Content uploaded by Jackson G. Lu
Author content
All content in this area was uploaded by Jackson G. Lu on Mar 10, 2023
Content may be subject to copyright.
Two Large-Scale Global Studies on COVID-19 Vaccine Hesitancy
Over Time: Culture, Uncertainty Avoidance, and
Vaccine Side-Effect Concerns
Jackson G. Lu
MIT Sloan School of Management, Massachusetts Institute of Technology
This article presents one of the largest and broadest investigations into COVID-19 vaccine hesitancy, a
burning issue that poses a global threat. First, I provide a timely review of the predictors of COVID-19
vaccine hesitancy identified by prior studies. More importantly, I advance a dynamic, cultural psychological
perspective to examine how the cultural dimension of uncertainty avoidance partly explains national
differences in initial vaccine hesitancy. To track global vaccine hesitancy over time, I leveraged a daily
survey of 979,971 individuals in 67 countries/territories (October 2020 to March 2021) and another daily
survey of over 11 million individuals in 244 countries/territories (December 2020 to March 2021). To
increase sample representativeness, both surveys used algorithms to correct for nonresponse bias and
coverage bias. Consistent with my theoretical perspective, people in higher (vs. lower) uncertainty
avoidance cultures had higher vaccine hesitancy initially (late 2020) as a function of greater vaccine
side-effect concerns, but these differences decreased over time as COVID-19 vaccine uptake became
prevalent. These findings were robust after controlling for other cultural dimensions, demographics,
COVID-19 severity, government response stringency, socioeconomic indicators, common vaccine cover-
age, and religiosity. Understanding cultural differences in vaccine hesitancy is important, as delaying
vaccination for even a short period can increase morbidity and mortality.
Keywords: cultural psychology, uncertainty avoidance, COVID-19, vaccine hesitancy, vaccine side-effect
concerns
Supplemental materials: https://doi.org/10.1037/pspa0000320.supp
As of May 2022, the coronavirus disease 2019 (COVID-19)
has caused over 6 million deaths worldwide and still claims
thousands of lives every day (World Health Organization,
n.d). Extensive research shows that “COVID-19 vaccines are
effective at preventing infection, serious illness, and death”
(Centers for Disease Control and Prevention, 2021). Vaccine
hesitancy, defined as “delay in acceptance or refusal of vaccina-
tion despite availability of vaccination services”(MacDonald &
the SAGE Working Group on Vaccine Hesitancy, 2015, p. 4161),
increases morbidity and mortality. To reduce the spread of
COVID-19, it is vital to understand the predictors of COVID-
19 vaccine hesitancy (Dror et al., 2020;Sallam, 2021). To this
end, this article (a) provides a timely review of the predictors of
COVID-19 vaccine hesitancy, (b) introduces a dynamic, cultural
perspective on vaccine hesitancy, and (c) presents one of the
largest and broadest investigations into COVID-19 vaccine hesi-
tancy, thereby offering both theoretical and empirical
contributions.
The Vaccine Hesitancy Determinants Matrix (MacDonald & the
SAGE Working Group on Vaccine Hesitancy, 2015) organizes
determinants of vaccine hesitancy into three categories: (a)
vaccine-specific factors, (b) individual and group factors, and (c)
contextual factors. Using these categories, Table 1 provides a
concise yet systematic review of the predictors of COVID-19
vaccine hesitancy identified by prior studies. In terms of vaccine-
specific factors, vaccine hesitancy tends to be low if COVID-19
vaccines are considered safe (Orangi et al., 2021), priced inexpen-
sively (Wang et al., 2020), or recommended by health care profes-
sionals (Reiter et al., 2020). In terms of individual and group factors,
vaccine hesitancy tends to be high among individuals who are
female (Lazarus et al., 2021), young (Fisher et al., 2020), less
educated (Robertson et al., 2021), low-income (de Figueiredo &
Larson, 2021), less knowledgeable about COVID-19 (Paul et al.,
2021), without a vaccination history (Schwarzinger et al., 2021),
or untrusting of the health care system (Murphy et al., 2021). In
terms of contextual factors, vaccine hesitancy tends to be low if
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This article was published Online First August 4, 2022.
Jackson G. Lu https://orcid.org/0000-0002-0144-9171
I am grateful to Kees van den Bos, John Carroll, Richa Chadha, Jane
Minyan Chen, Yan Chen, Yiming Huang, Peter Jin, Ruisi Li, Zhenyu Liao,
Priya Mehla, Kyra Rodriguez, Yuanyuan Shi, Jianghao Wang, Laura Chan-
glan Wang, JoAnne Yates, and Lu Zhang for their helpful feedback or
assistance. I thank the Massachusetts Institute of Technology COVID-19
research team for conducting Study 1’s survey in collaboration with
Facebook, Johns Hopkins University, the World Health Organization, and
the Global Outbreak Alert and Response Network. I also thank Carnegie
Mellon University and the University of Maryland for conducting Study 2’s
survey in collaboration with Facebook.
Correspondence concerning this article should be addressed to Jackson
G. Lu, MIT Sloan School of Management, Massachusetts Institute of
Technology, 100 Main Street, Cambridge, MA 02142, United States.
Email: lu18@mit.edu
Journal of Personality and Social Psychology:
Attitudes and Social Cognition
© 2022 American Psychological Association 2023, Vol. 124, No. 4, 683–706
ISSN: 0022-3514 https://doi.org/10.1037/pspa0000320
683
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 1
A Review of Predictors of COVID-19 Vaccine Hesitancy (Organized per the Vaccine Hesitancy Determinants Matrix)
Categories Factors Tendencies Example citations
Vaccine-specific factors Risk/benefit Vaccines perceived as risky and ineffective
are associated with higher vaccine
hesitancy.
Fisher et al. (2020),Katoto et al. (2022),Orangi et al. (2021),Paul
et al. (2021),Reiter et al. (2020), and Sherman et al. (2021)
Origin of vaccine In the United States, foreign (vs. domestic)
vaccines are associated with higher
vaccine hesitancy.
Kreps et al. (2020)
Costs Individuals who consider the price of
vaccine important have higher vaccine
hesitancy.
Wang et al. (2020)
Recommendation and/or
attitude of health
care professionals
Individuals whose health care providers
recommend vaccination have lower
vaccine hesitancy.
Reiter et al. (2020),Sherman et al. (2021), and Wang et al. (2020)
Individual and
group factors
Vaccination history Individuals who have received vaccines in
the past have lower vaccine hesitancy.
Caserotti et al. (2021),Fisher et al. (2020),Lackner and Wang
(2021),Paul et al. (2021),Pogue et al. (2020),Schwarzinger
et al. (2021),Sherman et al. (2021), and Wang et al. (2020)
Beliefs and attitudes
about COVID-19
Individuals who perceive themselves to be
less susceptible to COVID-19 have
higher vaccine hesitancy.
Caserotti et al. (2021),Gerretsen et al. (2021),Orangi et al.
(2021),Schwarzinger et al. (2021),Wang et al. (2020), and
Willis et al. (2021)
Knowledge/awareness Individuals who have poor knowledge of
COVID-19 have higher vaccine
hesitancy.
Mewhirter et al. (2022),Paul et al. (2021), and Sherman et al.
(2021)
Trust in the health care
system and government
Individuals who are less trusting of the
health care system and government have
higher vaccine hesitancy.
Fisher et al. (2020),Goodwin et al. (2022),Jennings et al. (2021),
Katoto et al. (2022),Lazarus et al., (2021),Murphy et al.
(2021),Paul et al. (2021), and Sherman et al. (2021)
Education Less educated individuals have higher
vaccine hesitancy.
de Figueiredo and Larson (2021),Fisher et al. (2020),
Khubchandani et al. (2021),Kreps et al. (2020),Lazarus et al.
(2021),Malik et al. (2020),Paul et al. (2021),Robertson et al.
(2021),Savoia et al. (2021), and Schwarzinger et al. (2021)
Gender Women have higher vaccine hesitancy. de Figueiredo and Larson (2021),Fisher et al. (2020),Latkin et al.
(2021),Malik et al. (2020),Murphy et al. (2021),Paul et al.
(2021),Robertson et al. (2021),Sallam et al. (2021),
Schwarzinger et al. (2021), and Wang et al. (2020)
Race/ethnicity Racial minorities (e.g., Black people) have
higher vaccine hesitancy.
Bell et al. (2020),Fisher et al. (2020),Grumbach et al. (2021),
Khubchandani et al. (2021),Latkin et al. (2021),Malik et al.
(2020),Nguyen et al. (2022),Paul et al. (2021),Reiter et al.
(2020),Robertson et al. (2021),Savoia et al. (2021), and Willis
et al. (2021)
Age Younger individuals have higher vaccine
hesitancy.
de Figueiredo and Larson (2021),Fisher et al. (2020),Lackner
and Wang (2021),Latkin et al. (2021),Lazarus et al. (2021),
Malik et al. (2020),Murphy et al. (2021),Robertson et al.
(2021), and Sherman et al. (2021)
Income Low-income individuals have higher
vaccine hesitancy.
Bell et al. (2020),Katoto et al. (2022),Khubchandani et al.
(2021),Lazarus et al. (2021),Murphy et al. (2021),Paul et al.
(2021),Roberts et al. (2022), and Willis et al. (2021)
Contextual factors Communication and media
environment
Exposure to vaccine misinformation and
conspiracy theories is positively
associated with vaccine hesitancy.
Transparent communication about negative
(positive) features of COVID-19 vaccines
increases (decreases) vaccine hesitancy.
For individuals strongly hesitant about
COVID-19 vaccines, providing
information on personal benefits reduces
hesitancy more than information on
collective benefits.
Allington et al. (2021),Fisher et al. (2020),Freeman et al. (2021),
Jennings et al. (2021),Katoto et al. (2022),Loomba et al.
(2021),Murphy et al. (2021),Petersen et al. (2021), and Romer
and Jamieson (2020)
Influential leaders and
immunization
program gatekeepers
Endorsement by the World Health
Organization and Anthony Fauci is
negatively associated with vaccine
hesitancy.
Bokemper et al. (2021),Kaplan and Milstein (2021), and Kreps
et al. (2020)
Politics In the United States, individuals with
conservative political leanings have
higher vaccine hesitancy.
Fridman et al. (2021),Hornsey et al. (2020),Khubchandani et al.
(2021),Latkin et al. (2021),Reiter et al. (2020), and Roberts
et al. (2022)
684 LU
COVID-19 misinformation is low (Loomba et al., 2021) and
vaccines are endorsed by affiliated political parties (Hornsey et
al., 2020), influential leaders (Bokemper et al., 2021), or trustworthy
organizations (e.g., World Health Organization; Kreps et al., 2020).
The present research not only examines several of these factors
in two global studies, but also identifies culture as a novel contextual
factor for vaccine hesitancy. Culture affects individuals profoundly,
as fundamental aspects of everyday life are contextualized within
culture (Kitayama & Cohen, 2010;Lu et al., 2023;Markus &
Kitayama, 2010). Although some studies have examined the role
of cultural variables in the spread of COVID-19 (Salvador et al.,
2020), COVID-19 morbidity and mortality (Gelfand et al., 2021;
Kumar, 2021), and mask use (Lu et al., 2021), little research has
examined how culture influences vaccine hesitancy, a burning issue
that poses a global threat. To fill this knowledge gap, I examine the
cultural dimension of uncertainty avoidance (Hofstede et al., 2010).
As I explain below, I focus on uncertainty avoidance because it is
theoretically pertinent to vaccine side-effect concerns, a key driver
of vaccine hesitancy (Karafillakis et al., 2017;Piot et al., 2019). To
ascertain the unique role of uncertainty avoidance in COVID-19
vaccine hesitancy, my studies control for other cultural dimensions
(e.g., individualism, tightness), demographics, COVID-19 severity,
government response stringency, socioeconomic indicators, common
vaccine coverage, and religiosity.
Methodologically, several constraints have limited the under-
standing of COVID-19 vaccine hesitancy. First, few empirical
studies have examined COVID-19 vaccine hesitancy on a global
scale (de Figueiredo & Larson, 2021;Hou et al., 2021;Lazarus et al.,
2021;World Economic Forum, 2021). For example, one of the
broadest studies on COVID-19 vaccine hesitancy only covered 32
countries/territories (N=26,759; de Figueiredo & Larson, 2021).
Therefore, more research is needed to understand how COVID-19
vaccine hesitancy varies across the world. In particular, a sample
with a large number of countries is needed to statistically test
how the cultural dimension of uncertainty avoidance relates to
vaccine hesitancy. Second, many studies on COVID-19 vaccine
hesitancy relied on convenience samples and failed to emphasize
sample representativeness, so the generalizability of their findings
is unclear. Third, past studies have mostly examined COVID-19
vaccine hesitancy at one or several specific points in time, such that
the observed levels of vaccine hesitancy could have been due to
temporary events (e.g., news that week). Thus, more research is
needed to examine how COVID-19 vaccine hesitancy changes over
a long and continuous span of time.
To address these methodological limitations, I leveraged two
large-scale global studies on COVID-19 vaccine hesitancy. In
collaboration with Facebook, Study 1 surveyed 979,971 individuals
in 67 countries/territories and Study 2 surveyed over 11 million
individuals in 244 countries/territories.
1
To track temporal trends
in vaccine hesitancy, both surveys were conducted daily and
spanned several months (Study 1: October 2020 to March 2021;
Study 2: December 2020 to March 2021). These methodological
strengths enable me to provide a theoretically novel and dynamic
perspective on COVID-19 vaccine hesitancy: People in higher (vs.
lower) uncertainty avoidance countries had higher vaccine hesitancy
initially (late 2020) as a function of greater side-effect concerns, but
these differences decreased over time as vaccine uptake became
prevalent. Thus, I contribute to the psychology of vaccine hesitancy
by identifying side-effect concerns as a mediator and time as a
moderator for the link between uncertainty avoidance and COVID-
19 vaccine hesitancy. These findings have actionable implications
for when and how governments and health organizations should
allocate their limited resources (e.g., focusing resources to address
vaccine side-effect concerns, especially when a vaccine is newly
introduced). By analyzing how people in different cultures reacted
to the pandemic, I spotlight the importance of cultural psychology
in coping with global crises.
A Dynamic, Cultural Perspective on
Vaccine Hesitancy
In cultural psychology, uncertainty avoidance is defined as “the
extent to which the members of a culture feel threatened by ambigu-
ous or unknown situations”(Hofstede et al., 2010, p. 191). Uncer-
tainty avoidance is a cultural dimension in Hofstede’s framework, “an
invaluable framework to explain variation in psychological tenden-
cies across societies”(Lawrie et al., 2020, p. 52). As indicated by
Hofstede’s index, countries differ considerably in uncertainty avoid-
ance (www.hofstede-insights.com/country-comparison).
Although no research has examined the role of uncertainty
avoidance in vaccine hesitancy, prior studies have consistently
shown that people in high uncertainty avoidance cultures tend to
be more hesitant about uncertain situations and unfamiliar things
(Chuang et al., 2022;Huynh, 2020;Kreiser et al., 2010;Van den
Bos et al., 2007,2013). For example, people in higher (vs. lower)
uncertainty avoidance cultures are less likely to adopt new products
(Yeniyurt & Townsend, 2003) and innovations (Van Everdingen &
Waarts, 2003). Relatedly, new products tend to take off more slowly
in higher uncertainty avoidance cultures (Tellis et al., 2003). Frijns
et al. (2013) found that chief executive officers (CEOs) of firms in
higher uncertainty avoidance cultures engaged less in cross-border
or cross-industry takeovers—which involved high risk and
uncertainty—and required higher premiums to pursue such take-
overs. Similarly, Kanagaretnam et al. (2014) found that banks in
higher uncertainty avoidance cultures were less risk-taking and
reported earnings more conservatively. Kozak et al. (2007) found
that travelers from higher uncertainty avoidance countries were
more inclined to change their travel plans if the destination had
elevated risk (e.g., new infectious disease, terrorist attacks). In the
context of COVID-19, Huynh (2020) found that the country-level
uncertainty avoidance index predicted more social distancing
behavior (to avoid infection). Together, these findings point to
the possibility that people in higher (vs. lower) uncertainty avoid-
ance cultures were more hesitant about newly developed COVID-19
vaccines.
Vaccine Side-Effect Concerns as a Mechanism
I propose that people in higher (vs. lower) uncertainty avoidance
cultures may initially have higher vaccine hesitancy because of
greater vaccine side-effect concerns. I focus on vaccine side-effect
concerns as a mediator because extensive research suggests that it is
an important reason why some people are hesitant to be vaccinated,
especially for newly developed vaccines. For example, researchers
identified vaccine side-effect concerns as a key driver of hesitancy
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1
For simplicity, I use “countries”to denote “countries/territories”in the
rest of the article.
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 685
about human papillomavirus (HPV) vaccines, which were approved
recently in 2014 (Karafillakis et al., 2017,2019;Piot et al., 2019).
Upon reviewing 29 articles on HPV vaccine hesitancy, Karafillakis
et al. (2017) noted that the most common concern about HPV
vaccine was safety, particularly given that unknown side effects
might develop long after vaccination. Due to the relative newness of
the HPV vaccine, many people were worried that “there were too
many uncertainties around long-term effectiveness of the vaccine”
(Karafillakis et al., 2017, p. 4843).
Similarly, although scientific studies suggest that COVID-19
vaccines are safe and effective (Centers for Disease Control and
Prevention, 2021), some people may worry about their unknown
side effects, especially because the vaccines were developed and
approved at an unprecedented speed (Kreps et al., 2020;Pogue et al.,
2020). As a fundamental cultural dimension affecting the everyday
life of a given culture, uncertainty avoidance might have influenced
vaccine hesitancy when COVID-19 vaccines were first introduced:
People in higher (vs. lower) uncertainty avoidance countries
might have had higher vaccine hesitancy because they were more
concerned that COVID-19 vaccines would have unknown side
effects, and preferred to wait and see whether early vaccine adopters
would experience any unexpected side effects. In other words,
vaccine side-effect concerns might have mediated the link between
uncertainty avoidance and vaccine hesitancy when COVID-19
vaccines were initially introduced.
A Dynamic Perspective on Vaccine Hesitancy
Beyond the mediation prediction, I advance the literature by
offering a dynamic perspective on the psychology of vaccine
hesitancy. Past research has mostly taken a static perspective
(e.g., Country A was higher than Country B in vaccine hesitancy
at a specific point in time), as most studies were cross-sectional
and unable to capture trends in vaccine hesitancy over a long period
of time. In line with recent research indicating the fluctuating nature
of vaccine hesitancy (de Figueiredo et al., 2020;Larson &
Broniatowski, 2021), I propose that people are sensitive to the
benefits and risks of COVID-19 vaccines, such that their vaccine
hesitancy changes over time.
As theorized above, when COVID-19 vaccines first arrived on
the scene, people in higher (vs. lower) uncertainty avoidance
countries might have had higher vaccine hesitancy because of
greater side-effect concerns. However, as more adopters received
the vaccine, it should have become clear that unknown side effects
are rare. As a result, even people in high uncertainty avoidance
cultures should have become less concerned about vaccine side
effects and thus less hesitant about getting vaccinated. Hence, I
hypothesize an interaction effect between uncertainty avoidance
and time on vaccine hesitancy: People in higher (vs. lower) uncer-
tainty avoidance countries would initially have higher vaccine
hesitancy as a function of greater side-effect concerns, but these
cultural differences would decrease over time as vaccine uptake
became prevalent.
To test my dynamic, cultural perspective on vaccine hesitancy, I
leveraged a daily survey of 979,971 individuals in 67 countries/
territories (October 2020 to March 2021) and another daily survey
of over 11 million individuals in 244 countries/territories (December
2020 to March 2021). To increase sample representativeness, both
surveys carefully applied algorithms to correct for nonresponse bias
and coverage bias.
Study 1: Vaccine Hesitancy in 67 Countries
Analyzing a daily COVID-19 survey conducted in 42 languages
in 67 countries (Table 2), I tested the hypothesized interaction effect
between uncertainty avoidance and time on vaccine hesitancy. I
predicted that people in higher (vs. lower) uncertainty avoidance
countries would have higher vaccine hesitancy initially, but these
cultural differences would decrease over time as vaccine uptake
became prevalent.
Method
Transparency and Openness
The survey was a collaboration between Facebook, the Massa-
chusetts Institute of Technology (MIT), and Johns Hopkins Univer-
sity, and received input from experts at the World Health
Organization and the Global Outbreak Alert and Response Network
(for details, see Collis et al., 2022;Moehring et al., 2022). It was
approved by the institutional review board of MIT (Protocol E-2294).
All participants were at least 18 years old and provided informed
consent. The data do not contain any identifying information. Data
and materials can be accessed at https://dataforgood.facebook.com/
dfg/tools/covid-19-preventative-health-survey#accessdata.
Sampling Methodology
With over 2.85 billion users, Facebook is the most popular social
media platform in the world. Every day, a new sample of adult
users received a survey invitation that appeared in their Facebook
News Feed (see Supplemental Figure S1). To provide geographic
coverage, the survey used stratified random sampling based on
administrative boundaries within countries.
A methodological strength is that the entire survey procedure was
in the user’s default Facebook language. For example, if a person
usually used Facebook in Korean, then both survey invitation and
survey content would automatically be in Korean. The survey was
translated into 42 languages (see Supplemental Table S1).
Weighting Methodology
Based on the “total survey error”framework (Groves & Lyberg,
2010), weights were calculated to minimize errors of representation
and increase sample representativeness in two stages (for details,
see Collis et al., 2022). As explained below, the first stage adjusted
for nonresponse bias (i.e., some users are more likely to respond
to the survey than others), and the second stage further adjusted for
coverage bias (i.e., not everyone in a country has a Facebook
account or uses it actively).
The first stage adjusted for nonresponse bias to make the
sample more representative of the Facebook Active User Base.
Inverse propensity score weighting (IPSW) was applied because
it is a well-established approach that allows for correcting many
covariates simultaneously. To model nonresponse, existing Face-
book user attributes (e.g., age, gender) were used as covariates.
For privacy reasons, the covariates were taken from internal
Facebook data rather than from individual survey responses.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
686 LU
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 2
Study 1: Descriptive Statistics of 67 Countries and Territories Across the Study Period
Country/territory ISO code Uncertainty avoidance (Hofstede) Sample size Mean age (categorical) % Male % Hesitant Vaccine hesitancy mean Vaccine hesitancy SD
Afghanistan AFG 3,130 2.57 93.5% 21.6% 1.32 0.65
Algeria DZA 70 3,242 3.28 77.4% 63.3% 1.98 0.85
Angola AGO 60 2,971 3.31 71.3% 38.2% 1.49 0.68
Argentina ARG 86 35,671 3.94 36.3% 37.5% 1.52 0.73
Australia AUS 51 3,122 4.53 43.5% 31.8% 1.44 0.70
Azerbaijan AZE 88 2,575 3.37 61.9% 65.7% 2.03 0.84
Bangladesh BGD 60 34,457 2.61 81.9% 23.6% 1.35 0.67
Bolivia BOL 87 3,191 3.08 52.2% 36.9% 1.53 0.75
Brazil BRA 76 35,769 3.58 37.8% 24.6% 1.34 0.65
Cambodia KHM 2,876 2.94 76.5% 25.7% 1.34 0.62
Cameroon CMR 3,211 2.92 71.6% 61.7% 2.02 0.89
Canada CAN 48 3,088 4.28 40.7% 35.8% 1.50 0.73
Chile CHL 86 3,336 3.82 42.0% 43.0% 1.61 0.77
Colombia COL 80 36,824 3.38 46.0% 35.1% 1.50 0.74
Ecuador ECU 67 3,295 3.12 49.7% 34.1% 1.47 0.72
Egypt EGY 55 38,722 3.12 70.3% 47.4% 1.70 0.81
Estonia EST 60 2,888 3.38 37.7% 49.9% 1.73 0.81
France FRA 86 36,949 4.21 41.5% 54.2% 1.82 0.84
Georgia GEO 85 3,296 3.54 44.8% 49.8% 1.71 0.80
Germany DEU 65 36,808 4.04 45.8% 38.9% 1.58 0.79
Ghana GHA 65 3,245 2.83 76.6% 38.2% 1.59 0.81
Guatemala GTM 98 3,532 2.96 51.1% 29.2% 1.41 0.69
Honduras HND 50 3,299 3.14 46.3% 31.5% 1.43 0.70
India IND 40 37,396 2.90 80.9% 27.9% 1.41 0.71
Indonesia IDN 48 35,673 3.27 65.6% 44.9% 1.64 0.79
Iraq IRQ 96 3,436 3.19 79.7% 38.0% 1.57 0.79
Italy ITA 75 35,420 4.17 45.4% 27.3% 1.37 0.65
Ivory Coast CIV 3,106 3.15 81.5% 44.0% 1.65 0.81
Jamaica JAM 13 2,780 3.70 34.8% 75.1% 2.11 0.77
Japan JPN 92 35,927 4.85 64.2% 42.9% 1.54 0.69
Kazakhstan KAZ 88 3,844 3.88 45.1% 72.8% 2.19 0.84
Kenya KEN 50 3,114 3.00 69.3% 27.7% 1.42 0.73
Korea (South) KOR 85 3,972 3.98 65.7% 26.1% 1.39 0.70
Malaysia MYS 36 37,415 3.48 58.0% 32.0% 1.42 0.67
Mexico MEX 82 36,621 3.22 46.8% 24.1% 1.32 0.61
Mongolia MNG 2,253 2.90 49.3% 32.9% 1.44 0.68
Morocco MAR 68 3,355 3.29 69.4% 59.0% 1.88 0.83
Mozambique MOZ 44 3,411 3.00 70.6% 37.6% 1.49 0.69
Myanmar MMR 3,561 2.91 72.1% 15.0% 1.21 0.53
Nepal NPL 40 3,262 2.65 76.6% 24.9% 1.35 0.66
Netherlands NLD 53 4,133 4.84 44.0% 39.8% 1.55 0.75
Nigeria NGA 55 36,550 3.19 78.7% 39.5% 1.61 0.82
Pakistan PAK 70 36,847 2.74 79.8% 35.0% 1.53 0.78
Peru PER 87 3,260 3.34 50.1% 35.6% 1.50 0.73
Philippines PHL 44 35,928 3.18 45.9% 47.7% 1.67 0.78
Poland POL 93 39,055 4.04 43.4% 46.9% 1.71 0.83
Portugal PRT 99 3,656 4.01 40.5% 46.3% 1.56 0.67
Romania ROU 90 36,650 3.94 48.8% 45.6% 1.70 0.83
Senegal SEN 55 2,020 3.25 73.9% 61.0% 1.99 0.88
Singapore SGP 8 3,156 3.80 60.9% 34.7% 1.48 0.72
South Africa ZAF 49 3,370 3.65 42.5% 41.5% 1.64 0.82
Spain ESP 86 3,778 4.08 40.3% 53.5% 1.74 0.77
Sri Lanka LKA 45 3,130 3.25 72.7% 32.4% 1.47 0.73
Sudan SDN 3,496 2.83 74.0% 33.9% 1.50 0.76
Taiwan TWN 69 3,725 3.71 54.5% 47.0% 1.69 0.81
Tanzania TZA 50 2,572 2.99 83.6% 36.4% 1.61 0.85
Thailand THA 64 40,167 3.83 54.6% 24.7% 1.36 0.67
Trinidad and Tobago TTO 55 3,442 3.84 40.7% 59.3% 1.85 0.80
Turkey TUR 85 36,834 3.72 74.1% 44.0% 1.65 0.80
Uganda UGA 3,075 2.73 73.3% 27.9% 1.45 0.76
Ukraine UKR 95 3,452 3.53 38.0% 59.2% 1.87 0.82
United Arab Emirates ARE 66 3,371 3.22 67.8% 34.6% 1.51 0.76
United Kingdom GBR 35 32,251 4.16 39.2% 22.0% 1.31 0.64
United States USA 46 31,462 4.49 36.4% 37.2% 1.56 0.78
Uruguay URY 98 3,446 4.12 30.4% 54.3% 1.76 0.79
Venezuela VEN 76 3,440 4.02 50.5% 44.1% 1.61 0.76
Vietnam VNM 30 37,662 2.80 59.4% 18.6% 1.24 0.55
Note. Hofstede uncertainty avoidance scores range from 8 (lowest) to 100 (highest). Age categories: 1 =under 20, 2 =20–30, 3 =31–40, ::: 7=71–80, 8 =over 80. Vaccine hesitancy: “If a
vaccine for COVID-19 becomes available, would you choose to get vaccinated?”(1 =yes, 2 =don’t know, 3 =no). % Hesitant =percentage of people who indicated “no”or “don’t know”.
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 687
Continuous variables (e.g., age) were transformed into categori-
cal variables to ensure that the sample matched their full dis-
tributions rather than only the mean values. In sum, correcting for
nonresponse bias yielded a more representative sample of the
Facebook Active User Base.
The second stage further adjusted for coverage bias to make the
sample more representative of the adult population in each country.
The IPSW output weights from the first stage were used as inputs
for post-stratification, a common survey methodology to correct for
known differences between the sample and the target population
(Little, 1993). Post-stratification was applied using publicly available
benchmarks. In sum, correcting for coverage bias yielded a more
representative sample of the adult population in a given country
(Collis et al., 2022).
Vaccine Hesitancy
Vaccine hesitancy was surveyed daily from October 28, 2020,
to March 29, 2021, which enabled me to track changes in vaccine
hesitancy over time. A total of 979,971 participants who had
not received a COVID-19 vaccine (44% female) responded to
a question: “If a vaccine for COVID-19 becomes available,
would you choose to get vaccinated?”(1 =yes, 2 =don’t
know, 3 =no). As many as 36.7% of the participants indicated
“no”or “don’tknow.”
Uncertainty Avoidance
I sourced Hofstede’s uncertainty avoidance index (www.hofste
de-insights.com/country-comparison), which ranges from 8 (lowest)
to 100 (highest). The index is available for 59 of the 67 countries
in Study 1 (Table 2).
Control Variables
To examine the role of uncertainty avoidance in COVID-19
vaccine hesitancy, I collected a broad set of individual-level and
country-level control variables. For detailed sources of country-
level variables in this article, see Supplemental Table S2.
Other Cultural Dimensions. To ascertain the unique effects
of uncertainty avoidance, I controlled for the other Hofstede cultural
dimensions (Hofstede et al., 2010;Lawrie et al., 2020): individual-
ism (how much a culture prioritizes personal interests over collec-
tive interests), indulgence (how much a culture allows relatively free
gratification of basic and natural human desires related to enjoying
life and having fun), long-term orientation (how much a culture
focuses on future-oriented values), masculinity (how much a culture
emphasizes achievement and material success), and power distance
(how much a culture accepts and expects unequally distributed
power). In addition, I controlled for cultural tightness–looseness
(Gelfand et al., 2011), or the degree to which a culture has “many
strongly enforced rules and little tolerance for deviance”(Harrington
& Gelfand, 2014, p. 7990).
Because the Hofstede indices for individualism, indulgence, long-
term orientation, masculinity, and power distance are unavailable
for over a dozen of the countries in Study 1 and the Gelfand tightness
index is unavailable for 45 of the countries, controlling for these
cultural dimensions in the initial regression models would cause
an unnecessary drop in the number of observations for analysis.
Thus, I controlled for these cultural dimensions only in the final
two models of each regression table (e.g., Models 4 and 5 in
Table 3).
COVID-19 Severity. Individuals might be more willing to get
vaccinated if COVID-19 is rampant in their country. Thus, I
controlled for daily COVID-19 severity for each country. I sourced
daily “new confirmed cases of COVID-19 (7-day smoothed) per
1,000,000 people”from Our World in Data. This variable denotes
the mean number of new COVID-19 cases per million population in
the preceding 7 days and is commonly used in the COVID-19
literature.
Government Response Stringency. I controlled for the government
response stringency index because countries vary in the stringency
of their COVID-19 policies (e.g., mask usage, private gathering
limits, school closures). For each country, I sourced this daily
stringency index from Our World in Data.
Population Density. I controlled for population density because
people living in sparser regions may perceive vaccination as less
necessary. I sourced population density (population per square kilo-
meters) from the United Nations. The variable was skewed, so log
transformation was applied.
GDP per Capita. I controlled for gross domestic product
(GDP) per capita (U.S. dollar) because it could be a confounding
variable simultaneously related to both uncertainty avoidance and
vaccine hesitancy. Research suggests that national wealth may be
related to uncertainty avoidance (Hofstede et al., 2010). Moreover,
people in less wealthy countries may be more hesitant to get
vaccinated because they are worried about the cost of vaccines
(MacDonald & the SAGE Working Group on Vaccine Hesitancy,
2015). Thus, I sourced data on GDP per capita (U.S. dollar) from
Our World in Data. The variable was skewed, so log transforma-
tion was applied.
Common Vaccine Coverage. Countries may differ in the
degree to which their citizens are familiar and comfortable with
vaccines in general. Thus, I controlled for country-level coverage
of one of the most common vaccines worldwide, DTP3 (diphtheria,
tetanus, and pertussis). As the Vaccine Alliance explains, “DTP3
coverage is a standard measure of the strength of immunisation and
health systems.”I sourced the “WHO-UNICEF estimates of DTP3
coverage”from the World Health Organization.
Religiosity. Past research suggests that religiosity may be posi-
tively related to vaccine hesitancy due to the tension between religious
doctrines and science (de Figueiredo et al., 2020;Lane et al., 2018;
Rutjens et al., 2018). Thus, I sourced the religiosity index from the
World Population Review, which indicates the percentage of a coun-
try’s citizens who consider religion important in daily life.
Education Level. Because individuals with higher education
may be more likely to appreciate COVID-19 vaccines (see Table 1),
I controlled for self-reported education (1 =less than primary
school, 2 =primary school, 3 =secondary school, 4 =college/
university, 5 =graduate school).
Age. Prior studies have found that younger individuals tend to
have higher COVID-19 vaccine hesitancy (see Table 1), perhaps
because COVID-19 is less lethal to them (Brodin, 2021). Thus, I
controlled for age (1 =under 20, 2 =20–30, 3 =31–40, ::: 7=71–
80, 8 =over 80).
Gender. Finally, I controlled for self-reported gender (1 =
female, 0 =male) because prior studies have shown that women
tend to have higher COVID-19 vaccine hesitancy (see Table 1).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
688 LU
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 3
Study 1: Multilevel Linear Regressions Predicting Vaccine Hesitancy
Variable
Model 1 Model 2 Model 3 Model 4 Model 5
βSE βSE βSE βSE βSE
Date (numerical) −.038*** (.004) −.041*** (.004) −.056*** (.004) −.056*** (.004) −.017** (.005)
Uncertainty avoidance .074*(.032) .074*(.033) .060 (.037) .087*(.041) .121
†
(.062)
Uncertainty avoidance ×Date −.118*** (.004) −.118*** (.004) −.126*** (.004) −.126*** (.004) −.093*** (.006)
Female .050*** (.001) .050*** (.001) .049*** (.001) .043*** (.001)
Age (categorical) −.078*** (.001) −.078*** (.001) −.081*** (.001) −.114*** (.002)
Education (categorical) −.046*** (.001) −.046*** (.001) −.047*** (.001) −.067*** (.001)
COVID-19 severity .031*** (.002) .031*** (.002) .014*** (.002)
Government response stringency −.060*** (.002) −.060*** (.002) −.036*** (.003)
Population density (log) −.034 (.032) −.052 (.034) .030 (.052)
GDP per capita (log) .007 (.039) .031 (.043) .132 (.124)
Common vaccine coverage .029 (.042) .005 (.043) −.128 (.130)
Religiosity .0001 (.046) .027 (.063) −.106 (.104)
Individualism .030 (.056) .148 (.116)
Indulgence −.032 (.052) −.149 (.088)
Long-term orientation .042 (.055) −.076 (.086)
Masculinity .012 (.035) −.015 (.056)
Power distance .082 (.051) .176
†
(.097)
Tightness .028 (.114)
Observations (person) 955,263 908,256 902,592 879,726 476,391
Number of countries 59 59 58 50 22
Akaike information criterion 2,132,613 2,014,372 2,001,477 1,951,109 1,040,705
Bayesian information criterion 2,132,684 2,014,477 2,001,653 1,951,342 1,040,937
Log likelihood −1,066,301 −1,007,177 −1,000,724 −975,534 −520,331
Note. Standardized regression coefficients are displayed, with standard errors in parentheses. GDP =gross domestic product; SE =standard error. Vaccine hesitancy: “If a vaccine for COVID-19 becomes
available, would you choose to get vaccinated?”(1 =yes, 2 =don’t know, 3 =no). Age categories: 1 =under 20, 2 =20–30, 3 =31–40, ::: 7=71–80, 8 =over 80. Education categories: 1 =less than primary
school, 2 =primary school, 3 =secondary school, 4 =college/university, 5 =graduate school.
†
p<.10. *p<.05. ** p<.01. *** p<.001.
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 689
Results
Descriptive statistics are displayed in Table 2 and bivariate
correlations in Supplemental Table S3. Because participants were
nested within countries, I conducted multilevel analyses to account
for (a) within-country statistical dependence and (b) the fact that
different countries had different sample sizes. To demonstrate the
robustness of uncertainty avoidance’s effects, each multilevel
regression table presents a progression of models, with additional
controls included at each step.
Cultural Differences in Vaccine Hesitancy
Decreased Over Time
As hypothesized, there was a significant Uncertainty Avoidance
×Date interaction effect on vaccine hesitancy in a multilevel linear
regression (Table 3 Model 1: β=−.118, SE =.004, z=−29.04,
p<.001). This effect remained robust after accounting for
individual-level controls (Model 2: β=−.118, SE =.004, z=
−28.56, p<.001), country-level controls (Model 3: β=−.126,
SE =.004, z=−30.30, p<.001), and other cultural dimensions
(Models 4 and 5). As illustrated in Figure 1, people in higher (vs.
lower) uncertainty avoidance countries had higher vaccine hesitancy
initially (October 2020), but these differences decreased over time
(as vaccine uptake became prevalent).
Robustness Checks
To ascertain the reliability of the above results, I conducted a
variety of robustness checks. First, results were robust when I
examined the interaction effect between uncertainty avoidance
and week (instead of between uncertainty avoidance and date)in
multilevel linear regressions—whether without controls (β=−.120,
SE =.004, z=−28.97, p<.001) or with controls (β=−.095, SE =
.006, z=−16.41, p<.001).
Second, I recoded the ordinal measure of vaccine hesitancy into a
binary variable (1 =“no”/“don’t know,”0=“yes”). In multilevel
logistic regressions, the Uncertainty Avoidance ×Date interaction
effect was robust—whether without or with controls (Supplemental
Table S4, all ps<.001).
Third, on March 11, 2021, Denmark and Norway paused delivery
of the Oxford–AstraZeneca vaccine over blood clot concerns, and
some other countries followed suit (The Washington Post, 2021).
Because this event could impact COVID-19 vaccine hesitancy, I
conducted a robustness check by limiting analyses to dates before
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 1
Study 1: Vaccine Hesitancy by High Versus Low Uncertainty Avoidance Countries (from October 28, 2020, to
March 29, 2021)
20%
30%
40%
50%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Week
% Hesitant People
High Uncertainty Avoidance Countries
Low Uncertainty Avoidance Countries
Note. Error bars indicate standard errors, which are small because of the large sample size. Based on the median uncertainty
avoidance score (i.e., 66), 51% of the countries are categorized as “high uncertainty avoidance countries”and 49% as “low
uncertainty avoidance countries.”Vaccine hesitancy: “If a vaccine for COVID-19 becomes available, would you choose to
get vaccinated?”(1 =yes, 2 =don’t know, 3 =no). Higher scores indicate higher vaccine hesitancy. % Hesitant people =
Percentage of people who indicated “no”or “don’t know”in a given week. See the online article for the color version of this
figure.
690 LU
March 11, 2021: The Uncertainty Avoidance ×Date interaction
effect on vaccine hesitancy was robust—whether without controls
(β=−.128, SE =.004, z=−30.12, p<.001) or with controls (β=
−.110, SE =.006, z=−18.19, p<.001).
Fourth, besides the above analyses wherein the unit of analysis
is the individual (979,971 individuals), I also conducted analyses
by aggregating individual responses to the country level for
each day (i.e., 67 countries over 153 days): The Uncertainty
Avoidance ×Date interaction effect on vaccine hesitancy was
still robust—whether without controls (β=−.401, SE =.037,
z=−10.74, p<.001) or with controls (β=−.384, SE =.050, z=
−7.74, p<.001).
Other Predictors of Vaccine Hesitancy
As illustrated in Figure 2 and detailed in Table 3, women, younger
individuals, and less educated individuals had higher vaccine
hesitancy on average (all ps<.001). These results are consistent
with prior studies on vaccine hesitancy (Fisher et al., 2020;
Robertson et al., 2021).
Moreover, people in countries whose governments responded
more stringently to COVID-19 had lower vaccine hesitancy on
average (Table 3). This finding suggests that governmental
response to COVID-19 might relate to people’s attitudes toward
vaccine uptake.
Discussion
By analyzing a daily survey of 979,971 participants in 67 countries,
Study 1 supported my dynamic, cultural perspective on vaccine
hesitancy. Consistent with the hypothesized interaction effect between
uncertainty avoidance and time, people in higher (vs. lower) uncer-
tainty avoidance countries had higher vaccine hesitancy initially, but
these differences decreased over time (as vaccine uptake became
prevalent).
Study 2: Vaccine Hesitancy in 244 Countries
Study 2 had two important aims. First, I tested whether Study 1’s
findings were replicable in an even larger and broader survey
of over 11 million participants in all 244 countries (Table 4)
where Facebook is available. Study 2 enabled me to leverage
Hofstede’s uncertainty avoidance index more fully because the
survey covered 115 of all 118 countries for which Hofstede’s
index is available.
2
Second, in line with my theoretical perspec-
tive, I tested whether vaccine side-effect concerns mediated the
interaction effect between uncertainty avoidance and time on
vaccine hesitancy.
Method
Transparency and Openness
I analyzed data from the COVID-19 Trends and Impact Survey
(for details, see Astley et al., 2021;Barkay et al., 2020;Kreuter
et al., 2020;Salomon et al., 2021), which was approved by the
institutional review boards of Carnegie Mellon University
(Protocol 2020_00000162) and University of Maryland (Proto-
col 1587016-7). All participants were at least 18 years old and
provided informed consent. The data do not contain any identifying
information. Data and materials can be accessed at https://data
forgood.facebook.com/dfg/tools/covid-19-trends-and-impact-surve
y#accessdata.
As in Study 1, Facebook applied algorithms to correct for nonre-
sponse bias and coverage bias to increase sample representativeness. The
survey was translated into 57 languages (see Supplemental Table S5).
Vaccine Hesitancy
From December 21, 2020, to March 1, 2021, a total of 11,123,364
participants (55% female) who had not received a COVID-19
vaccine answered a question about vaccine hesitancy: “If a vaccine
to prevent COVID-19 were offered to you today, would you choose
to get vaccinated?”(1 =yes, definitely, 2 =yes, probably, 3 =no,
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 2
Study 1: Percentage of Hesitant Individuals by Gender, Age, and
Education Level
33.3%
40.6%
0%
10%
20%
30%
40%
50%
Men Women
% Hesitant People
37% 38.2% 38.4% 37.2% 35.3%
31.1%
27.1% 28.8%
0%
10%
20%
30%
40%
50%
< 20 20 to 30 31 to 40 41 to 50 51 to 60 61 to 70 71 to 80 > 80
Age
% Hesitant People
39% 39.4%
35.5% 33%
0%
10%
20%
30%
40%
50%
<= Primary school Secondary school College/University Graduate school
Education
% Hesitant People
Note. Error bars indicate standard errors, which are small because of the
large sample size. Vaccine hesitancy: “If a vaccine for COVID-19 becomes
available, would you choose to get vaccinated?”(1 =yes, 2 =don’t know,
3=no). Higher scores indicate higher vaccine hesitancy. % hesitant people =
Percentage of people who indicated “no”or “don’t know”. See the online
article for the color version of this figure.
2
The only three places missing are mainland China, Iran, and Syria
because Facebook is unavailable in those places.
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 691
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4
Study 2: Descriptive Statistics of 244 Countries and Territories Across the Study Period
Country/territory ISO code
Uncertainty
avoidance
(Hofstede) Sample size
Mean age
(categorical) % Male % Hesitant
Vaccine
hesitancy
mean
Vaccine
hesitancy
SD
Side-effect
concerns
mean
Side-effect
concerns SD
Afghanistan AFG 8,167 2.57 89.0% 35.5% 2.18 1.21 2.60 1.19
Aland Islands ALA 934 3.63 55.1% 27.1% 1.92 1.12 2.33 1.14
Albania ALB 70 5,861 3.00 58.5% 39.2% 2.27 1.10 2.72 1.08
Algeria DZA 70 25,175 3.14 74.2% 59.5% 2.75 1.12 2.70 1.16
American Samoa ASM 1,145 3.18 65.6% 48.7% 2.54 1.24 2.65 1.30
Andorra AND 1,515 3.43 53.9% 46.6% 2.47 1.23 2.75 1.24
Angola AGO 60 4,116 2.98 76.9% 33.7% 2.16 1.10 2.98 1.09
Anguilla AIA 376 3.71 63.6% 47.9% 2.50 1.25 2.65 1.27
Antarctica ATA 405 3.99 61.1% 63.5% 2.92 1.25 2.58 1.33
Antigua and Barbuda ATG 301 3.33 58.7% 44.2% 2.44 1.26 2.63 1.31
Argentina ARG 86 189,367 3.43 37.6% 26.5% 1.95 1.00 2.78 1.10
Armenia ARM 88 2,661 3.26 51.6% 58.1% 2.78 1.09 2.87 1.08
Aruba ABW 419 2.90 76.5% 43.9% 2.40 1.22 2.80 1.18
Australia AUS 51 112,001 3.66 39.7% 22.0% 1.81 0.95 2.44 1.06
Austria AUT 70 54,336 3.72 47.9% 28.3% 1.93 1.12 2.34 0.97
Azerbaijan AZE 88 5,906 2.89 64.1% 56.1% 2.68 1.06 2.92 1.10
Bahamas BHS 267 3.68 57.5% 47.9% 2.48 1.25 2.55 1.27
Bahrain BHR 348 3.04 64.4% 39.1% 2.24 1.21 2.53 1.26
Bangladesh BGD 60 31,194 2.06 85.2% 26.8% 1.92 1.01 2.81 1.12
Barbados BRB 119 3.95 70.5% 50.4% 2.55 1.21 2.51 1.31
Belarus BLR 95 16,341 3.46 39.2% 58.6% 2.83 1.09 3.16 0.93
Belgium BEL 94 54,813 3.74 45.8% 25.1% 1.87 1.01 2.37 1.02
Belize BLZ 280 3.58 40.8% 31.4% 2.04 1.16 2.58 1.25
Benin BEN 1,523 2.50 85.6% 47.5% 2.46 1.18 3.04 1.14
Bermuda BMU 119 3.87 68.8% 52.9% 2.63 1.29 2.61 1.33
Bhutan BTN 28 104 2.94 68.1% 43.3% 2.38 1.23 2.60 1.35
Bolivia BOL 87 47,067 2.70 52.3% 26.6% 2.01 0.97 3.04 1.02
Bonaire, Sint Eustatius and Saba BES 83 3.70 57.5% 34.9% 2.19 1.22 2.32 1.23
Bosnia and Herzegovina BIH 87 8,918 3.17 51.5% 43.0% 2.36 1.08 2.51 1.12
Botswana BWA 147 3.99 69.0% 59.9% 2.79 1.31 2.71 1.35
Bouvet Island BVT 30 4.44 53.3% 40.0% 2.33 1.21 3.00 1.26
Brazil BRA 76 877,735 3.00 39.9% 16.9% 1.62 0.91 2.26 1.14
British Indian Ocean Territory IOT 150 3.39 61.1% 34.7% 2.07 1.18 2.51 1.31
Brunei BRN 152 3.00 71.4% 33.6% 2.08 1.18 2.60 1.21
Bulgaria BGR 85 31,238 3.82 44.7% 50.3% 2.52 1.16 2.61 1.08
Burkina Faso BFA 55 2,588 2.86 83.9% 48.1% 2.50 1.14 2.95 1.14
Burundi BDI 85 3.71 70.7% 43.5% 2.53 1.26 2.83 1.30
Cambodia KHM 3,590 2.74 74.3% 25.9% 1.98 0.95 2.90 1.05
Cameroon CMR 4,376 2.52 75.5% 58.5% 2.74 1.13 3.22 1.10
Canada CAN 48 181,312 3.75 39.2% 20.1% 1.71 0.97 2.33 1.06
Cape Verde CPV 40 83 3.57 52.9% 43.4% 2.35 1.25 2.85 1.17
Cayman Islands CYM 76 4.10 67.5% 47.4% 2.41 1.25 2.42 1.26
Central African Republic CAF 89 3.28 56.0% 42.7% 2.34 1.22 2.56 1.29
Chad TCD 97 3.44 83.0% 35.1% 2.11 1.12 2.63 1.12
(table continues)
692 LU
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4 (continued)
Country/territory ISO code
Uncertainty
avoidance
(Hofstede) Sample size
Mean age
(categorical) % Male % Hesitant
Vaccine
hesitancy
mean
Vaccine
hesitancy
SD
Side-effect
concerns
mean
Side-effect
concerns SD
Chile CHL 86 88,824 3.32 41.3% 24.2% 1.87 1.00 2.91 1.10
Christmas Island CXR 84 3.79 62.9% 41.7% 2.30 1.23 2.70 1.16
Cocos (Keeling) Islands CCK 50 4.03 88.5% 36.0% 2.26 1.17 2.85 1.18
Colombia COL 80 182,595 2.90 47.5% 22.3% 1.86 0.96 2.96 1.09
Comoros COM 45 3.68 56.5% 40.0% 2.38 1.25 2.67 1.11
Congo, Democratic Republic COD 2,582 2.95 87.9% 53.9% 2.65 1.18 3.08 1.13
Congo, Republic COG 148 3.08 61.0% 58.8% 2.77 1.19 2.97 1.24
Cook Islands COK 33 4.25 53.3% 36.4% 2.27 1.23 2.75 1.49
Costa Rica CRI 86 31,245 3.15 44.9% 15.6% 1.63 0.88 2.57 1.11
Croatia HRV 80 23,755 3.58 41.7% 40.4% 2.25 1.08 2.36 1.08
Curaçao CUW 47 4.09 48.4% 25.5% 1.98 1.03 3.08 1.31
Cyprus CYP 256 3.02 59.2% 39.1% 2.22 1.08 2.49 1.10
Czech Republic CZE 74 71,730 3.60 43.1% 34.2% 2.10 1.10 2.13 0.99
Denmark DNK 23 94,271 4.50 40.2% 7.4% 1.34 0.69 2.08 0.98
Djibouti DJI 71 3.64 61.8% 45.1% 2.41 1.30 2.53 1.34
Dominica DMA 284 2.94 46.0% 40.8% 2.39 1.12 3.14 1.17
Dominican Republic DOM 45 25,516 2.99 49.2% 34.5% 2.16 1.05 3.10 1.08
Ecuador ECU 67 66,354 2.64 50.7% 27.3% 2.00 1.00 3.07 1.03
Egypt EGY 55 83,867 2.45 69.6% 39.9% 2.27 1.08 2.91 1.05
El Salvador SLV 94 29,866 2.74 49.6% 20.6% 1.85 0.92 2.98 1.05
Equatorial Guinea GNQ 63 3.97 40.6% 38.1% 2.21 1.14 3.10 1.18
Eritrea ERI 84 3.13 63.8% 36.9% 2.23 1.25 3.10 1.06
Estonia EST 60 181 3.34 46.9% 32.0% 2.07 1.09 2.75 1.09
Eswatini SWZ 80 3.51 60.0% 36.2% 2.20 1.12 3.13 0.99
Ethiopia ETH 55 6,912 2.64 90.9% 26.0% 1.90 1.08 2.88 1.07
Falkland Islands FLK 47 3.43 45.5% 42.6% 2.32 1.18 2.57 1.44
Faroe Islands FRO 46 3.12 70.4% 37.0% 2.24 1.16 2.57 1.31
Federated States
of Micronesia
FSM 42 4.15 50.0% 47.6% 2.57 1.27 3.27 1.03
Fiji FJI 48 104 3.98 70.8% 40.4% 2.26 1.28 2.75 1.31
Finland FIN 59 26,164 3.79 39.9% 14.2% 1.55 0.86 2.17 0.97
France FRA 86 307,792 3.75 42.0% 39.7% 2.25 1.10 2.57 1.09
French Guiana GUF 105 3.07 50.7% 31.4% 2.16 1.12 2.33 1.12
French Polynesia PYF 41 3.18 47.4% 51.2% 2.56 1.12 3.05 1.00
French Southern Territories ATF 65 4.69 66.7% 46.2% 2.34 1.14 2.97 1.16
Gabon GAB 210 4.75 43.5% 38.1% 2.25 1.03 2.88 1.09
Gambia GMB 62 3.78 56.2% 41.9% 2.29 1.18 3.11 1.03
Georgia GEO 85 135 3.47 50.0% 39.3% 2.37 1.20 2.92 1.16
Germany DEU 65 330,406 3.55 45.5% 25.9% 1.88 1.05 2.44 0.93
Ghana GHA 65 15,076 2.46 76.1% 37.9% 2.24 1.11 3.42 0.98
Gibraltar GIB 69 3.67 54.3% 40.6% 2.20 1.28 2.94 1.09
Greece GRC 100 55,374 3.61 52.0% 25.2% 1.90 1.00 2.46 0.99
Greenland GRL 79 3.83 58.5% 35.4% 2.15 1.17 2.57 1.24
Grenada GRD 47 3.79 38.5% 34.0% 2.13 1.19 2.95 1.20
Guadeloupe GLP 178 3.98 38.9% 39.3% 2.24 1.09 2.70 1.05
Guam GUM 54 3.23 69.4% 37.0% 2.19 1.13 2.59 1.14
(table continues)
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 693
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4 (continued)
Country/territory ISO code
Uncertainty
avoidance
(Hofstede) Sample size
Mean age
(categorical) % Male % Hesitant
Vaccine
hesitancy
mean
Vaccine
hesitancy
SD
Side-effect
concerns
mean
Side-effect
concerns SD
Guatemala GTM 98 27,480 2.54 54.6% 21.4% 1.86 0.96 2.98 1.05
Guernsey GGY 36 3.46 54.2% 19.4% 1.81 0.89 2.61 0.98
Guinea GIN 1,686 2.49 91.0% 29.5% 2.04 1.10 2.62 1.21
Guinea-Bissau GNB 37 3.53 53.3% 48.6% 2.54 1.22 2.71 1.31
Guyana GUY 45 3.29 50.0% 24.4% 1.98 1.03 3.00 0.94
Haiti HTI 2,650 2.60 74.7% 57.0% 2.74 1.11 3.04 1.16
Heard Island and
McDonald Islands
HMD 42 4.35 70.6% 54.8% 2.76 1.23 2.28 1.32
Honduras HND 50 22,948 2.75 48.0% 20.1% 1.82 0.94 3.06 1.05
Hong Kong HKG 29 15,664 3.51 62.4% 37.8% 2.21 0.98 2.79 1.02
Hungary HUN 82 114,933 4.12 44.3% 29.6% 1.99 1.09 2.42 1.05
Iceland ISL 50 128 3.49 57.1% 32.0% 2.09 1.21 2.58 1.17
India IND 40 237,122 2.46 83.1% 29.7% 1.99 1.06 2.57 1.16
Indonesia IDN 48 142,399 2.63 65.2% 26.6% 2.00 0.98 2.61 1.07
Iraq IRQ 96 36,634 2.58 79.4% 42.9% 2.33 1.16 2.63 1.10
Ireland IRL 35 48,862 3.77 38.0% 13.8% 1.53 0.86 2.40 1.04
Isle of Man IMN 45 3.39 50.0% 40.0% 2.27 1.18 2.45 1.32
Israel ISR 81 30,467 3.73 47.6% 27.3% 1.88 1.07 2.74 1.12
Italy ITA 75 331,565 3.62 45.5% 13.6% 1.57 0.83 2.57 0.93
Ivory Coast CIV 6,310 2.95 81.0% 42.3% 2.36 1.14 2.93 1.15
Jamaica JAM 13 125 3.80 67.1% 38.4% 2.34 1.12 2.88 1.19
Japan JPN 92 592,328 4.32 63.7% 23.8% 2.06 0.73 2.97 0.75
Jersey JEY 62 3.61 64.3% 30.6% 2.10 1.08 2.96 1.02
Jordan JOR 65 22,137 2.95 68.6% 53.6% 2.59 1.17 2.72 1.12
Kazakhstan KAZ 88 10,787 3.45 46.7% 66.4% 3.00 1.06 3.08 1.08
Kenya KEN 50 28,108 2.70 72.3% 27.2% 1.94 1.06 3.32 1.01
Kiribati KIR 29 3.88 53.3% 34.5% 2.14 1.06 3.00 1.22
Korea (South) KOR 85 49,476 3.19 66.1% 23.6% 1.89 0.86 2.77 0.93
Kosovo KOS 140 3.20 51.2% 32.1% 2.12 1.12 3.00 0.95
Kuwait KWT 80 5,129 3.06 76.3% 32.4% 2.07 1.06 2.73 1.09
Kyrgyzstan KGZ 4,060 3.16 47.3% 54.8% 2.71 1.08 2.92 1.09
Laos LAO 2,289 2.48 74.0% 22.8% 1.90 0.94 2.73 1.05
Latvia LVA 63 147 3.23 53.6% 43.5% 2.32 1.27 2.59 1.20
Lebanon LBN 57 16,008 2.94 61.2% 37.7% 2.19 1.08 2.77 1.07
Lesotho LSO 54 3.09 53.3% 44.4% 2.37 1.20 3.27 0.94
Liberia LBR 91 3.22 79.2% 38.5% 2.16 1.21 2.68 1.30
Libya LBY 67 15,005 2.83 75.4% 32.1% 2.04 1.11 2.55 1.09
Liechtenstein LIE 66 3.59 43.9% 51.5% 2.44 1.23 2.28 1.03
Lithuania LTU 65 125 2.83 60.4% 41.6% 2.28 1.23 2.40 1.12
Luxembourg LUX 70 259 3.29 52.0% 30.5% 2.01 1.12 2.42 1.04
Macao MAC 155 4.72 38.6% 36.1% 2.19 1.12 2.48 1.16
Madagascar MDG 1,433 3.38 62.3% 48.8% 2.50 1.08 2.96 1.03
Malawi MWI 50 103 3.03 74.2% 36.9% 2.32 1.22 3.06 1.20
Malaysia MYS 36 63,759 2.77 55.7% 25.0% 1.96 0.91 3.12 0.97
Maldives MDV 141 3.25 50.6% 34.0% 2.16 1.20 2.40 1.18
Mali MLI 2,769 2.64 88.2% 40.4% 2.30 1.16 2.80 1.19
(table continues)
694 LU
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4 (continued)
Country/territory ISO code
Uncertainty
avoidance
(Hofstede) Sample size
Mean age
(categorical) % Male % Hesitant
Vaccine
hesitancy
mean
Vaccine
hesitancy
SD
Side-effect
concerns
mean
Side-effect
concerns SD
Malta MLT 96 132 3.23 65.2% 34.1% 2.15 1.19 2.43 1.19
Marshall Islands MHL 18 4.33 50.0% 33.3% 2.11 1.32 3.40 1.34
Martinique MTQ 118 3.61 51.2% 41.5% 2.32 1.14 2.71 1.13
Mauritania MRT 905 2.66 87.8% 40.2% 2.27 1.19 2.66 1.16
Mauritius MUS 63 3.12 50.0% 23.8% 1.95 0.96 2.69 1.02
Mayotte MYT 67 3.71 44.2% 44.8% 2.28 1.13 2.78 1.13
Mexico MEX 82 655,970 2.72 47.0% 11.6% 1.55 0.81 2.79 1.03
Moldova MDA 95 8,204 3.08 42.5% 52.7% 2.64 1.13 3.09 1.01
Monaco MCO 64 3.77 50.0% 32.8% 2.14 1.10 2.80 1.40
Mongolia MNG 67 3.58 58.8% 53.7% 2.69 1.25 2.62 1.23
Montenegro MNE 90 85 3.08 54.2% 58.8% 2.75 1.13 2.94 1.06
Montserrat MSR 23 3.62 61.5% 52.2% 2.39 1.12 3.75 0.46
Morocco MAR 68 17,414 2.94 72.7% 29.4% 2.02 1.05 2.47 1.11
Mozambique MOZ 44 7,056 2.60 71.8% 28.1% 2.04 0.96 3.41 0.89
Myanmar MMR 15,379 2.55 69.3% 15.2% 1.64 0.88 2.59 1.01
Namibia NAM 45 56 4.29 63.2% 33.9% 2.11 1.14 2.39 1.20
Nauru NRU 31 4.13 46.2% 35.5% 2.16 1.37 2.40 1.24
Nepal NPL 40 17,421 2.08 74.9% 28.6% 2.00 1.00 2.86 1.12
Netherlands NLD 53 124,819 4.25 46.4% 17.9% 1.65 0.96 2.18 0.88
New Caledonia NCL 30 3.06 63.6% 36.7% 2.17 1.12 2.58 1.24
New Zealand NZL 49 43,997 3.96 36.2% 24.2% 1.85 0.95 2.48 1.06
Nicaragua NIC 16,204 2.83 50.7% 21.2% 1.85 0.94 3.04 1.05
Niger NER 72 2.95 70.3% 51.4% 2.43 1.31 2.81 1.21
Nigeria NGA 55 41,029 2.76 78.2% 35.3% 2.17 1.12 3.16 1.11
Niue NIU 13 3.25 25.0% 38.5% 2.31 1.18
Norfolk Island NFK 18 3.30 12.5% 27.8% 2.28 1.07 2.89 1.36
North Macedonia MKD 87 167 4.37 40.6% 42.5% 2.26 1.21 2.79 1.22
Northern Mariana Islands MNP 25 4.15 83.3% 24.0% 1.88 1.05 3.00 1.15
Norway NOR 50 86,243 4.46 43.1% 12.2% 1.52 0.80 1.98 0.89
Oman OMN 2,539 3.00 75.7% 28.9% 2.00 1.06 2.61 1.11
Pakistan PAK 70 40,191 2.33 81.0% 34.9% 2.13 1.12 2.73 1.14
Palau PLW 57 4.46 29.6% 43.9% 2.28 1.18 2.95 1.13
Palestine PSE 8,663 2.82 72.0% 47.2% 2.44 1.15 2.76 1.07
Panama PAN 86 17,227 3.27 42.3% 22.6% 1.85 0.96 2.91 1.08
Papua New Guinea PNG 56 3.06 79.3% 32.1% 2.07 1.14 2.30 1.26
Paraguay PRY 85 17,570 2.75 46.0% 32.6% 2.13 1.04 2.75 1.10
Peru PER 87 91,543 2.81 51.2% 22.8% 1.88 0.97 2.96 1.03
Philippines PHL 44 104,743 2.49 46.6% 36.8% 2.22 1.00 3.35 0.92
Pitcairn Islands PCN 21 4.08 63.6% 42.9% 2.43 1.36
Poland POL 93 100,561 3.30 49.6% 31.6% 2.02 1.12 2.56 1.03
Portugal PRT 99 123,966 3.49 42.8% 16.2% 1.67 0.85 2.77 0.92
Puerto Rico PRI 38 35,815 4.15 35.2% 15.7% 1.58 0.90 2.80 1.13
Qatar QAT 80 4,934 2.87 70.7% 23.5% 1.83 0.99 2.68 1.11
Réunion REU 179 3.72 52.0% 43.0% 2.32 1.12 2.61 1.12
Romania ROU 90 78,944 3.38 50.3% 30.1% 1.99 1.12 2.62 1.10
Russia RUS 95 86,014 3.65 42.6% 58.0% 2.83 1.14 2.94 1.06
(table continues)
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 695
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4 (continued)
Country/territory ISO code
Uncertainty
avoidance
(Hofstede) Sample size
Mean age
(categorical) % Male % Hesitant
Vaccine
hesitancy
mean
Vaccine
hesitancy
SD
Side-effect
concerns
mean
Side-effect
concerns SD
Rwanda RWA 69 3.12 86.1% 40.6% 2.19 1.22 2.97 1.26
Saint Barthélemy BLM 49 4.97 40.0% 34.7% 2.14 1.31 2.80 1.21
Saint Helena SHN 19 3.44 37.5% 26.3% 2.00 1.20 2.90 1.10
Saint Kitts and Nevis KNA 18 3.11 25.0% 33.3% 2.00 1.19 2.60 1.34
Saint Lucia LCA 27 3.36 46.2% 48.1% 2.37 1.24 2.67 1.15
Saint Martin MAF 53 3.94 58.8% 45.3% 2.32 1.12 2.57 1.14
Saint Pierre and Miquelon SPM 13 3.75 57.1% 30.8% 1.77 1.24 2.60 1.52
Saint Vincent and
the Grenadines
VCT 11 3.40 60.0% 18.2% 1.45 1.04 2.33 1.53
Samoa WSM 33 2.31 46.2% 39.4% 2.06 1.12 2.88 1.09
San Marino SMR 91 3.49 55.0% 34.1% 2.11 1.17 2.75 1.22
Sao Tome and Principe STP 70 19 3.73 72.7% 26.3% 2.05 1.27 2.56 1.51
Saudi Arabia SAU 64 23,784 3.02 82.3% 29.7% 1.99 1.06 2.56 1.14
Senegal SEN 55 2,623 3.04 70.9% 46.5% 2.46 1.15 2.84 1.17
Serbia SRB 92 28,054 3.47 46.7% 43.5% 2.34 1.11 2.41 1.12
Seychelles SYC 26 3.39 56.2% 34.6% 2.08 1.13 2.89 1.17
Sierra Leone SLE 50 34 3.48 61.9% 23.5% 2.06 1.13 3.00 1.10
Singapore SGP 8 11,897 3.20 59.1% 21.0% 1.82 0.88 2.80 1.03
Sint Maarten SXM 8 4.67 16.7% 25.0% 2.12 1.25
Slovakia SVK 51 40,645 3.53 47.5% 25.5% 1.88 1.05 2.20 1.01
Slovenia SVN 88 15,173 3.96 45.8% 34.3% 2.09 1.11 2.61 1.00
Solomon Islands SLB 18 3.00 28.6% 44.4% 2.28 1.32 2.17 1.47
Somalia SOM 122 2.58 85.9% 50.8% 2.51 1.28 2.76 1.25
South Africa ZAF 49 79,623 3.17 41.5% 37.1% 2.19 1.12 3.04 1.09
South Georgia and the
South Sandwich Islands
SGS 9 5.00 40.0% 33.3% 2.11 1.27 3.00 1.00
South Sudan SSD 203 2.53 70.8% 24.6% 1.87 1.08 2.39 1.13
Spain ESP 86 211,431 3.59 41.8% 15.7% 1.65 0.85 2.70 1.04
Sri Lanka LKA 45 10,075 2.84 72.2% 25.6% 1.92 1.00 3.07 1.05
Sudan SDN 13,301 2.41 76.6% 33.0% 2.07 1.16 2.55 1.17
Suriname SUR 92 34 3.84 66.7% 23.5% 1.91 0.97 2.53 1.18
Svalbard and Jan Mayen SJM 31 4.42 43.5% 22.6% 1.74 1.03 2.07 0.88
Sweden SWE 29 210,819 4.49 41.6% 12.2% 1.50 0.80 2.18 1.00
Switzerland CHE 58 50,342 3.97 45.1% 33.9% 2.09 1.09 2.46 1.00
Taiwan TWN 69 137,516 3.00 53.8% 33.3% 2.20 0.81 2.85 0.95
Tajikistan TJK 64 3.08 82.9% 29.7% 2.12 1.08 2.44 1.32
Tanzania TZA 50 5,207 2.72 86.2% 38.5% 2.25 1.18 2.96 1.18
Thailand THA 64 163,930 3.05 52.6% 17.8% 1.80 0.85 2.81 0.99
Timor-Leste TLS 87 3.10 70.2% 26.4% 2.05 1.11 2.71 1.17
Togo TGO 95 2.96 72.0% 46.3% 2.35 1.27 2.86 1.20
Tokelau TKL 18 4.60 100.0% 27.8% 2.11 1.18 2.80 1.10
Tonga TON 46 3.14 58.8% 21.7% 1.91 0.98 2.67 1.11
Trinidad and Tobago TTO 55 102 3.43 55.6% 33.3% 2.08 1.07 2.94 1.21
Tunisia TUN 75 24,517 3.23 60.2% 50.9% 2.54 1.10 2.93 1.04
Turkey TUR 85 114,833 3.31 70.5% 33.9% 2.14 1.03 2.78 1.02
Turks and Caicos Islands TCA 62 4.21 63.4% 40.3% 2.23 1.09 2.47 1.22
(table continues)
696 LU
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4 (continued)
Country/territory ISO code
Uncertainty
avoidance
(Hofstede) Sample size
Mean age
(categorical) % Male % Hesitant
Vaccine
hesitancy
mean
Vaccine
hesitancy
SD
Side-effect
concerns
mean
Side-effect
concerns SD
Tuvalu TUV 39 2.57 87.0% 35.9% 2.13 1.24 2.64 1.26
Uganda UGA 233 3.11 68.5% 27.9% 1.93 1.12 3.13 1.13
Ukraine UKR 95 105,608 3.33 40.5% 53.0% 2.71 1.12 3.11 0.97
United Arab Emirates ARE 66 18,684 2.84 62.4% 26.2% 1.91 0.99 2.78 1.10
United Kingdom GBR 35 197,069 3.39 38.2% 13.0% 1.50 0.84 2.14 1.03
United States USA 46 2,658,589 4.15 32.9% 27.9% 1.90 1.10 2.49 1.10
United States Minor
Outlying Islands
UMI 58 2.78 46.9% 36.2% 2.19 1.16 2.58 1.32
Uruguay URY 98 44,334 3.74 31.3% 34.4% 2.16 1.02 2.97 1.04
Uzbekistan UZB 6,043 3.21 56.8% 52.1% 2.65 1.06 2.84 1.11
Vanuatu VUT 44 3.97 66.7% 29.5% 1.77 1.08 2.90 1.18
Vatican City VAT 49 3.81 71.4% 59.2% 2.80 1.19 2.33 1.30
Venezuela VEN 76 53,700 3.64 52.2% 28.0% 2.02 1.00 3.18 1.03
Vietnam VNM 30 184,100 2.12 59.9% 13.9% 1.72 0.80 2.64 1.03
Virgin Islands, U.K. VGB 121 3.80 39.5% 33.1% 2.02 1.15 2.62 1.28
Virgin Islands, U.S. VIR 83 3.40 69.0% 36.1% 2.17 1.23 2.88 1.19
Wallis and Futuna WLF 58 3.90 58.6% 32.8% 2.21 1.18 2.52 1.34
Western Sahara ESH 94 3.67 68.5% 51.1% 2.54 1.25 3.00 1.21
Yemen YEM 5,325 2.51 90.2% 45.2% 2.38 1.24 2.25 1.19
Zambia ZMB 50 189 2.85 69.3% 49.7% 2.52 1.23 2.87 1.24
Zimbabwe ZWE 614 3.14 62.5% 44.1% 2.42 1.20 3.01 1.22
Note. Hofstede uncertainty avoidance scores range from 8 (lowest) to 100 (highest). Age categories: 1 =18–24, 2 =25–34, 3 =35–44, 4 =45–54, 5 =55–64, 6 =65–74, 7 =75 or older. Vaccine hesitancy:
“If a vaccine to prevent COVID-19 were offered to you today, would you choose to get vaccinated?”(1 =yes, definitely, 2 =yes, probably, 3 =no, probably not, 4 =no, definitely not). % Hesitant =Percentage
of people who indicated “no, probably not”or “no, definitely not”. Side-effect concerns: “How concerned are you that you would experience a side effect from a COVID-19 vaccination?”(1 =not at all
concerned, 2 =slightly concerned, 3 =moderately concerned, 4 =very concerned). Understandably, some smaller countries had fewer participants, so it is important to interpret their descriptive statistics with
caution.
CULTURE, UNCERTAINTY AVOIDANCE, VACCINE HESITANCY 697
probably not, 4 =no, definitely not). Higher scores indicate higher
vaccine hesitancy.
Figure 3 maps the mean vaccine hesitancy across the world on
December 21, 2020 (the first day of the study). Notably, among the
countries for which Hofstede’s uncertainty avoidance scores are
available, the ten countries with the highest mean vaccine hesitancy
all had high uncertainty avoidance scores (e.g., Kazakhstan, Algeria,
Armenia), indicating the potential role of uncertainty avoidance in
COVID-19 vaccine hesitancy.
Vaccine Side-Effect Concerns (Mediator)
Starting from January 14, 2021, a question was added to the
global survey: “How concerned are you that you would experience a
side effect from a COVID-19 vaccination?”(1 =not at all con-
cerned, 2 =slightly concerned, 3 =moderately concerned, 4 =very
concerned). 6,498,946 participants who had not received a COVID-19
vaccine answered this question.
Control Variables
I used the same control variables as in Study 1: the other cultural
dimensions (individualism, indulgence, long-term orientation, mas-
culinity, power distance, tightness), demographics, COVID-19
severity, government response stringency, population density,
GDP per capita, common vaccine coverage, and religiosity. The
only exception is that Study 2’s survey measured education with an
open-ended question: “How many years of education have you
completed?”The responses were understandably noisy, so I decided
not to use this variable to be conservative.
Results
Descriptive statistics are displayed in Table 4 and bivariate
correlations in Supplemental Table S6. Because participants were
nested within countries, I conducted multilevel analyses to account
for (a) within-country statistical dependence and (b) the fact that
different countries had different sample sizes.
Cultural Differences in Vaccine Hesitancy
Decreased Over Time
Replicating Study 1’sfindings, there was a significant Uncertainty
Avoidance ×Date interaction effect on vaccine hesitancy in a
multilevel linear regression (Table 5 Model 1: β=−.071, SE =
.001, z=−64.83, p<.001). This effect remained robust after
accounting for individual-level controls (Model 2: β=−.068, SE =
.001, z=−55.90, p<.001), country-level controls (Model 3:
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Figure 3
Study 2: Mean Vaccine Hesitancy in 244 Countries/Territories on December 21, 2020 (the First Day When Vaccine Hesitancy Was Surveyed)
1 4
Note. Vaccine hesitancy: “If a vaccine to prevent COVID-19 were offered to you today, would you choose to get vaccinated?”(1 =yes, definitely, 2 =yes,
probably, 3 =no, probably not, 4 =no, definitely not). Darker colors (higher scores) indicate higher vaccine hesitancy. Blue areas indicate countries/territories
where data were not available on December 21, 2020. See the online article for the color version of this figure.
698 LU