Psychological predictors of vaccination intentions among U.S. undergraduates and online panel workers during the 2020 COVID-19 pandemic.

OBJECTIVE: Availability of safe and effective vaccines against COVID-19 is critical for controlling the pandemic, but herd immunity can only be achieved with high vaccination coverage. The present research examined psychological factors associated with intentions to receive COVID-19 vaccination and whether reluctance towards novel pandemic vaccines are similar to vaccine hesitancy captured by a hypothetical measure used in previous research.
METHOD: Study 1 was administered to undergraduate students when COVID-19 was spreading exponentially (February-April 2020). Study 2 was conducted with online panel workers toward the end of the first U.S. wave (July 2020) as a pre-registered replication and extension of Study 1. In both studies, participants (total N = 1,022) rated their willingness to receive the COVID-19 vaccination and to vaccinate a hypothetical child for a fictitious disease, and then responded to various psychological measures.
RESULTS: In both studies, vaccination intentions were positively associated with past flu vaccine uptake, self-reported vaccine knowledge, vaccine confidence, and sense of collective responsibility. Complacency (not perceiving disease as high-risk), anti-vaccine conspiracy beliefs, perceived vaccine danger, and mistrust in science/scientists were negative correlates of vaccination intentions. Constraints (psychological barriers), calculation (extensive information-searching), analytical thinking, perceived disease vulnerability, self-other overlap, and conservatism were weakly associated with vaccination intentions but not consistently across both studies or vaccine types. Additionally, similar factors were associated with both real and hypothetical vaccination intentions, suggesting that conclusions from pre-COVID vaccine hesitancy research mostly generalize to the current pandemic situation.
CONCLUSION: Encouraging flu vaccine uptake, enhancing confidence in a novel vaccine, and fostering a sense of collective responsibility are particularly important as they uniquely predict COVID-19 vaccination intentions. By including both actual pandemic-related hesitancy measures and hypothetical hesitancy measures from past research in the same study, this work provides key context for the generalizability of earlier non-pandemic research.

Introduction

The rapid spread of COVID-19 has resulted in a global health crisis. As of November 2021, over 767,000 American lives have been lost [1]. For most of 2020, despite safety precautions (e.g., physical distancing, frequent hand-washing), the lack of effective antiviral treatment or vaccines resulted in a serious burden to the health care system. Additionally, resulting from local and state governments issuing lockdowns, the U.S. economy has been strongly negatively impacted [2]. Effective COVID-19 vaccines were hoped to both reduce detrimental health impacts and facilitate economic recovery by lifting pandemic restrictions. However, vaccines can only protect communities when enough people choose to vaccinate [3]. The early stages of the pandemic provided an unexpected opportunity to study vaccination intentions when effective treatments or vaccines against COVID-19 had not yet been developed or deployed, and social norms regarding the new vaccine had not yet been established. Whereas vaccine hesitancy research prior to the pandemic examined attitudes about vaccination for diseases that are hypothetical [4, 5], relatively low-threat (e.g., seasonal flu) [6], or mostly eradicated in the U.S. (e.g., polio) [7], the present research examined which psychological factors are associated with intention to receive a novel, pandemic vaccine in a context where no one was immune to a highly contagious and life-threatening disease.

Building on prior research that has identified individual, contextual, and vaccine-specific issues [7-10], the goal of our research was to examine whether these predictors generalize to a novel situation in which individuals decide whether to receive a newly vaccine designed to fight a previously unknown disease. Because the COVID-19 vaccine was developed and authorized in record time amidst unprecedented socioeconomic disruptions, there may be unique factors influencing vaccination intentions. Additionally, psychological research prior to the pandemic has documented anti-vaccine conspiracy beliefs and perceived dangers of vaccines as negative correlates of vaccination intentions, but findings were based on attitudes toward vaccinating a hypothetical child for a fictitious disease [4, 5, 11]. Using hypothetical scenarios is effective for conducting experiments because researchers can tightly control specific information about diseases or vaccines, but such methods suffer from ecological validity concerns. Thus, by measuring vaccination intentions for both COVID-19 (a real disease) and a fictitious child disease, we sought to examine whether previously identified factors associated with hypothetical immunization decisions are also predictive of real vaccination intentions.

While this article was under review, various studies on COVID-19 vaccination intention were published (see [12] for a review). Below, we briefly present findings from selected U.S. studies assessing similar measures as the current study. A cross-sectional study conducted in October 2020 using the 5C psychological antecedents of vaccination [13], a previously validated measure the current study also administered, found that perceived benefits (vaccine confidence and collective responsibility) as well as barriers to vaccination (complacency and constraints) were independently associated with COVID-19 vaccination intention [14]. Other cross-sectional surveys conducted in May-June 2020 showed that willingness to receive COVID-19 vaccine was associated with past flu vaccination behavior, perceived likelihood of infection, perceived severity of COVID-19, altruism, liberal political leaning, and having higher income [15-17]. Additionally, a longitudinal study conducted in March and July of 2020 showed increasing hesitancy over time and that conspiracy beliefs were inversely related to perceptions of vaccine safety and COVID-19 vaccination intention [18]. Another longitudinal survey conducted in March-August 2020 showed diverging ideological trajectories of COVID-19 vaccination intentions, where intentions declined among Republicans but not Democrats [19]. These studies highlight the importance of vaccination messages for different people and situations, and that decisions to accept a novel vaccine are complex and multi-faceted [20]. The current research adds to the collection of studies aimed to improve our understanding of attitudes about vaccinations.

The present research

The present research examined vaccine hesitancy when the pandemic was spreading exponentially in the U.S and vaccine development was initiating (Study 1; February-April 2020) and during Phase II clinical trials at the end of the first U.S. wave (Study 2; July 2020). Specifically, our research aimed to predict vaccination intentions for COVID-19 and for a hypothetical childhood disease (dysomeria) used in past research conducted before the pandemic (see Measures). We hypothesized that both COVID-19 and dysomeria vaccination intentions would be positively associated with confidence and collective responsibility aspects of the 5C psychological antecedents of vaccination [13], vaccine knowledge, analytical thinking, perceived disease vulnerability, self-other overlap, and prior flu vaccination. We also hypothesized that vaccination intentions would be negatively associated with complacency, constraints, and calculation aspects of the 5C, anti-vaccine conspiracy beliefs, perceived vaccine danger, mistrust in science/scientists, and political conservatism.

Verbatim materials and supplementary analyses are reported in the S1 File. Beyond what is reported here, additional measures were collected for both samples and that are reported only in the S1 File. For succinctness, analyses reported here focus on measures that were collected for both samples, but interested readers are encouraged to consult the S1 File for results of other highly relevant variables (e.g., media consumption). All data and analysis codes are available at https://osf.io/rd2s6/. Data were analyzed using R version 4.0.5 [21]. This research was approved by the Institutional Review Board at the University of Illinois, Urbana-Champaign. Waiver of documentation of informed consent was approved for this study.

Study 1

Participants and procedure

Study 1 participants included 346 students recruited from the undergraduate course-credit subject pool of the University of Illinois at Urbana-Champaign in Spring 2020. Undergraduate students in psychology courses who are 18 or older have the option to enroll in the subject pool to participate in research studies in exchange for course credit or to complete an alternative assignments with a similar duration. Participants in the subject pool voluntarily and individually sign up to complete studies and can withdraw from studies at any time. This data collection method is a standard practice in psychology [22, 23]. Sample size was determined by availability of credits. For Study 1, data collection began on February 24, 2020, and ended on April 21, 2020. After reading the informed consent and consenting to participate, participants responded to a number of measures before providing demographics and being debriefed. The university transitioned to fully remote learning on March 23, but the study was always administered online.

Participants’ data were excluded from analyses if they failed one or more attention-check questions, and/or spent M ≤ 10s per screen. After exclusions (see the S1 File for details), 308 participants remained (76.95% female, Mage = 19.65, SDage = 1.35). Sensitivity analyses showed that with α = .05 (two-tailed), we had 80% power to detect r = |0.16|. Reported racial/ethnic identities were Asian/Asian American (18.51%), Black/African American (5.52%), Hispanic/Latino (17.21%), White/European American (49.35%), more than one (7.14%), and other/prefer not to say (2.27%).

Measures

Unless noted, all items used 7-point scales ranging from 1 = Strongly disagree to 7 = Strongly agree.

Vaccination intention (COVID-19)

Participants read CDC-provided information about COVID-19, including total U.S. cases and deaths and that there were no antiviral treatments or vaccines yet available (see S1 File for details). Participants were asked whether they would receive the COVID-19 vaccine if it were available. Response options were: “I would…” 1 = not receive the vaccination even if it’s free ($0), 2 = receive the vaccination only if it’s free ($0), 3 = pay up to $10 to receive the vaccination, 4 = pay up to $25…, 5 = pay up to $50…, 6 = pay up to $100…, and 7 = pay more than $100 to receive the vaccination. We chose this measure because willingness to incur greater financial cost for a novel vaccine should reflect higher commitment to being vaccinated.

Vaccination intention (dysomeria)

To assess vaccination intention for a childhood disease, participants read about dysomeria—a (fictitious) disease spread by droplet infection, causing serious symptoms [11, 24]. Participants were asked to imagine that they had an 8-months-old infant. They were informed that vaccination against dysomeria was recommended by the CDC but that adverse events were reported 12% of the time. Participants then indicated their intention to vaccinate their hypothetical child (1 = Definitely not vaccinate, 7 = Definitely vaccinate).

Past flu vaccination uptake

Participants indicated their past flu vaccination uptake by responding to the question: “This past season, did you receive the flu shot?” with response options: “Yes,” “No,” and “prefer not to answer.” Responses were coded such that 0 = No/prefer not to answer and 1 = Yes.

Vaccine knowledge

Three questions assessed self-rated vaccine knowledge (e.g., “Compared to your peers, how much do you know about how vaccines work?”) using 7-point scales (1 = I am not at all knowledgeable, 7 = I am extremely knowledgeable).

5C psychological antecedents of vaccination

The 5C Psychological Antecedents of Vaccination measure [13] contains five subscales with three items each: confidence (e.g., “I am completely confident that vaccines are safe”), collective responsibility (e.g., “Vaccination is a collective action to prevent the spread of diseases”), complacency (e.g., “Vaccination is unnecessary because vaccine-preventable diseases are not common anymore”), constraint (e.g., “Everyday stress prevents me from getting vaccinated”), and calculation (e.g., “When I think about getting vaccinated, I weigh benefits and risks to make the best decision”). Items in each subscale were averaged to form composite scores. A programming error led to one item from the 5C’s preliminary scale being presented and was omitted from Study 1 analysis. Study 2 used the correct item from the final scale.

Anti-vaccine conspiracy beliefs

Anti-vaccine conspiracy beliefs [11] were assessed with seven items (e.g., “Immunizations allow governments to track and control people”). One item from the original measure regarding the flu vaccine was not included due to programming error.

Perceived vaccine danger

Perceived dangers of vaccines [11] were assessed with eight items (e.g., “I feel uncertain about the potential side-effects of immunizations”).

Mistrust in science/scientists

To assess mistrust in science/scientist, we drew five items with the highest inter-item correlations from the Trust in Science and Scientists Inventory [25]. These items (e.g., “Scientists ignore evidence that contradicts their work.”) were measured on 5-point scales (1 = Strongly disagree, 5 = Strongly agree).

Analytical thinking

The ten highest-loading items (e.g., “I enjoy intellectual challenges”) of the rational subscale of the Rational-Experimental Multimodal Inventory (REIm) [26] were used to assess analytic thinking (1 = Completely false, 5 = Completely true).

Perceived disease vulnerability

Perceived disease vulnerability was measured with the 7-item infectibility subscale (e.g., “I have a history of susceptibility to infectious disease”) of the Perceived Vulnerability to Disease Scale [27].

Self-other overlap

An adapted version of the Inclusion-of-Other-in-the-Self scale [28] asked participants to select one of seven Venn-diagrams that best represented their relationship with acquaintances. Higher numbers indicate greater self-other overlap.

Ideological conservatism

Participants indicated their ideological conservatism with one item: “Where would you place yourself on the following ideological spectrum?” (1 = Extremely liberal, 2 = Moderately liberal, 3 = Slightly liberal, 4 = Middle of the road, 5 = Slightly conservative, 6 = Moderately conservative, 7 = Extremely conservative).

Results

Descriptive statistics

A non-negligible proportion of participants from Study 1 indicated hesitancy regarding the COVID-19 vaccine. Specifically, 5.52% indicated they would not receive the vaccine even if it were free, and 6.49% indicated they would receive the vaccine only if it cost $0. The amount of money participants were willing to pay for a COVID-19 vaccine was: up to $10 (6.82%), up to $25 (14.94%), up to $50 (18.18%), up to $100 (15.58%), and more than $100 (32.47%). For the hypothetical dysomeria vaccine, hesitancy was less obvious with only 1.30% indicating that they would “definitely not vaccinate” the child and the majority (62.01%) endorsing the highest end of the scale indicating that they would “definitely vaccinate” the child. For prior vaccination behavior, 56.82% reported receiving the seasonal flu shot.

Bivariate correlations

Table 1 shows means, standard deviations, and correlations between the vaccination intention variables (COVID-19 and dysomeria) and predictor variables. The full correlation matrices of all measured variables are available in the S1 File. COVID-19 and dysomeria vaccination intentions were positively correlated (r = .31, p < .001). The correlational analyses revealed similar results for both types of vaccination (see Table 1) and support most of our hypotheses. Specifically, past flu vaccine uptake, vaccine knowledge, vaccine confidence, and collective responsibility consistently emerged as positive correlates of COVID-19 and dysomeria vaccination intentions. Complacency, constraint, calculation, anti-vaccine conspiracy beliefs, perceived vaccine danger, and mistrust in science/scientists were all negatively associated with.

Table 1

Reliability, means, and standard deviations of variables, and correlations between vaccination intention outcome variables (COVID-19 and dysomeria) and predictor variables for Study 1 and Study 2.

Variables	Study 1 (undergraduates)					Study 2 (online panel workers)
	α/r	M	SD	COVID	Dysomeria	α/r	M	SD	COVID	Dysomeria
COVID-19 vaccine intention	-	5.10	1.83	-	.31**	-	4.57	1.96	-	.45**
Dysomeria vaccine intention	-	6.36	1.09	.31**	-	-	5.65	1.82	.45**	-
Past flu vaccine uptake	-	0.57	0.50	.25**	.18*	-	0.56	0.50	.29**	.27**
Vaccine knowledge	.88	4.97	1.17	.22**	.19**	.92	4.93	1.37	.21**	.23**
5C-Confidence	.61	5.74	0.95	.41**	.42**	.73	5.05	1.45	.50**	.56**
5C-Collective responsibility	.44	6.42	0.93	.38**	.49**	.71	5.73	1.34	.36**	.51**
5C-Complacency	.52	2.04	1.05	-.23**	-.37**	.82	2.39	1.51	-.22**	-.42**
5C-Constraint	.69	1.94	1.08	-.19**	-.27**	.83	2.33	1.49	-.06	-.22**
5C-Calculation	.78	3.86	1.66	-.20**	-.29**	.77	5.30	1.38	-.07	-.03
Conspiracy beliefs	.83	1.95	0.90	-.32**	-.34**	.91	2.57	1.43	-.27**	-.45**
Perceived vaccine danger	.90	2.78	1.18	-.33**	-.46**	.92	3.39	1.44	-.35**	-.45**
Mistrust in science/scientists	.86	1.77	0.64	-.27**	-.24**	.94	2.22	1.05	-.33**	-.42**
Analytical thinking	.83	3.78	0.58	.08	.17*	.88	3.67	0.77	.11*	.10
Disease vulnerability	.88	3.73	1.17	.05	.06	.84	3.55	1.19	.16**	.14**
Self-other overlap (IOS)	-	4.28	1.56	.09	-.08	-	3.47	1.91	.19**	.09
Conservatism	-	3.18	1.49	-.05	-.21**	-	4.03	1.89	-.24**	-.24**

COVID = COVID-19 vaccination intention, Dysomeria = hypothetical child vaccination intention, 5C = 5C Antecedents of Vaccination, IOS = inclusion of other (acquaintances for Study 1, community members for Study 2) in the self. Past flu vaccine uptake is a dummy variable with receiving the seasonal flu shot coded as 1. COVID and Dysomeria columns report correlations with other variables. Bolded p < .05

* p < .01

** p < .001.

COVID-19 and dysomeria vaccination intentions. Analytical thinking was positively associated with dysomeria but not COVID-19 vaccination intentions, and political conservatism was negatively associated with dysomeria but not COVID-19 vaccination intentions. Bivariate associations were non-significant for perceived disease vulnerability and self-other overlap in both types of vaccination intentions.

Regression analyses

All measures were standardized prior to analyses. To examine which constructs uniquely predicted vaccination intentions, in two separate models, we regressed COVID-19 and dysomeria vaccination intentions on all 14 measures. We used semipartial rs (the unique contribution of each predictor; that is, the correlation between the outcome and the part of the focal predictor that is uncorrelated with the other predictors in the model) as estimates of predictor effect size. We then added gender, ethnicity, and log-transformed COVID U.S. deaths as control variables. Results are shown in Models 1a-1b (COVID-19) and Models 2a-2b (dysomeria) in Table 2.

Table 2

OLS regression models predicting vaccination intentions in Study 1 (undergraduates).

	COVID-19 Vaccine			Hypothetical Child Vaccine
	Model 1a		Model 1b	Model 2a		Model 2b
Predictor	β	sr	β	Β	sr	β
Past vaccination behavior	.25	.11	.23	.00	.00	-.01
(1 = flu vaccine received)	[.03, .47]		[.01, .46]	[-.20, .20]		[-.21, .20]
Vaccine knowledge	.05	.04	.06	-.01	-.01	-.01
	[-.07, .17]		[-.06, .18]	[-.12, .10]		[-.12, .10]
5C: Confidence	.22*	.16	.23*	.09	.07	.10
	[.09, .36]		[.09, .37]	[-.03, .22]		[-.03, .23]
5C: Collective	.18	.13	.18	.30**	.22	.29**
	[.04, .31]		[.04, .32]	[.17, .42]		[.17, .42]
5C: Complacency	.01	.01	.01	-.11†	-.08	-.10
	[-.12, .14]		[-.12, .15]	[-.23, .02]		[-.22, .02]
5C: Constraint	.02	.02	.02	.00	.00	.00
	[-.10, .14]		[-.09, .14]	[-.11, .10]		[-.11, .11]
5C: Calculation	-.10†	-.09	-.10†	-.11	-.10	-.11
	[-.21, .01]		[-.21, .02]	[-.21,.-01]		[-.22, -.01]
Conspiracy belief	-.07	-.05	-.08	.08	.05	.08
	[-.22, .08]		[-.23, .07]	[-.06, .22]		[-.06, .22]
Vaccine danger	.02	.01	.01	-.28**	-.17	-.28**
	[-.14, .19]		[-.15, .18]	[-.43, -.13]		[-.43, -.13]
Science mistrust	-.05	-.04	-.05	.10	.07	.10
	[-.18, .08]		[-.18, .08]	[-.02, .22]		[-.02, .22]
Analytic thinking	-.03	-.03	-.02	.11	.10	.11
	[-.14, .08]		[-.14, .09]	[.01, .21]		[.01, .21]
Disease vulnerability	.03	.03	.03	.01	.01	.01
	[-.07, .14]		[-.08, .14]	[-.09, .11]		[-.09, .11]
Self-other overlap	.07	.07	.08	-.10	-.10	-.10
	[-.03, .17]		[-.02, .19]	[-.19, -.01]		[-.20, -.01]
Conservatism	.05	.05	.06	-.04	-.04	-.04
	[-.06, .16]		[-.05, .18]	[-.14, .06]		[-.14, .07]
Gender (1 = female)			.07			.06
			[-.19, .33]			[-.18, .30]
Ethnicity (1 = non-White)			.14			-.02
			[-.08, .37]			[-.23, .18]
COVID US deaths (log)			-.06
			[-.16, .04]
N	308		307	308		307
R2 /Adjusted R2	0.25/0.21		0.26/0.21	0.37/0.34		0.37/0.33

Coefficients [95% CI] are standardized. sr = semipartial r.

† p < .10, bolded p < .05

* p < .01

** p < .001.

The adjusted R2 of these models were 21% (COVID-19) and 33%-34% (dysomeria), suggesting that the set of predictors, together, explained substantial variance in vaccination intentions. Past flu vaccination behavior (semipartial r = .11), vaccine confidence (semipartial r = .16), and collective responsibility (semipartial r = .13) emerged as unique predictors of COVID-19 vaccination intentions with and without control variables. Collective responsibility (semipartial r = .22), calculation (semipartial r = -.10), perceived vaccine danger (semipartial r = -.17), and analytic thinking (semipartial r = .10) uniquely predicted decisions regarding hypothetical child vaccine. Self-other overlap (semipartial r = -.10) was another unique predictor for dysomeria, but the association was in the opposite direction than our prediction. Thus, although similar associations were observed for both types of vaccines in bivariate correlations, only collective responsibility emerged as a unique predictor for both vaccines in Study 1.

Study 2

Study 2 was a replication and extension of Study 1, remedying some technical issues from the first study and collecting data from a more age-diverse sample of online panel workers. Study 2 was preregistered at https://osf.io/rycjd. We conducted an a priori power analysis (see preregistration) and aimed to collect complete data from a sample of 652 CloudResearch Prime Panel participants. Study 2 data were collected on July 20, 2020. We received 848 complete responses and after applying preregistered exclusion criteria (e.g., incomplete responses, attention check failures; see the S1 File for details), the final sample size was 676. Sensitivity analyses showed that with α = .05 (two-tailed), we had 80% power to detect r = |0.11|. Procedures were identical to Study 1, except the 5C measures were assessed after the vaccination intention items in Study 2, but the order of remaining questionnaire was the same.

Participants were 62.87% female with mean age of 51.95 (SD = 18.05). Reported racial/ethnic identities were Asian/Asian American (6.07%), Black/African American (6.07%), Hispanic/Latino (4.14%), Native American/Pacific Islander (1.04%), White/European American (79.29%), more than one (1.78%), and other/prefer not to say (1.63%). Participants’ annual household income was as follows: $0-$25,000 (21.30%), $25,001-$50,000 (26.48%), $50,001-$75,000 (19.97%), $75,001-$100,000 (14.35%), $100,001-$125,000 (6.21%), $125,001-$150,000 (4.88%), $150,001-$175,000 (1.92%), $175,001-$200,000 (1.78%), and more than $200,000 (3.11%). Additionally, 60.21% reported being a parent. When asked if they had been tested for COVID-19, the majority (88.61%) responded “No,” 0.44% preferred not to respond, and 10.95% responded “Yes.” Current residence of this sample included all 50 states (see S5 Table in S1 File). Demographic characteristics of Prime Panel samples tend to be more comparable to a nationally-representative sample than Amazon Mechanical Turk [29].

The same measures from Study 1 were used in Study 2 with a few exceptions we note below.

Participants read CDC-provided information about COVID-19, including total U.S. cases and deaths and that there were no antiviral treatments or vaccines yet available (see S1 File for details). COVID-19 vaccination intentions were assessed with one discrete measure, followed by a continuous measure. Participants were first asked whether they would receive the COVID-19 vaccine if it were available, with response options: A = I would not receive the vaccination even if it’s free ($0), B = I would receive the vaccination only if it’s free ($0); if I need to pay money, I would not receive the vaccination, and C = I would pay money to receive the vaccination. On the next page, participants moved a slider scale ($0-$500) to indicate the maximum amount of money they would personally pay for the vaccine.

As an exploratory measure, participants were also shown a list of eight common concerns (e.g., “The vaccine is too new”) about the vaccine and were asked to indicate their agreement with each statement (1 = Strongly disagree to 7 = Strongly agree) and whether or not each concern mattered for their decision regarding COVID-19 vaccination [30].

Participants indicated their past flu vaccination uptake by responding to the question: “This past flu season (October 2019-April 2020), did you receive the flu shot?” Responses were coded such that 0 = No and 1 = Yes.

The same self-other overlap scale as Study 1 was used except the “other” circle represented other community members instead of acquaintances.

Hesitancy toward COVID-19 vaccine was relatively high in Study 2 with 18.05% selecting A (I would not receive the vaccination even if it’s free), 27.51% selecting B (if I need to pay money, I would not receive the vaccination), and only 54.44% selecting C (I would pay money to receive the vaccination). The maximum amount of money participants indicated they would personally pay for a COVID-19 vaccine varied substantially, ranging from $0 to $500 (M = 102.81, SD = 142.50, Median = 40.00).

We used the discrete measure and continuous measure to create a new variable similar to Study 1’s measure, where 1 = selected option A and $0 on the slider, 2 = selected option B and $0 on the slider, 3 = $1-$10 on the slider, 4 = $11-$25, 5 = $26-$50, 6 = $51-$100, and 7 = indicated more than $100 on the slider. With this variable, 10.95% were willing to pay $0 and refused (i.e., they would not receive the vaccination even if it were free), 6.80% were willing to pay $0 and hesitant (i.e., they would not receive the vaccination if not free), and the amount of money other participants were willing to pay was: $1-$10 (11.09%), $11-$25 (17.01%), $26-$50 (16.57%), $51-$100 (15.09%), and more than $100 (22.49%). This combined measure of COVID-19 vaccination intention was used in correlational and regression analyses. Supplementary analyses treating the two items as separate measures are reported in the S1 File.

In Study 2, participants were asked about various concerns related to the COVID-19 vaccine. Table 3 shows the means and standard deviations of ratings for these concerns as a function of the discrete COVID-19 vaccination intention measure. When asked whether or not each concern mattered for participants’ decision to vaccinate, “I worry about the side effects” (45%), “the vaccine is too new” (44%), and “the vaccine will not protect me” (35%) were the top three concerns (see S3 Fig in S1 File). These results corroborate other recent work on this topic [30].

Table 3

Means and standard deviations of ratings for concerns influencing vaccination as a function of COVID-19 vaccination intention in Study 2.

COVID-19 Vaccination Intention	Refusal		Hesitant		Willing to pay
Concern	M	SD	M	SD	M	SD
The vaccine is too new	6.01	1.41	4.91	1.70	3.92	1.90
I worry about the side effects	5.89	1.68	4.95	1.80	3.90	1.96
The vaccine will not protect me	4.70	1.73	3.29	1.60	2.39	1.54
I avoid most vaccines	4.37	2.22	2.65	1.80	1.80	1.48
COVID-19 is not severe enough to concern me	3.93	2.15	2.46	1.80	1.68	1.39
A doctor has recommended no vaccines	3.58	2.12	3.45	1.97	2.86	2.11
I will not have access to the vaccine	3.02	1.78	3.41	1.65	2.66	1.58
My religion prevents vaccination	2.00	1.63	1.54	1.26	1.54	1.32

Refusal = “I would not receive the vaccination even if it’s free ($0).” Hesitant = “I would receive the vaccination only if it’s free ($0); If I need to pay money, I would not receive the vaccination.” Willing to pay = “I would pay money to receive the vaccination.”

For the hypothetical dysomeria vaccine, 6.66% indicated that they would “definitely not vaccinate” the child, and half of the sample (49.26%) endorsed the highest end of the scale that they would “definitely vaccinate” the child. For prior vaccination behavior, 55.77% reported receiving the seasonal flu shot.

Table 1, presented earlier, shows Study 2’s means, standard deviations, and correlations between the vaccination intention variables (COVID and dysomeria) and predictor variables. The full correlation matrices of all measured variables are available in the S1 File. The correlation between COVID-19 and dysomeria vaccination intentions was moderately high (r = .45, p < .001). Consequently, the correlational analyses revealed similar results for both types of vaccination in Study 2. Notably, results from Study 2 mostly replicated those from Study 1 with a few exceptions (see Table 1).

As in Study 1, correlational analyses support most of our hypotheses. Specifically, past flu vaccine uptake, vaccine knowledge, vaccine confidence, collective responsibility, analytical thinking, disease vulnerability, and self-other overlap consistently emerged as positive correlates of COVID-19 and dysomeria vaccination intentions, although the effect sizes for the last three variables were somewhat modest (see Table 1). Complacency, anti-vaccine conspiracy beliefs, perceived vaccine danger, and mistrust in science/scientists were all negatively associated with COVID-19 and dysomeria vaccination intentions. Notably, political conservatism was negatively associated with COVID-19 vaccination in Study 2 but not in Study 1, which may be indicative of increased polarization of opinions regarding COVID-19 vaccination over time [19] or differences between student and online sampling frames. Constraint was negatively associated with dysomeria but not COVID-19 vaccination intentions. Calculation had non-significant associations with both types of vaccines in Study 2.

All measures were standardized prior to analyses. Analytical methods were the same as Study 1, except we included gender, ethnicity, age, income, log-transformed COVID state deaths, and parenthood as additional demographic controls. Results are shown in Models 1a-1b (COVID-19) and Models 2a-2b (Dysomeria) in Table 4. The adjusted R2 of these models were 30%-38% (COVID-19) and 39%-41% (dysomeria), suggesting that the set of predictors explained substantial variance in vaccination intentions.

Table 4

OLS regression models predicting vaccination intentions in Study 2 (online panel workers).

	COVID-19 Vaccine			Hypothetical Child Vaccine
	Model 1a		Model 1b	Model 2a		Model 2b
Predictor	β	sr	β	β	sr	β
Past vaccination behavior	.25*	.11	.24**	.00	.00	.03
(1 = flu vaccine received)	[.10, .40]		[.10, .38]	[-.14, .14]		[-.11, .17]
Vaccine knowledge	.04	.03	.00	.11*	.09	.09
	[-.04, .11]		[-.07, .07]	[.04, .19]		[.02, .16]
5C: Confidence	.32**	.23	.27**	.35**	.24	.32**
	[.23, .42]		[.18, .35]	[.26, .43]		[.23, .41]
5C: Collective	.06	.04	.09†	.13*	.09	.13*
	[-.03, .16]		[.00, .19]	[.04, .23]		[.04, .23]
5C: Complacency	.05	.03	.03	-.12	-.07	-.14*
	[-.06, .15]		[-.07, .13]	[-.22, -.03]		[-.23, -.04]
5C: Constraint	.10	.08	.06	.03	.02	-.02
	[.02, .18]		[-.02, .14]	[-.05, .11]		[-.10, .06]
5C: Calculation	-.04	-.04	-.03	.01	.01	.03
	[-.11, .03]		[-.10, .03]	[-.05, .07]		[-.03, .10]
Conspiracy belief	.04	.02	.00	-.09	-.05	-.12
	[-.08, .16]		[-.11, .12]	[-.21, .02]		[-.23, -.01]
Vaccine danger	-.10	-.05	-.04	.00	.00	.02
	[-.22, .02]		[-.15, .07]	[-.11, .11]		[-.09, .13]
Science mistrust	-.06	-.04	-.08	-.03	-.02	-.02
	[-.16, .04]		[-.17, .02]	[-.13, .06]		[-.12, .07]
Analytic thinking	.03	.03	.00	-.04	-.03	-.06†
	[-.04, .11]		[-.07, .07]	[-.11, .03]		[-.13, .01]
Disease vulnerability	.10*	.09	.12**	.03	.03	.03
	[.03, .17]		[.06, .19]	[-.03, .10]		[-.03, .10]
Self-other overlap	.09	.08	.07	-.03	-.03	-.04
	[.02, .15]		[.00, .13]	[-.09, .03]		[-.10, .03]
Conservatism	-.08	-.07	-.10*	-.04	-.03	-.05
	[-.15, -.01]		[-.17, -.03]	[-.11, .03]		[-.12, .02]
Gender (1 = female)			-.21*			-.19*
			[-.35, -.08]			[-.32, -.05]
Ethnicity (1 = non-White)			.00			-.16
			[-.16, .16]			[-.32, .00]
Age			-.02			-.11*
			[-.10, .05]			[-.18, -.03]
Income			.26**			.08
			[.19, .32]			[.02, .14]
COVID state deaths (log)			-.01
			[-.07, .06]
Parenthood (1 = parent)						.04
						[-.09, .17]
N	674		670	674		670
R2 /Adjusted R2	0.32/0.30		0.40/0.38	0.40/0.39		0.42/0.41

Coefficients [95% CI] are standardized. sr = semipartial r.

† p < .10, bolded p < .05

* p < .01

** p < .001.

Replicating Study 1, past flu vaccination behavior (semi-partial r = .11) and vaccine confidence (semipartial r = .23) emerged as unique predictors of COVID-19 vaccination intentions with and without control variables. Although collective responsibility was another unique predictor in Study 1, this trend was weaker in Study 2 with semi-partial r = .04. However, disease vulnerability (semipartial r = .09), self-other overlap (semipartial r = .08), and conservatism (semipartial r = -.07) emerged as unique predictors of COVID-19 vaccination intention in Study 2. Replicating Study 1, collective responsibility (semipartial r = .09) again emerged as a significant unique predictor of dysomeria vaccination intention, with and without controls. Additionally, vaccine knowledge (semipartial r = .09), confidence (semipartial r = .24), and complacency (semipartial r = -.07) each uniquely predicted dysomeria vaccination intention and were robust to inclusion of control variables. Vaccine confidence and collective responsibility—though latter to a lesser extent—were unique predictors of intentions for both vaccines in Study 2. Moreover, gender and income uniquely predicted intentions for both types of vaccines in Study 2 (see Table 4)—a finding consistent with other recent COVID-19 research [14, 17].

Discussion

Although COVID-19 vaccines have become widely available in the U.S., vaccine hesitancy and refusal persist [31]. Hence, combatting vaccine hesitancy remains an extraordinary challenge for public health experts, and understanding the psychological roots of vaccine hesitancy during the ongoing pandemic is important to prepare for similar health crises in the future [32]. This work therefore remains timely in its identification of correlates of vaccine hesitancy, particularly since the rate of COVID-19 vaccination has begun to slow [33]. In the current research, a number of variables were associated with both actual (COVID-19) and hypothetical child (dysomeria) vaccine hesitancy.

Key findings

Among both U.S. undergraduates (Study 1) and online panel workers (Study 2), we found that vaccination intentions were positively associated with past flu vaccination behavior, self-reported vaccine knowledge, confidence, and sense of collective responsibility. Complacency, anti-vaccine conspiracy beliefs, perceived vaccine danger, and mistrust in science/scientists were negative correlates. For constraint, calculation, analytic thinking, perceived disease vulnerability, self-other overlap, and political conservatism, strong associations were not consistently observed across vaccine types and samples. Consistent with prior research [30, 34, 35], previous vaccination against seasonal influenza was uniquely and strongly associated with COVID-19 vaccination intention. Confidence in the safety/effectiveness of vaccines also emerged as a unique predictor. Study 2 further revealed that worries about side effects, the vaccine being too new, and the vaccine not being protective were most prevalent concerns regarding the COVID-19 vaccine.

Implications and contributions

To the extent that people who receive flu shots are more likely to also accept the COVID-19 vaccine, interventions promoting seasonal influenza vaccination in advance of a future health crisis may be one way to increase pandemic vaccine uptake. Given the strong association of vaccine confidence to intention, another useful avenue for intervention, particularly for a novel vaccine, may be to target confidence in the product by emphasizing its safety and highlighting a positive ratio of benefits to potential side effects, and communicating these points effectively using trusted parties. Although not all variables uniquely predicted vaccine hesitancy when accounting for shared variance among the larger set, individual predictors identified in zero-order correlations should not be discounted. Individually, collective responsibility, complacency, vaccine knowledge, conspiracy beliefs, perceived vaccine danger, and science mistrust were consistently associated with COVID-19 vaccination intentions in both studies. Given the strong overall predictive power of the variable set as a whole, it appears that there are multiple routes to tackling hesitancy.

To our knowledge, no prior study has simultaneously measured vaccine attitudes for both a hypothetical disease and a real disease during a pandemic. Measuring both together is important for providing context for past work regarding how well they generalize to actual intentions during a crisis period. That is, studying vaccination attitudes about unknown, hypothetical diseases [4, 5, 11] lacks ecological validity, and studying vaccination intentions for real, more “established” diseases for which vaccines have existed for years (and most people are inoculated) may not provide the same insights as attitudes towards a novel pandemic vaccine. The current work overcame these limitations and shows that factors associated with immunization decisions for a hypothetical disease are also mostly predictive of real vaccination intentions for a novel disease. By asking about both hypothetical and real diseases, this research has improved our understanding of the utility of previous work in its application to the current pandemic.

Limitations

We acknowledge a few limitations of the current work. First, we cannot infer causality from our cross-sectional studies. Although both studies were conducted during critical times in the pandemic when effective treatments or vaccines were not yet available, longitudinal studies examining changes in vaccination attitudes across the pandemic timeline would be informative. Second, our samples of college students and online panel workers are not representative of all U.S. residents. Although a larger and nationally-representative sample is needed to capture COVID-19 vaccination intentions of the general U.S. population, we expect our results to be reproducible with different samples of students from similar undergraduate subject pools and Prime Panel workers. Moreover, understanding attitudes about pandemic vaccines and identifying potential reasons for hesitancy, even among a subset of the populations, is critical to address the broader goal of achieving sufficient vaccine coverage. Third, COVID-19 vaccine intention was operationalized as self-reported willingness to pay money for the vaccine. Although this item should reflect commitment to vaccinate, future studies could benefit from measuring actual delay in vaccinating or vaccination behavior as indicators of hesitancy. Despite these limitations, the present work documents psychological factors associated with vaccine hesitancy in the context of a life-changing pandemic.

Online supplementary materials.

Study materials, supplementary measures and analyses.

(DOCX)

Click here for additional data file.

1 Oct 2021

PONE-D-21-22539Psychological predictors of vaccination intentions among U.S. undergraduates and online panel workers during the 2020 COVID-19 PandemicPLOS ONE

Dear Dr. Watanabe,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Nov 15 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Ismaeel Yunusa, PharmD, PhD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

2. Please include your full ethics statement in the ‘Methods’ section of your manuscript file. In your statement, please include the full name of the IRB or ethics committee who approved or waived your study, as well as whether or not you obtained informed written or verbal consent. If consent was waived for your study, please include this information in your statement as well.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This manuscript adds an increasing list of similar researches on psychological predictors of covid-19 vaccination intentions, and the results are not surprisingly consistent with what others have been found, but still it's good to see the authors collected two new samples on vaccine intentions (the college undergrads and the online workers).

My main questions are:

1) There is no justification of the representativeness of the sample for their surveys relative to the general US populations, which makes it hard to assess whether the study is reliably capturing the attitudes of the general population toward covid-19 vaccination. The sample size for each of the two samples seemed too small to be much representative.

2) There is a concern that some of the reported statistical analyses may not be valid, due to likely violations of the assumptions of a linear regression model. This is especially for the college student sample. If the students surveyed have certain clustering effect, e.g., some are in the class, in the same department or program, then the data is not independent. If this is the case, some kind of linear mixed model with random effects is more appropriate than the multiple regression model that the authors employed.

Reviewer #2: Thank you for the opportunity to review this manuscript. I appreciate the amount of work that has been put into this study; the authors should be commended for their efforts to study the psychological roots of vaccine hesitancy during the current pandemic. Below are my concerns

•I have major problems with the way the data are presented, and the manuscript is written and I recommend major revision to have a more concise manuscript with a clear flow and more clear conclusion. I include a couple of examples on issues that need to be further clarified below.

•I do not see any benefit of splitting the study samples into 1& 2, and if there was, it was not made clear in the presentation of the study. I understand that sample 2 was needed to have a more age-diverse sample, but why not combine them? If it was because you expected differences between the two samples, then explain why you would put this assumption? On what basis? This needs to be explained or changed. There are also differences in the questions asked to each sample which were not explained; on what basis did the author choose to do so? Any benefits? Wouldn't that increase variability of the results among the samples?

•The manuscript reads more like a thesis than a concise manuscript it lacks “straight to the point” presentation of data and clear conclusions. I recommend rewriting the manuscript, especially the introduction, which should be one coherent piece of work, also the methods and the results. The conclusion in the abstract will need reconsideration as it starts with a vague sentence and does not give any concise conclusion

•The questionnaires used are very lengthy, and despite the attention checks the authors have described, no information has been provided regarding how many participants were excluded due to failing the attention check; they only report a general total number of 38 excluded for different reasons for sample one and same regarding sample 2 were 172 were excluded for several reasons.

•Line 85: "Whereas previous research has examined attitudes about vaccination 86 for diseases that are hypothetical (e.g., Haase et al., 2015; Jolley & Douglas, 2017), relatively 87 low-threat (e.g., seasonal flu; Zhang et al., 2010) or mostly eradicated in the U.S. (e.g., Dubé et 88 al., 2015), the present research investigates vaccination intentions during a unique period of uncertainty involving a highly contagious disease when effective treatments or vaccines had not yet been developed." The authors should report the published evidence on vaccine intentions in the COVID-19 era

•The authors mention that” Sample 1 included 346 undergraduate students from the University of Illinois at Urbana183 Champaign recruited in exchange for course credit” Can you elaborate more on what do you mean here? Were the participants given a choice to join? How did you guarantee that participants were aware that their participation is optional?

•The discussion section can be improved by highlighting the take-home message from the study and take the reader through what has been done and what does that mean rather than displaying the results. As an example in line 418: "knowledge, no prior study has simultaneously measured vaccine attitudes for both a hypothetical disease and a real disease during a pandemic. Measuring both together and determining which variables predict well across measures is important, as it provides context for past work and can bolster confidence in how well past research applies to actual intentions during a crisis period…… The current work shows that factors associated with immunization decisions for a hypothetical disease are also mostly predictive of real vaccination intentions." I do not see the point here? Can you elaborate on the reasons behind studying vaccine attitudes for both hypothetical disease and a real disease during a pandemic? Not at any point of the manuscript, this was discussed clearly

•The tables are confusing with too many details; I suggest editing them to highlight the most important message from each table and move any unnecessary information to the supplementary file.

•In the supplementary file, Why was the attention check different between samples 1 &2? What did you base this on?

For Sample 1, the following item was asked prior to the demographics questionnaire:

•Thank you for your responses. You are almost finished with this study. Before answering a few questions about yourself, we have one last question. Do you think we should include your responses in our study? That is, did you take the study seriously and respond thoughtfully? Your credit assignment for this study does not depend on your response to this question. [Yes, I responded thoughtfully / No, I did not respond thoughtfully.]

Sample 2 Only

For Sample 2, the following item was embedded in the optimism bias measure:

If you are reading this, please select 67 on

this sliding scale.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: PONE review.docx

Click here for additional data file.

7 Nov 2021

Author Response to Reviewer Comments

Thank you for offering us the opportunity to revise and resubmit our manuscript “Psychological predictors of vaccination intentions among U.S. undergraduates and online panel workers during the 2020 COVID-19 Pandemic” to PLOS ONE. We sincerely thank you for the thoughtfulness and care with which you read and evaluated our work. Our detailed responses to your comments can be found below, in bold blue print. Thank you again for your time.

Comments from Reviewer 1

1. This manuscript adds an increasing list of similar researches on psychological predictors of covid-19 vaccination intentions, and the results are not surprisingly consistent with what others have been found, but still it's good to see the authors collected two new samples on vaccine intentions (the college undergrads and the online workers).

RESPONSE: Thank you for your positive evaluation of our manuscript. We agree that COVID-19 vaccine intention is an interesting, important, and growing area of research.

2. My main questions are:

1) There is no justification of the representativeness of the sample for their surveys relative to the general US populations, which makes it hard to assess whether the study is reliably capturing the attitudes of the general population toward covid-19 vaccination. The sample size for each of the two samples seemed too small to be much representative.

Response: Thank you for bringing up this important point regarding the representativeness of our samples. We agree that our samples of undergraduate students and online panel workers are not necessarily representative of the U.S. population. In the original submission, we had acknowledged this limitation in the General Discussion on p. 23 (line 434-435): “…the present study was conducted online in the U.S. using convenience samples, which were not representative of all U.S. residents.” Based on your feedback, we have discussed this limitation further on p. 25 (lines 495-501) of the revised manuscript and made this more salient by creating a “limitations” section. While we continue to acknowledge that our samples are not necessarily representative of the general U.S. population, we have added that we expect the results to be reproducible with students from similar undergraduate subject pools and with Prime Panel workers because our samples are representative of target populations from which we sampled. Moreover, we believe that identifying potential reasons for hesitancy, even among a subset of the population, is critical to address the broader goal of achieving sufficient vaccine coverage. Additionally, we noted on pp. 17 (lines 335-337) that our online sample was drawn from all 50 states and that Prime Panels are more representative of the country than other online platforms (e.g., Amazon Mechanical Turk) across demographic variables including age, marital status, number of children, political affiliation, and religious devotion (Chandler et al., 2019).

3. 2) There is a concern that some of the reported statistical analyses may not be valid, due to likely violations of the assumptions of a linear regression model. This is especially for the college student sample. If the students surveyed have certain clustering effect, e.g., some are in the class, in the same department or program, then the data is not independent. If this is the case, some kind of linear mixed model with random effects is more appropriate than the multiple regression model that the authors employed.

Response: In Study 1, all participants were recruited from an undergraduate psychology course-credit subject pool at the University of Illinois Urbana-Champaign in Spring 2020. Undergraduate students who are 18 or older taking one or more psychology courses have the option to enroll in the subject pool to participate in research studies for course credit or to complete an alternative assignment. Participants in the subject pool individually sign up to complete studies on their own time (i.e., outside of class). Although participants in the subject pool consist of students with different majors, year in college, classes, etc., a) our data represent individual responses of these students who are all drawn from same participant pool, and b) we did not collect data on students’ home department, year in college, classes taken, etc. Hence, no clustering effect can be modeled. As we discuss in our response to Reviewer 2’s comment regarding the subject pool (#10 below), we additionally clarify in the revised manuscript on p. 8 (lines 175-181) that data collection using undergraduate subject pools is a standard practice in psychology and social sciences more broadly (Rossell et al., 2005; Sadeghiyeh et al., 2021).

Comments from Reviewer 2

4. Thank you for the opportunity to review this manuscript. I appreciate the amount of work that has been put into this study; the authors should be commended for their efforts to study the psychological roots of vaccine hesitancy during the current pandemic.

RESPONSE: Thank you. We appreciate your positive feedback as well as your critical insights that have allowed us to improve the work.

5. Below are my concerns:

I have major problems with the way the data are presented, and the manuscript is written and I recommend major revision to have a more concise manuscript with a clear flow and more clear conclusion. I include a couple of examples on issues that need to be further clarified below.

RESPONSE: Thank you for this feedback and your detailed comments. In our revision, we believe we have addressed these concerns (see below).

6. I do not see any benefit of splitting the study samples into 1& 2, and if there was, it was not made clear in the presentation of the study. I understand that sample 2 was needed to have a more age-diverse sample, but why not combine them? If it was because you expected differences between the two samples, then explain why you would put this assumption? On what basis? This needs to be explained or changed. There are also differences in the questions asked to each sample which were not explained; on what basis did the author choose to do so? Any benefits? Wouldn't that increase variability of the results among the samples?

Response: Thank you for your comments regarding the presentation of the data. In the original manuscript, we described Sample 1 as the undergraduate student sample and Sample 2 as the online panel worker sample rather than describe them as participants from two distinct studies because we sought to simplify the discussion of methods/measures (which were nearly identical) and results. Based on your comment, however, we have decided to present the data as two separate “studies” (Study 1 corresponding to the undergraduate sample and Study 2 corresponding to the Prime Panel sample). Although we did not hypothesize specific differences to emerge between the two samples, there are several notable differences between the two sets of data that warrant separate analyses.

First, Study 2 was a pre-registered study intended to replicate the effects observed in Study 1 with a new sample. Second and relatedly, we fixed minor technical issues from the first survey and added new/modified items to Study 2. For example, our primary dependent measure of COVID-19 vaccination intention was asked with two items in Study 2 instead of using a single-item measure (see p. 17 lines 341-350). Third and most importantly, these studies were conducted at different time points during the Pandemic. Study 1 was conducted during the initial months of the Pandemic (February-April 2020) when there was much uncertainty and general lack of information about the virus (e.g., infectiousness, symptoms, etc.) and while U.S. cases were growing exponentially and lockdowns had just begun; in contrast, Study 2 was collected on July 20, 2020—roughly 6 months after the first U.S. laboratory-confirmed COVID case and cases were steadily growing. Notably, Study 2 was conducted during Phase II clinical trials toward the end of the first U.S. wave when more information about the virus and about vaccine development were available, and when people were likely starting to get used to the “new normal.” Because willingness to receive a COVID vaccine has fluctuated over time in the U.S. population with reported percentages varying substantially depending on when the poll was taken (e.g., Fridman et al., 2021; Funk & Tyson, 2020; Tyson et al., 2020), an analysis combining the two datasets may produce potentially misleading conclusions.

In sum, for succinctness, we presented the results using measures that were collected for both samples, which may have given the impression that we “split” the sample. However, these data were collected in two distinct studies conducted at different points during the pandemic. To clarify that the data are distinct in time and sample characteristics, we now treat them as two separate “studies” in our revised manuscript and further emphasize that Study 2 is a replication of Study 1 (see p. 3 line 62, p. 16 lines 315-318). Moreover, we highlight the improvements made in the second survey based on the results from the first survey (see p. 10, lines 228-229) and include a new variable about concerns regarding COVID-19 vaccine (see p. 17 lines 351-354), which was originally described in the Online Supplementary Materials. Finally, we have restructured the manuscript such that participants, procedures, measures, and results for Study 1 and Study 2 are described separately.

7. The manuscript reads more like a thesis than a concise manuscript it lacks “straight to the point” presentation of data and clear conclusions. I recommend rewriting the manuscript, especially the introduction, which should be one coherent piece of work, also the methods and the results. The conclusion in the abstract will need reconsideration as it starts with a vague sentence and does not give any concise conclusion

Response: Thank you for your feedback. We have made substantial revision to the manuscript following your suggestions. Specifically, the introduction now succinctly describes the objectives and motivation for the research. We deleted descriptions of related works that were not essential to the background of the current research, reducing the length of the introduction from 922 words to 606. As mentioned earlier, we have entirely restructured the methods and results section such that the two studies are now described separately. Moreover, we have added the “key findings” section in the General Discussion to highlight the take-home message (see p. 23 lines 451-463). We also revised the conclusion in the abstract, and now it read as follows: “Encouraging flu vaccine uptake, enhancing confidence in a novel vaccine, and fostering a sense of collective responsibility are particularly important as they uniquely predict COVID-19 vaccination intentions.” (pp. 3-4 lines 75-77).

8. The questionnaires used are very lengthy, and despite the attention checks the authors have described, no information has been provided regarding how many participants were excluded due to failing the attention check; they only report a general total number of 38 excluded for different reasons for sample one and same regarding sample 2 were 172 were excluded for several reasons.

Response: Thank you for your feedback. As described in the supplementary file (OSM p. 10), there were two attention checks in Study 1. The first item asked: If you’re reading this, please check “Strongly Agree.” Out of 346 respondents, n=34 failed this question and selected a response other than “Strongly Agree.” The second item was displayed at the end of the survey and asked whether they took the study seriously and responded thoughtfully. Out of 346 respondents, n=10 responded that they did not take the study seriously. In summary, 28 respondents failed the first item only, 4 respondents failed the second item only, and 6 respondents failed both items, leaving us with the final sample size of N=346-28-4-6=308 consisting of participants who passed both attention check items.

For Study 2, we had three pre-registered exclusion criteria: respondents who 1) do not finish the survey, 2) fail to answer one or more attention checks correctly, and 3) fail to spend more than 10 seconds on pages involving reading and/or responding to multiple items. We received 848 complete responses in Study 2. The first attention check item was identical to Study 1. Out of 848 respondents, n=117 failed to select “Strongly Agree.” The second attention check item asked respondents to select “67” on a sliding scale. Out of 848 respondents, n=109 provided a response other than 67. In summary, 58 respondents failed the first item only, 50 respondents failed the second item only, and 59 respondents failed both items, leaving us with n=848-58-50-59=681 respondents who passed both attention check items in Study 2. Finally, we checked the average reading time and excluded n=5 respondents who spent less than 10s per page. Thus, the final sample size for Study 2 is N=681-5=676.

Because footnotes are not allowed in PLOS ONE and to further increase concision, we direct readers in the main text (p. 8 line 187, p. 16 lines 321-322) to see the supplementary file, which now contains the above details regarding exclusions for Study 1 (OSM p. 18) and Study 2 (OSM pp. 20-21).

9. Line 85: "Whereas previous research has examined attitudes about vaccination for diseases that are hypothetical (e.g., Haase et al., 2015; Jolley & Douglas, 2017), relatively low-threat (e.g., seasonal flu; Zhang et al., 2010) or mostly eradicated in the U.S. (e.g., Dubé et 88 al., 2015), the present research investigates vaccination intentions during a unique period of uncertainty involving a highly contagious disease when effective treatments or vaccines had not yet been developed." The authors should report the published evidence on vaccine intentions in the COVID-19 era

Response: Thank you for your suggestion. In this statement, we wanted to make a point that our research (as well as other research conducted during the pandemic) is different from past research on vaccine hesitancy conducted prior to the pandemic. We have modified this sentence as follows (p. 5, lines 111-116): “Whereas vaccine hesitancy research prior to the pandemic examined attitudes about vaccination for diseases that are hypothetical, relatively low-threat (e.g., seasonal flu), or mostly eradicated in the U.S. (e.g., polio), the present research examined which psychological factors are associated with intention to receive a novel, pandemic vaccine in a context where no one was immune…”

However, we agree that more research on vaccine hesitancy specific to COVID-19 has been published since we originally drafted this manuscript. Therefore, we have incorporated these recent works in the introduction (see pp. 6-7 lines 131-149), results (see p. 19 lines 387-388, p. 20 lines 412-415, p. 21 lines 438-439), and discussion (p. 23 lines 458-460).

10. The authors mention that” Sample 1 included 346 undergraduate students from the University of Illinois at Urbana Champaign recruited in exchange for course credit” Can you elaborate more on what do you mean here? Were the participants given a choice to join? How did you guarantee that participants were aware that their participation is optional?

Response: In Study 1, all participants were recruited from an undergraduate psychology course-credit subject pool at the University of Illinois Urbana-Champaign in Spring 2020. Undergraduate students who are 18 or older taking one or more psychology courses have the option to enroll in the subject pool to participate in research studies for course credit or to complete an alternative assignment with similar duration. This information is provided in the course syllabus and the informed consent that participants read prior to starting the study. Participants in the subject pool individually sign up to complete studies on their own time. Participation is voluntary, and students are allowed to withdraw from the study for any reason without any penalty. We have clarified our recruitment procedure for Study 1 on p. 8 (lines 175-185) of the revised manuscript. Moreover, we note that data collection using undergraduate subject pools is a standard practice in psychology and related fields within the social sciences more broadly (e.g., business, consumer research, economics).

11. The discussion section can be improved by highlighting the take-home message from the study and take the reader through what has been done and what does that mean rather than displaying the results. As an example in line 418: "knowledge, no prior study has simultaneously measured vaccine attitudes for both a hypothetical disease and a real disease during a pandemic. Measuring both together and determining which variables predict well across measures is important, as it provides context for past work and can bolster confidence in how well past research applies to actual intentions during a crisis period…… The current work shows that factors associated with immunization decisions for a hypothetical disease are also mostly predictive of real vaccination intentions." I do not see the point here? Can you elaborate on the reasons behind studying vaccine attitudes for both hypothetical disease and a real disease during a pandemic? Not at any point of the manuscript, this was discussed clearly

Response: Thank you for this suggestion. We have added a “key findings” section in the General Discussion to emphasize the take-home message on p. 23 (lines 451-463). In addition, we have revised the paragraph you referred to above (p. 24, lines 478-489) to clarify our point regarding the importance of measuring vaccination attitudes for both a hypothetical disease and a real disease. We argue that measuring both together is important for providing context for past work regarding how well they generalize to actual intentions during a crisis period. That is, studying vaccination attitudes about a novel yet hypothetical disease (e.g., Haase et al., 2015; Jolley & Douglas, 2017) lacks ecological validity, and studying vaccination intentions for real, more “established” diseases for which vaccines have existed for years (and most people are inoculated) may not provide the same insights as attitudes towards a novel pandemic vaccine. The current work overcame these limitations, also finding that factors associated with immunization decisions for a hypothetical disease are mostly predictive of real vaccination intentions. By asking about both hypothetical and real diseases, this research has improved our understanding of the utility of previous work in its application to the current pandemic. We also articulate this point earlier on in the manuscript (see pp. 5-6 lines 117-130).

12. The tables are confusing with too many details; I suggest editing them to highlight the most important message from each table and move any unnecessary information to the supplementary file.

Response: Thank you for your feedback. We have moved the Table 1-2 from the original version to the supplementary file, and we now present a simpler version as Table 1 (p. 13). We bolded the significant correlations so that they are easily noticeable. Similarly, we have simplified the tables describing the regression models (Table 2 on p. 15 and Table 4 on p. 22). The original tables each contained results of 30 distinct models, but the revised tables each contain 4 only models.

13. In the supplementary file, Why was the attention check different between samples 1 &2? What did you base this on?

For Sample 1, the following item was asked prior to the demographics questionnaire: Thank you for your responses. You are almost finished with this study. Before answering a few questions about yourself, we have one last question. Do you think we should include your responses in our study? That is, did you take the study seriously and respond thoughtfully? Your credit assignment for this study does not depend on your response to this question. [Yes, I responded thoughtfully / No, I did not respond thoughtfully.]

Sample 2 Only

For Sample 2, the following item was embedded in the optimism bias measure:

If you are reading this, please select 67 on this sliding scale.

Response: We appreciate your inquiry regarding the attention checks and for taking the time to review the supplementary materials. There are three reasons why we decided to alter our approach in Study 2. First, we noticed from our previous work that has used online panel workers that almost all participants claim that they responded thoughtfully even when we explicitly tell them that their payment will not be affected on the basis of their response. Thus, we decided to drop this particular question which did not seem to be providing much information about actual attentiveness. Second, for a question with two response options (“Yes, I responded thoughtfully” vs. “No, I did not respond thoughtfully”), the chance of answering “Yes” is 50% even if a participant were to select a random response. With a slider scale like above, there is only 1% chance that a non-attentive respondent will accidentally move the slider to the correct number. To ensure obtaining quality data, we decided to have a stricter attention check item in Study 2. Third, there were other measures (including the main COVID-19 vaccination intention item) that used a slider scale in Study 2, whereas all questions were in multiple-choice or Likert-type scale formats in Study 1. An ancillary reason for using the slider scale as an attention check was to confirm that participants were able to indicate their desired response using the slider format. We have noted these reasons in the supplementary file (OSM p. 10). Finally, and importantly, we also note that these attention check items were specified in our pre-registration.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

9 Nov 2021

Psychological predictors of vaccination intentions among U.S. undergraduates and online panel workers during the 2020 COVID-19 Pandemic

PONE-D-21-22539R1

Dear Dr. Watanabe,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Ismaeel Yunusa, PharmD, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for your contributions!

Reviewers' comments:

17 Nov 2021

PONE-D-21-22539R1

Psychological predictors of vaccination intentions among U.S. undergraduates and online panel workers during the 2020 COVID-19 Pandemic

Dear Dr. Watanabe:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Ismaeel Yunusa

Academic Editor

PLOS ONE