COVID-19 Vaccine Hesitancy

Last registered on July 25, 2022


Trial Information

General Information

COVID-19 Vaccine Hesitancy
Initial registration date
January 27, 2022

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
January 28, 2022, 11:01 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
July 25, 2022, 12:00 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.



Primary Investigator

University of Arkansas

Other Primary Investigator(s)

PI Affiliation
University of Nottingham
PI Affiliation
University of Nottingham
PI Affiliation
University of Nottingham
PI Affiliation
University of Nottingham

Additional Trial Information

Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Amidst continually increasing evidence in favour of vaccination, hesitancy is sharply divided along political lines. This project will quantitatively assess to what degree vaccine hesitancy can be attributed to (1) a lack of information on COVID-19 and its vaccines, and (2) politically motivated reasoning and exposure to abundant misinformation. Using a large-scale randomized control trial, we will test scalable interventions that provide basic information about vaccines and other interventions designed to combat misinformation and biased reasoning on vaccination. Ultimately, we hope to deliver communication tools that allow healthcare providers to identify which individuals will be receptive to good-faith conversations on the safety and efficacy of COVID-19 vaccination as well as productively engage those who are otherwise determined to reject any information that is perceived to be pro-vaccination. Our approach will adapt the communication tools of “Paradoxical Reasoning” which were developed to unfreeze entrenched beliefs in the context of the most intractable conflicts (e.g., the Israeli-Palestinian border conflict).
External Link(s)

Registration Citation

Brownback, Andy et al. 2022. "COVID-19 Vaccine Hesitancy." AEA RCT Registry. July 25.
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details


The first intervention is the viewing of tailored informational videos on COVID vaccines designed to fill informational gaps. The second intervention is a ``paradoxical reasoning'' protocol that asks respondents to engage with reductio-ad-absurdum arguments to soften their positions on COVID vaccination.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Our primary outcome variables are derived from the COVID-19 Perceptions in Section 3.1. We will use these to construct three primary indices for evaluating perceptions. These are listed below in descending order of importance:

1. Perception1: Vaccination intentions and concerns: (5, 6)
2. Perception2: Vaccination efficacy: (1, 2)
3. Perception3: Vaccination side effects: (7)

Each index will be constructed by taking the mean of the relevant variables, and then standardizing this mean value across subjects. Our indices will have a mean of 0 and standard deviation of 1. Perception1 − Perception3 will form our primary analysis.

Perception1 is our most important outcome. Vaccination intentions and concerns are our primary policy-relevant concerns. Next, we hope to influence beliefs about vaccine efficacy (Perception2)—primarily with the hope that it encourages greater vaccine take-up. Concerns about side effects of vaccination (Perception3) are a second-order concern but important in how they may limit vaccination efforts. We note that only subjects who elect to view extra videos will receive information about side effects, so we interpret results around this index with caution.

We exclude certain perception questions from these indices because they may result in ambiguous predictions. For example, for item (3), a subject may become less concerned about the severity of COVID-19 because they have become convinced of the effectiveness of COVID-19 vaccines. For supplementary analysis, we will include a more detailed approach that breaks apart each perception question individually. We will also explore a fourth index, Perception4: Trust in institutions: (8, 9, 10).

We are also interested in the information-seeking behaviors of our subjects. To capture this, two of our video treatments (V1 and V2) allow subjects to elect to watch additional videos after their first mandatory video. We measure “information-seeking” as the number of subsequent videos watched. As a secondary measure, we will also analyze the amount of time spent watching those videos.

1. Seeking1: Number of additional videos watched (0 to 4)
2. Seeking2: Time spent watching additional videos (supplementary)
Primary Outcomes (explanation)
Our perception index consists of questions from the following list:

1. Effectiveness of vaccines (infection)
2. Effectiveness of vaccines (hospitalization) 3. Risk perception of COVID-19 (severity) 4. Risk perception of COVID-19 (likelihood) 5. Preference for “natural immunity”
6. Vaccine intentions
7. Concern about vaccine side effects
8. Trust (in doctors)
9. Trust (in FDA)
10. Trust (in CDC)

Secondary Outcomes

Secondary Outcomes (end points)
See attached document
Secondary Outcomes (explanation)
See attached document

Experimental Design

Experimental Design
In this large-scale online experiment, we will explore the impact of two primary interventions. The first intervention is the viewing of tailored informational videos on COVID vaccines designed to fill informational gaps. The second intervention is a “paradoxical reasoning” protocol that asks respondents to engage with reductio-ad- absurdum arguments to soften their positions on COVID vaccination.
Experimental Design Details
Randomization Method
Cross-randomization to ensure subjects are assigned evenly across all possible combinations of treatments.
Randomization Unit
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
4000 subjects

(Update for February 16th): 3000 additional subjects who have not previously participated.
Sample size: planned number of observations
4000 subjects in the initial main survey + 4000 subjects in a follow-up to the initial survey. All 4000 will be included in a follow-up survey, but we anticipate some attrition, so the total number of observations will likely be short of 8000. (Update for February 16th): We will be conducting an additional survey to better identify effects of certain interventions. This will involve an additional 3000 subjects who have not previously participated.
Sample size (or number of clusters) by treatment arms
Video Treatments:
1. Video Control Condition (VC )(N ≈ 1333):
2. Video Optional Condition (V1)(N ≈ 1333):
3. Video Required Condition (V2)(N ≈ 1333):

Paradox treatments:

1. Paradox Control Condition (PC )(N ≈ 1000):
2. Live Virus Paradox (PLV )(N ≈ 1000):
3. Long-Run Paradox (PLR)(N ≈ 1000):
4. Both Paradoxes (PB)(N ≈ 1000):

So, each cross-randomization cell should have a sample of approximately 333 subjects.

Motivated reasoning randomization:

1. GENERAL (N ≈ 1333)
2. MYOCARDITIS (N ≈ 1333)
3. THROMBOSIS (N ≈ 1333)

So, the full 3X4X3 factorial design should have approximately 111 subjects in each cell. However, the even cross-randomization will allow us to identify average treatment effects for each treatment without being restricted to such small samples.

(Update for February 16th): In our additional survey, we will simplify the design to better understand the additivity of interventions. These 3000 additional subjects will be assigned using a 2x2 randomization.

Paradox treatment random assignment:
1. Paradox Control Condition (PC ) (N ≈ 1500)
2. Both Paradoxes (PB) (N ≈ 1500)

Video treatment independent random assignment:
1. Video Control Condition (VC ) (N ≈ 1500)
2. Video Required (VR) (N ≈ 1500)

Thus, each cell of the 2X2 for the additional study will contain approximately 750 subjects.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
In pilot studies of approximately 400 subjects, our primary research questions were already marginally statistically significant. So, if our primary sample exhibits similar characteristics, our sample sizes are far in excess of what we would need. However, our main sample will be recruited through a different survey firm. So, the uncertainties about statistical significance arise more from differences across survey firms rather than from sample sizes.

Institutional Review Boards (IRBs)

IRB Name
University of Nottingham
IRB Approval Date
IRB Approval Number
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials