Page 1 of 14

Advances in Social Sciences Research Journal – Vol. 9, No. 3

Publication Date: March 25, 2022

DOI:10.14738/assrj.93.11934. Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors.

Advances in Social Sciences Research Journal, 9(3). 70-83.

Services for Science and Education – United Kingdom

Analyzing the Effect of Exposure to COVID-19 Misinformation on

Health Behaviors

Meiling Laurence

Student, Science Research Department,

Paul D. Schreiber Senior High School, NY, USA

Olivia Platt

Student, Science Research Department,

Paul D. Schreiber Senior High School, NY, USA

Valerie A. Earnshaw

Associate Professor, Human Development & Family Sciences,

University of Delaware, DE, USA

ABSTRACT

American news media have become more diverse, siloed, and politically polarized

in recent decades, leading Americans with increasingly varied news consumption

patterns to receive disparate information. In addition, the accessibility of social

media platforms to unreliable users has resulted in misinformation-sharing on

many online platforms. This project aims to elucidate the origins of individuals’

COVID-19 pro- and anti-health behaviors — particularly in the context of their news

and information consumption — by investigating relationships between self- reported exposure to true and false statements about COVID-19, belief in such

statements, and COVID-19 pro- and anti-health behaviors. Furthermore, the effects

of primary news sources, time spent on social media, and demographics on these

variables were analyzed. COVID-19 statements were drawn from the C.D.C., W.H.O.,

and public polls by Pew Research Center. To collect data, a 27-question survey was

distributed to 250 panelists from the general U.S. population. Relationships

between pairs of variables were analyzed using chi-square tests, which showed that

exposure to COVID-19 statements had a significant (P < 0.05) effect on belief in

statements, exposure to true statements had a significant effect on COVID-19 health

behavior scores, and belief in statements had a significant effect on health behavior

scores. In addition, time spent on social media, age, and political affiliation, had

significant effects on these variables. This project adds to the growing body of

research on the effects of information sources on health behaviors and sheds light

on the genesis of COVID-19 pro- and anti-health behaviors.

Keywords: COVID-19; health misinformation; health behaviors; pro-health behaviors;

anti-health behaviors; news consumption; survey

INTRODUCTION

Americans’ sources of information were once relatively centralized and homogeneous, but in

recent decades, American news media have become more diverse, siloed, and politically

polarized, leading information accessed by the U.S. population to become increasingly varied.

Page 2 of 14

71

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

For much of the twentieth century, the Big Three broadcast news networks (NBC, ABC, and

CBS) predominated [1] and many Americans consequently consumed similar information from

this relatively small selection of sources. However, events in the second half of the century such

as the advent of narrowcasting in the 1960s and the 1987 repeal of the Fairness Doctrine

marked a trend toward increased variance in Americans’ information consumption, as news

coverage became less balanced and more niche [2]. Today, Americans’ sources for news often

vary depending on demographic factors, such as political party and age [3]. Moreover, social

media and the Internet have also led to increased variance in information consumed — the

accessibility of social media platforms to unreliable users [4] and its siloed nature due to

automated algorithms have resulted in widespread online misinformation-sharing [5].

When COVID-19 was declared a pandemic in 2020 by the World Health Organization (W.H.O.),

the abundance of online misinformation and other disparities in the quality and accuracy of

information consumed by the U.S. population elicited the question, what is the effect of exposure

to true and false COVID-19 information on individuals’ COVID-19 health behaviors? Pre-pandemic

research had indicated that exposure to and belief in health misinformation can affect health

behaviors; in 2014, Oliver et al. administered a survey whose results showed that 49% of

Americans agreed with at least one medical conspiracy theory and 18% agreed with three or

more, and suggested that conspiracism correlates with greater use of alternative medicine and

the avoidance of traditional medicine [6]. Likewise, in 2016, Earnshaw et al. (2016)

administered a survey showing that participants who more strongly endorsed conspiracy

beliefs reported that they would be less likely to seek care for Ebola and were less supportive

of quarantining people returning from West Africa [7].

Upon the start of the COVID-19 pandemic, novel studies were conducted focused on the COVID- 19 pandemic specifically, examining the relationships between COVID-19 information, beliefs,

and behaviors; in April of 2020, Earnshaw et al. (2020) demonstrated that COVID-19 conspiracy

beliefs may act as a barrier to some COVID-19 pro-health behaviors and policy support: survey

respondents who believed in COVID-19 conspiracies reported that their intentions to vaccinate

were 3.9 times lower and indicated less support for COVID-19 public health policies than

respondents who disbelieved conspiracies [8].

However, as COVID-19 is a novel disease, these studies were relatively few in number and some

results were inconsistent with those of pre-pandemic studies. The results of Earnshaw et al.

(2020) [8] diverged from those of Oliver et al. [6] in showing no differences between

cooperation with public health recommendations by conspiracy belief endorsement in a

multivariable regression analysis, for example. In addition, Chou et al. (2020) [5] recognized

that still little evidence is available regarding the extent to which misinformation exposure

online affects health-related behaviors, attitudes, knowledge, and outcomes, and that more

empirical research is necessary to elucidate the full extent of the real-world consequences of

misinformation exposure. Furthermore, while the effects of exposure to and endorsement of

misinformation and conspiracies have often been explored in past research, little research

exists on the effects of exposure to true health-related information on health behaviors. More

comprehensive research on the associations between exposure to true and false COVID-19

information, belief in such information, and COVID-19 health behaviors is thus necessary.

Page 3 of 14

72

Advances in Social Sciences Research Journal (ASSRJ) Vol. 9, Issue 3, March-2022

Services for Science and Education – United Kingdom

Consequently, this study investigates the association between exposure to and belief in false

and true COVID-19 statements, and how exposure, belief, and other variables — primary news

source, time spent on social media, gender, age, household income, U.S. census region,

education level, and political affiliation — affect COVID-19 health behaviors. It was

hypothesized that there would be an association between exposure to (true and false) COVID- 19 statements and belief in such statements, exposure to COVID-19 statements and COVID-19

health behavior scores, and belief in statements and COVID-19 health behavior scores. It was

further hypothesized that primary media source, time spent on social media, age, U.S. census

region, education level, and political affiliation would affect exposure to COVID-19 statements,

belief in statements, and COVID-19 health behavior scores.

METHODOLOGY

A 27-question survey was conducted and distributed to 250 panelists from the general U.S.

population via SurveyMonkey’s Audience feature. The survey comprises seven sections,

Informed Consent, News Consumption, Social Media, Exposure to Statements, Belief in

Statements, COVID-19 Health Behaviors, and Demographics. Survey questions were modeled

after public polls by the Pew Research Center and studies by social science researchers Mitchell

et al. [9] and Earnshaw et al. (2020) [8]. In order to build a substantial question set, additional

original questions were created by mimicking the syntax used by Mitchell et al. [9] and

Earnshaw et al. (2020) [8] and drawing text from statements by the C.D.C. and W.H.O as well as

devising original statements based on observation.

Materials

Over 2 million people take surveys on SurveyMonkey’s platform every day. Panelists comprised

a random selection of people from the general U.S. population who chose to participate in

SurveyMonkey’s Contribute or Rewards programs, through which they were given chances to

win sweepstakes prizes, gift cards, or money to donate to charities by participating in optional

surveys. After panelists completed the survey, SurveyMonkey automatically filtered out people

who did not complete it (non-responses). The total cost of survey distribution was $562.50.

Section I- Informed Consent

At the beginning of the survey, panelists were presented with a question that requested their

consent for participation in the study. The consent form asked them for permission to use the

data collected from the survey in this study. Panelists clicked "I agree" in order to indicate that

they had read the description of the survey and agreed to the terms as described.

Section II- News Consumption

To measure news consumption, a list was compiled of popular U.S. media sources [10, 11, 12,

13, 14] that fall under seven categories defined by Mitchell et al. [9] Panelists were prompted

to indicate their primary news source based on the list of different categories. Examples from

each media source category were included (compiled from lists of the most popular media

outlets and balanced by partisan bias) to clarify the purview of each, given that the panelists

may not have been familiar with the name of the category into which their primary news source

falls.

Furthermore, a follow-up free response question was included for panelists to specify their

particular primary news source in order to gain a clearer picture of their news consumption.

Page 4 of 14

73

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

When statistical analysis was conducted, however, this free response question was excluded

because it was determined to be too difficult to derive significance from the large pool of varied

responses.

Section III - Social Media

A question was asked to test the relationship between social media usage (i.e., hours spent on

social media each day and social media platforms used) and exposure to/belief in COVID-19

statements and COVID-19 health behaviors.

Section IV - Exposure to Statements

In this section, panelists indicated the extent to which they had been exposed to particular true

and false statements about COVID-19 (as defined by the C.D.C. , W.H.O., and social science

researchers). This was done in order to evaluate their exposure to misinformation and truths.

Panelists were asked, “How much, if anything, have you heard about the following?” after being

shown each statement. They responded with “a lot,” “a little,” or “nothing at all.”

COVID-19 Statements:

1) Taking vitamin C to prevent getting COVID-19 [9].

2) Becoming infected with COVID-19 from respiratory droplets when an infected person

coughs, sneezes, or talks [15].

3) Being able to hold your breath for 10 seconds or more without coughing or feeling

uncomfortable as an indicator that you don’t have COVID-19 [16].

4) Wearing masks for too long causing people to not get enough oxygen and experience

CO2 intoxication [16].

5) Becoming infected with COVID-19 by coming into close contact with a person who has

COVID-19 [15].

6) A connection between 5G mobile-phone technology and COVID-19 [9].

Of the statements that were created by drawing text from the C.D.C. and W.H.O., some contained

slightly simplified text that served to keep the questions at an eighth grade reading level.

Statements two and five are factual statements while the rest of the statements express

information deemed as false by the C.D.C and W.H.O. The factual statements were used as

positive controls when evaluating the panelists' exposure to COVID-19 misinformation. The

order of the true and false statements were randomized in order to prevent bias and promote

the authenticity of participant responses. Participants were compelled to rely on their own

knowledge regarding COVID-19 when factual statements were mixed with statements

containing misinformation.

Section V- Belief in Statements

In this section, panelists indicated the extent to which they believed in or agreed with each

statement. They were given the options “strongly agree,” “agree,” “disagree,” or “strongly

disagree.” An “I don’t know” option was included in the selection of responses because it was

considered to be plausible that some panelists would not have known or heard of the

information contained within the statements and would not have had an opinion on their

validity [17].

Page 5 of 14

74

Advances in Social Sciences Research Journal (ASSRJ) Vol. 9, Issue 3, March-2022

Services for Science and Education – United Kingdom

Section VI- COVID-19 Health Behaviors

In this section, panelists indicated the extent to which they engaged in pro-health COVID-19

behaviors — such as staying at home, avoiding public places, handwashing, mask-wearing, and

social-distancing — and anti-health COVID-19 behaviors — such as gathering maskless indoors

and dining indoors — which served as controls. They responded with “never,” “rarely,” “often,”

or “always.” Pro- and anti-health COVID-19 behaviors were assigned points in opposite ways

during the creation of health behavior scores for the statistical analysis, as is described in

greater detail in “Pre-Statistical Analysis: Creating Behavior Scores.” In this section, some

statements were drawn from Earnshaw et al. (2020) [8], while others mimicked the syntax of

Earnshaw et al. (2020) statements and included text drawn from Kramer et al. [18], or were

original statements devised based on observations.

COVID-19 Health Behavior Statements:

1) In the past month, how often, if ever, have you stayed at home as much as possible

because of the COVID-19 outbreak? [8]

2) In the past month, how often, if ever, have you avoided public places because of the

COVID-19 outbreak? [8]

3) In the past month, how often, if ever, have you gathered indoors with people who are

not a part of your household without wearing masks?

4) In the past month, how often, if ever, have you washed your hands with soap and water

for 20 seconds? [8]

5) In the past month, how often, if ever, have you worn a mask or face covering when in

stores or other businesses because of the COVID-19 outbreak? [18]

6) In the past month, how often, if ever, have you stayed 6 feet or more away from other

people when you go out in public because of the COVID-19 outbreak? [8]

7) In the past month, how often, if ever, have you dined indoors at a restaurant?

After responses were collected from all 250 panelists, question three was determined to have

demonstrated atypical results when compared to the rest of the behavioral statements; many

panelists whose answers to each of the other questions indicated extremely low engagement in

COVID-19 anti-health behaviors oddly answered that they “often” or “always” “gathered

indoors with people who are not outside of [their] household without wearing masks.” For this

reason, it was presumed that panelists possibly misinterpreted the question, and the question

was consequently excluded from calculation of panelists’ health behavior scores that were used

in the statistical analysis.

Section VII- Demographics

SurveyMonkey already collects demographic information; they always ask panelists for

information on their gender, age, household income, type of device used to take the survey, and

the U.S. census region of every survey panelist [19, 20]. As a result, only three additional

demographic questions were asked about state, education level, and political affiliation. There

was no question regarding race/ethnicity due to the limitations posed by the question quota;

this variable was deemed to be the least important to the context of this research. These

questions were asked at the end of the survey in order to reduce bias. Accordingly, panelists

were not asked about state, education level, or political affiliation until they already answered

the other survey questions.

Page 6 of 14

75

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

Institutional Review Board Approval

The plan for this research was approved by Paul D. Schreiber Senior High School’s Institutional

Review Board on February 2, 2021.

Data Coding: Creating COVID-19 Health Behavior Scores

Point values from 0-4 were assigned to the responses to Section VI questions, as shown in Table

I. The point value of 2 was skipped in order to account for what was deemed to be the

considerable difference between the responses “Rarely” and “Often.” “Never” for an anti-health

behavior awarded 0 points and “Always” for a pro-health behavior also awarded 0 points. The

point values that panelists received for each response were summed to determine their total

health behavior scores. The total point values were separated into four separate categories —

0-2, 3-4, 5-8, 9-20 — based on the first, second, and third quartiles of the data. Lower total point

values indicated pro-health behavioral tendencies, and thus panelists who scored in the 0-2

point category were considered to be the most COVID-19 risk-averse due to their high levels of

engagement in pro-health behaviors and limited engagement in anti-health behaviors.

Table I: COVID-19 health behavior score tables.

Pro-Health Behaviors Anti-Health Behaviors Point Value

Never Always 4

Rarely Often 3

Often Rarely 1

Always Never 0

Statistical Analysis

Chi-square tests for independence (performed via the “CHITEST” function on Google Sheets)

were used to compare pairs of variables in a contingency table. The chi-square test was chosen

due to the categorical nature of the data collected in this study. The chi-square equation is

shown in Figure I. In order to analyze a large number of individual variable relationships, 100

separate chi-square tests were conducted to evaluate the relationship between each individual

pair of variables. P < 0.05 was used to determine significance. Two tables were created to

display the overall p-values from each test (p-values were defined as the resultant value from

performing the “CHITEST” function on Google Sheets). The first table (Table II) depicts

relationships between behavior scores, media source category, and demographic questions

from the survey. In both tables, significant p-values are marked with single or double asterisks.

The second table (Table III) displays relationships involving the false and true COVID-19

statements described in the question methodology.

�! = � $�"_ − �")

!

�"

Figure I: Chi-square equation

�! = chi squared

�" = observed value

�" = expected value

Page 8 of 14

77

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

Age

Exposure

to

statement p < 0.01* p > 0.6

p <

0.0001** p < 0.001** p > 0.8 p < 0.01*

Age

Belief in

statement p < 0.05* p > 0.1 p < 0.001**

p <

0.0001** p > 0.4 p < 0.006*

US Census

Region

Exposure

to

statement p > 0.9 p > 0.2 p > 0.7 p > 0.1 p > 0.7 p > 0.7

US Census

Region

Belief in

statement p > 0.7 p > 0.8 p > 0.1 p > 0.3 p > 0.1 p > 0.2

Education

Exposure

to

statement p > 0.9

p <

0.00001** P > 0.4 P > 0.1 p < 0.05* p > 0.1

Education

Belief in

statement p > 0.2

p <

0.0001** p < 0.05* p < 0.003* p < 0.005* p > 0.1

Political

affiliation

Exposure

to

statement p > 0.1 p > 0.08 p > 0.3 p > 0.2 p > 0.5 p > 0.1

Political

affiliation

Belief in

statement p < 0.05* p > 0.2 p > 0.1

p <

0.000001** p > 0.07 p < 0.01*

*p<0.05, **p<0.001

Key: False Statement = FS; True Statement = TS

RESULTS

After chi-square tests were conducted, 45 of the p-values were found to be significant (P < 0.05).

Some of these p-values were highly significant, with values less than 0.0001. The relationship

between exposure to statements and belief in statements demonstrated the highest significance

out of all tested relationships. For the Vitamin C False Statement, Respiratory Droplets True

Statement, Holding Breath False Statement, Wearing Masks False Statement, and 5G False

Statement, p-values were less than 0.0001.

Table IV displays a sample chi-square test, specifically between exposure to the false Vitamin C

Statement (“Taking vitamin C to prevent getting COVID-19”) and time spent on social media.

The observed number of occurrences of the response “a lot” were much lower than the expected

number among those who used social media for a relatively few number of hours each day (as

indicated by the blue highlighting), suggesting that people who used social media for a fewer

amount of hours each day were less exposed to the Vitamin C False Statement. On the other

hand, the people who used social media for greater amounts each day of time were more

exposed to the Vitamin C False Statement, as suggested by observed values that were much

higher than expected values (indicated by the red highlighting). The large differences in

observed and expected values throughout this test resulted in a significant overall p-value of

0.014.

Page 10 of 14

79

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

In addition, time spent on social media had a significant effect on exposure to each of the

statements except for the Respiratory Droplets True Statement and the Wearing Mask False

Statement. In accordance with the aforementioned association between exposure and belief,

time spent on social media also had a significant effect on belief in each of the statements. The

ways in which time spent on social media affected belief and behavior are yet unclear.

Furthermore, age had a significant effect on time spent on social media: younger panelists were

more likely to spend greater amounts of time on social media each day and older panelists were

more likely to spend lesser amounts of time on social media each day. There was also a

significant relationship between age and exposure to false statements: younger panelists were

more likely to have been exposed to the Vitamin C, Holding Breath, Wearing Masks, and 5G

False Statements, and they were also more likely to report agreement with these statements

than older panelists.

Lastly, there were no significant relationships between political affiliation and exposure to any

of the statements. However, political affiliation did notably have a significant relationship with

belief in three of the false statements (Vitamin C, Wearing Masks, and 5G): Republicans were

more likely to report higher levels of agreement with the statements, while Democrats were

more likely to report lower levels of agreement. COVID-19 health behavior scores were also

affected by political affiliation: Republicans were more likely to exhibit higher health behavior

scores — indicating greater self-reported engagement in anti-health COVID-19 behaviors and

lesser engagement in pro-health behaviors — while Democrats were more likely to exhibit

lower health behavior scores, indicating greater self-reported engagement in pro-health

behaviors and lesser engagement in anti-health behaviors.

DISCUSSION

Before the start of this research, it was hypothesized that 1) there would be an association

between exposure to statements and belief in statements, 2) there would be an association

between exposure to statements and COVID-19 health behavior scores, 3) there would be an

association between belief in statements and COVID-19 health behavior scores, and 4) that

primary media source, time spent on social media, age, U.S. census region, education level, and

political affiliation would affect exposure to statements, belief in statements, and COVID-19

health behavior scores.

The results from the statistical analysis supported hypothesis one: there were significant

relationships between exposure to each of the six statements and belief in those statements,

with greater reported exposure being associated with greater agreement with statements.

Hypothesis two was partially supported: while self-reported extent of exposure to the false

COVID-19 statements did not have any significant effect on health behavior scores, higher levels

of exposure to the true COVID-19 statements were associated with lower health behavior

scores, which indicated greater engagement in pro-health COVID-19 behaviors and lesser

engagement in anti-health COVID-19 behaviors.

Hypothesis three was mostly supported: belief in five out of the six statements (all except the

Close Contact True Statement) had significant effects on health behavior scores; higher levels

of agreement with the false statements were associated with greater engagement in anti-health

Page 12 of 14

81

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

hindered the accuracy of their responses; multiple responses were completed in under 1.5

minutes, which is half of SurveyMonkey’s estimated completion time for the survey (3 minutes).

If statistical analysis were to be repeated, it would be prudent to exclude these responses from

the data. All health behaviors, as well as exposure to and belief in each of the statements, within

the survey were self-reported, which could have resulted in inaccuracies. There is also a

limitation in the definition of true and false information used in this study. Although the true

and false COVID-19 statements in the survey were defined as such by prominent health

organizations (i.e., the C.D.C. and W.H.O.), COVID-19 is still a novel disease and the scientific

consensus around it may continue to evolve and change as more information is unearthed.

Lastly, although this survey sampled the general U.S. population, it was not nationally

representative, as the proportion of respondents from each state did not correspond to the

proportion of the total U.S. population that each state represents. Future research should

conduct nationally representative surveys and/or weigh results to reflect the population of

each U.S. state.

Since pandemic norms are ever changing due to factors such as vaccination campaigns, the

advent and spread of COVID-19 variants, and discovery of new information, future studies

should account for potential changes in pandemic behavioral mores and expectations as the

pandemic progresses and scientific consensus evolves. The methodology used in this study may

also be applied to studies of the effects of misinformation on health behaviors related to other

diseases, including HIV/AIDS, morbillivirus (measles), and Ebola. While the data used in this

study is limited to self-reported information collected from a survey, future studies should

expand on this research by analyzing real-world data such as the number of COVID-19 cases,

hospitalizations, or deaths in specific communities [5]. Moreover, several of the findings from

this research merit future investigation: future research should investigate the relationships

between political affiliation and exposure to/belief in true/false information, and health

behaviors; the results of this study showed that political affiliation had effects on belief in

statements and COVID-19 health behaviors but not exposure to statements, and the reasons for

this are yet unclear. In addition, the effect of exposure to true information on health behaviors

should also be further explored, as this study showed significant associations between higher

levels of self-reported exposure to true COVID-19 statements and lower health behavior scores

(which indicate higher engagement in pro-health COVID-19 behaviors/lower engagement in

anti-health COVID-19 behaviors). This result is potentially consistent with past studies that

have shown that rapidly disseminating truths can have positive effects during a crisis [21], and

that using trusted authority figures to address conspiracy theories before people are exposed

to them may benefit COVID-19 prevention efforts [8]. Lastly, as research in this field remains

relatively novel, more repetitions are necessary to reduce error and solidify results.

CONCLUSIONS

This research adds to the growing body of literature on the effects of health misinformation on

health behaviors. To the knowledge of the contributors to this study, this research is novel in

examining the effects of both true and false information on health-related beliefs and behaviors,

rather than focusing on misinformation and conspiracies, as past studies have. In addition, this

research examines both panelists’ reported pro- and anti-health behaviors, whereas past

studies have generally focused on policy support and anti-health behaviors. Altogether, this

project sheds light on the effects of information sources on health-related beliefs and behaviors,

and the genesis of COVID-19 pro- and anti-health behaviors.

Page 14 of 14

83

Laurence, M., Platt, O., & Earnshaw, V. A. (2022). Analyzing the Effect of Exposure to COVID-19 Misinformation on Health Behaviors. Advances in

Social Sciences Research Journal, 9(3). 70-83.

URL: http://dx.doi.org/10.14738/assrj.93.11934

[18] Kramer, S. (2020, October 20). More Americans say they are regularly wearing masks in stores and other

businesses. Retrieved December 15, 2020, from https://www.pewresearch.org/fact-tank/2020/08/27/more- americans-say-they-are-regularly-wearing-masks-in-stores-and-other-businesses/

[19] Demographic segmentation made easy with Surveymonkey audience. (n.d.). Retrieved March 07, 2021, from

https://www.surveymonkey.com/curiosity/demographic-segmentation-made-easy-surveymonkey-audience/.

[20] Buying responses with surveymonkey audience. (n.d.). Retrieved March 07, 2021, from

https://help.surveymonkey.com/articles/en_US/kb/SurveyMonkey-Audience#Demographic

[21] Countering False Information on Social Media in Disasters and Emergencies, March 2018. (n.d.). Retrieved

December 11, 2020, from https://www.dhs.gov/sites/default/files/publications/SMWG_Countering-False-Info- Social-Media-Disasters-Emergencies_Mar2018-508.pdf