub https://github.com/human-nature-lab/TSCSMethods.jl (2022).

  105. R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2021).

  106. Strochak, S., Ueyama, K. & Williams, A. urbnmapr: state and county shapefiles in sf and tibble format. GitHub https://github.com/UrbanInstitute/urbnmapr (2023).

  107. Bernal, J. L., Cummins, S. & Gasparrini, A. Interrupted time series regression for the evaluation of public health interventions: a tutorial. Int. J. Epidemiol. 46, 348–355 (2017).

    PubMed  Google Scholar 

  108. O’Neill, S., Kreif, N., Grieve, R., Sutton, M. & Sekhon, J. S. Estimating causal effects: considering three alternatives to difference-in-differences estimation. Health Serv. Outcomes Res. Methodol. 16, 1–21 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  109. Abadie, A. Semiparametric difference-in-differences estimators. Rev. Econ. Stud. 72, 1–19 (2005).

    Article  Google Scholar 

  110. Zeldow, B. & Hatfield, L. A. Confounding and regression adjustment in difference-in-differences studies. Health Serv. Res. 56, 932–941 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  111. Lechner, M. The estimation of causal effects by difference-in-difference methods. Found. Trends Econ. 4, 165–224 (2011).

    Article  Google Scholar 

  112. Taddeo, M. M., Amorim, L. D. & Aquino, R. Causal measures using generalized difference-in-difference approach with nonlinear models. Stat. Interface 15, 399–413 (2022).

    Article  Google Scholar 

  113. Siedner, M. J. et al. Social distancing to slow the US COVID-19 epidemic: longitudinal pretest–posttest comparison group study. PLoS Med. 17, e1003244 (2020).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  114. Murray, T. Stay-at-home orders, mobility patterns, and spread of COVID-19. Am. J. Public Health 111, 1149–1156 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  115. Buckee, C. O. et al. Aggregated mobility data could help fight COVID-19. Science 368, 145–146 (2020).

    Article  PubMed  Google Scholar 

  116. Bhattacharjee, S., Liao, S., Paul, D. & Chaudhuri, S. Inference on the dynamics of COVID-19 in the United States. Sci. Rep. 12, 2253 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  117. Abouk, R. & Heydari, B. The immediate effect of COVID-19 policies on social-distancing behavior in the United States. Public Health Rep. 136, 245–252 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

  118. Bonvini, M., Kennedy, E. H., Ventura, V. & Wasserman, L. Causal inference for the effect of mobility on COVID-19 deaths. Ann. Appl. Stat. 16, 2458–2480 (2022).

    Article  Google Scholar 

  119. Kraemer, M. U. G. et al. The effect of human mobility and control measures on the COVID-19 epidemic in China. Science 368, 493–497 (2020).

    Article  CAS  PubMed  Google Scholar 

  120. Zhang, M. et al. Human mobility and COVID-19 transmission: a systematic review and future directions. Ann. GIS 28, 501–514 (2022).

    Article  Google Scholar 

  121. Wikle, N. B. et al. SARS-CoV-2 epidemic after social and economic reopening in three U.S. states reveals shifts in age structure and clinical characteristics. Sci. Adv. 8, eabf9868 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  122. Goldman, N., Pebley, A. R., Lee, K., Andrasfay, T. & Pratt, B. Racial and ethnic differentials in COVID-19-related job exposures by occupational standing in the US. PLoS ONE 16, e0256085 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  123. Rosenbaum, P. R. & Rubin, D. B. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am. Stat. 39, 33–38 (1985).

    Google Scholar 

  124. Cochran, W. G. & Rubin, D. B. Controlling bias in observational studies: a review. Sankhya Indian J. Stat. A 35, 417–446 (1973).

    Google Scholar 

  125. Austin, P. C. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm. Stat. 10, 150–161 (2011).

    Article  PubMed  Google Scholar 

  126. Abadie, A. & Imbens, G. W. On the failure of the bootstrap for matching estimators. Econometrica 76, 1537–1557 (2008).

    Article  Google Scholar 

  127. Kass, R. E. & Raftery, A. E. Bayes factors. J. Am. Stat. Assoc. 90, 773–795 (1995).

    Article  Google Scholar 

  128. Morey, R. D. & Rouder, J. N. Bayes factor approaches for testing interval null hypotheses. Psychol. Methods 16, 406–419 (2011).

    Article  PubMed  Google Scholar 

  129. Morey, R. et al. BayesFactor (CRAN, 2022); https://richarddmorey.github.io/BayesFactor/

  130. Forastiere, L., Airoldi, E. M. & Mealli, F. Identification and estimation of treatment and interference effects in observational studies on networks. J. Am. Stat. Assoc. 116, 901–918 (2021).

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We thank B. Snyder for help with data development and G. King for helpful comments. This research was supported by the Robert Wood Johnson Foundation. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Author information

Authors and Affiliations

  1. Yale Institute for Network Science, Yale University, New Haven, CT, USA

    Eric Feltham, Laura Forastiere, Marcus Alexander & Nicholas A. Christakis

  2. Department of Sociology, Yale University, New Haven, CT, USA

    Eric Feltham & Nicholas A. Christakis

  3. Department of Biostatistics, Yale School of Public Health, New Haven, CT, USA

    Laura Forastiere

  4. Frank H. Netter MD School of Medicine, Quinnipiac University, North Haven, CT, USA

    Marcus Alexander

  5. Department of Statistics and Data Science, Yale University, New Haven, CT, USA

    Nicholas A. Christakis

  6. Department of Medicine, Yale School of Medicine, New Haven, CT, USA

    Nicholas A. Christakis

Authors
  1. Eric Feltham
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Laura Forastiere
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Marcus Alexander
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Nicholas A. Christakis
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

E.F., L.F., M.A. and N.A.C. conceptualized the project. E.F., L.F., M.A. and N.A.C. developed the methodology. E.F. and L.F. performed statistical analysis. N.A.C. acquired funding. E.F., L.F., M.A. and N.A.C. wrote the paper.

Corresponding author

Correspondence to Eric Feltham.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Human Behaviour thanks Engy Ziedan, Yesim Tozan and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Pre-outcome-window ATTs.

ATTs for the death and case rates, from 30 days before treatment up to the 9 days after treatment. The light green cell is the day of treatment, and the dark grey cell is the reference day for the ATT calculation (see Methods). In each panel, the error bars indicate the 95% CIs. The panels present the ‘average treatment effect on the treated’ (ATT) estimates for the (a) omnibus analysis, (b) the primary elections, (c) GA special election, (d) NJ & VA gubernatorial elections, (e) Donald Trump’s political rallies, and (f) the BLM protests. In each case, we observe results that are similar to those for the corresponding main analysis; that is, generally non-significant results, which is expected for the period prior to treatment.

Extended Data Fig. 2 Omnibus estimates for the effect of political events on the death rate.

(a) Overall ATT estimates and covariate balance before matching refinement. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. There are 2135 treated units present. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county. (c) Overall ATT estimates and covariate balance before matching refinement and after the application of a caliper. (d) Overall ATT estimates and covariate balance after the application of a caliper, and after matching refinement, the observed balance scores are, on average over the matching window, within the threshold of 0.1, indicating sufficient similarity between the treated and matched counties. On average, for estimates over the outcome window, 1449.4 treated units remain, with 6084.5 matches.

Extended Data Fig. 3 Omnibus estimates for the effect of political events on the case rate.

(a) Overall ATT estimates and covariate balance before matching refinement. (The ATTs represent the average difference in the change in case rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. There are 2135 treated units present. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county. (c) Overall ATT estimates and covariate balance before matching refinement and after the application of a caliper. (d) Overall ATT estimates and covariate balance after the application of a caliper, and after matching refinement, the observed balance scores are, on average over the matching window, within the threshold of 0.1, indicating sufficient similarity between the treated and matched counties. On average, for estimates over the outcome window, 1473.5 treated units remain, with 6305.4 matches.

Extended Data Fig. 4 Overall estimates for the primary elections.

In each panel, the error bars indicate the 95% CIs. (a) Overall ATT estimates and covariate balance before matching refinement. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. There are 1173 treated units present. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county. (c) Overall ATT estimates and covariate balance before matching refinement and after the application of a caliper. (d) Overall ATT estimates and covariate balance after the application of a caliper, and after matching refinement, the observed balance scores are, on average over the matching window, within the threshold of 0.1, indicating sufficient similarity between the treated and matched counties. On average, for estimates over the outcome window, 961.2 treated units remain, with 4283.8 matched units.

Extended Data Fig. 5 Overall estimates for the GA elections.

In each panel, the error bars indicate the 95% CIs. (a) Overall ATT estimates and covariate balance before matching refinement. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. There are 159 treated units present. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county. (c) Overall ATT estimates and covariate balance before matching refinement and after the application of a caliper. (d) Overall ATT estimates and covariate balance after the application of a caliper, and after matching refinement, the observed balance scores are, on average over the matching window, within the threshold of 0.1, indicating sufficient similarity between the treated and matched counties. On average, for estimates over the outcome window, 137 treated units remain, with 566 matched units.

Extended Data Fig. 6 Overall estimates for the NJ and VA gubernatorial elections.

In each panel, the error bars indicate the 95% CIs. (a) Overall ATT estimates and covariate balance before matching refinement. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. There are 154 treated units present. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county. (c) Overall ATT estimates and covariate balance before matching refinement and after the application of a caliper. (d) Overall ATT estimates and covariate balance after the application of a caliper, and after matching refinement, the observed balance scores are, on average over the matching window, within the threshold of 0.1, indicating sufficient similarity between the treated and matched counties. On average, for estimates over the outcome window, 141 treated units remain, with 649 matched units.

Extended Data Fig. 7 Estimates for Donald Trump’s rallies, stratified by exposure (without caliper).

(a) Overall ATT estimates and covariate balance before matching refinement, for each stratum. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. The number of treated units in each stratum are on average, over the outcome window. For Treatment, there are 67 treated units; for Degree 1, there are 397 treated units; for Degree 2, there are 788 treated units; for Degree 3, there are 1156 treated units. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county, for each stratum.

Extended Data Fig. 8 Estimates for Donald Trump’s rallies, stratified by exposure (with caliper).

In both panels, the error bars indicate the 95% CIs. (a) Overall ATT estimates and covariate balance before matching refinement, for each stratum. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an election. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. The number of treated units in each stratum are on average, over the outcome window. For Treatment, 58.2 treated units remain; for Degree 1, 363.4 treated units remain; for Degree 2, 713.3 treated units remain; for Degree 3, 1055.1 treated units remain. Respectively, with 251.8, 1679, 3290.9, 4280.4 matches. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county, for each stratum.

Extended Data Fig. 9 Overall estimates for the BLM protests. In each panel, the error bars indicate the 95% CIs.

In each panel, the error bars indicate the 95% CIs. (a) Overall ATT estimates and covariate balance before matching refinement. (The ATTs represent the average difference in the change in death rates, from the day before treatment to 10 to 40 days after an event. The pre-refinement covariate balance for each matching covariate. All covariates are measurements at the county level. The balance score is the average standardized mean difference between the treated and control units, over a matching period from 30 days before to 1 day before treatment. On average, for estimates over the outcome window, 658 treated units are present. (b) Overall ATT estimates and covariate balance after matching refinement, to no more than the five best matches to each treated county. (c) Overall ATT estimates and covariate balance before matching refinement and after the application of a caliper. (d) Overall ATT estimates and covariate balance after the application of a caliper, and after matching refinement, the observed balance scores are, on average over the matching window, within the threshold of 0.1, indicating sufficient similarity between the treated and matched counties. On average, for estimates over the outcome window, 450.5 treated units remain, with 1901.7 matched units.

Extended Data Fig. 10 Political Event sizes.

(a) Overall distribution of event sizes in the data, across event type. Large frequencies for specific values reflect the thresholding procedure used to estimate crowd sizes from different reports (see Methods). (b) Overall distribution of event sizes in the data, across each event type, represented as the percentage of the county population in which the event takes place. (c) Event sizes over the roughly two-year period that constitutes our study horizon, colored by event type. Event sizes are plotted on the natural log scale, labelled on the original scale (persons at event).

Supplementary information

Supplementary Information

Supplementary Discussion, Tables 1–10 and Figs. 1–77.

Reporting Summary

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Feltham, E., Forastiere, L., Alexander, M. et al. Mass gatherings for political expression had no discernible association with the local course of the COVID-19 pandemic in the USA in 2020 and 2021. Nat Hum Behav 7, 1708–1728 (2023). https://doi.org/10.1038/s41562-023-01654-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41562-023-01654-1

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing