Skip to main content

The effect of social media communication on intention to vaccinate against COVID-19: A relationship mediated by perceptions of vaccine efficacy and safety

Journal of Marketing Communications, 2025

Dooyeon Park, Bruno Schivinski, Hilary Duns-McKay, Bartholomew Heeren, Elizabeth McLachlan

Journal of Marketing Communications, volume 31, issue 1, pages 1–20, 2025 (online 27 March 2023)

  • Open access
  • 428 adults in Australia
  • Online survey
  • Structural equation modelling
  • Mediation analysis
  • COVID-19
Download the full text (PDF)Published version

Free full text: the open access article (CC BY-NC-ND 4.0), re-typeset as a clean PDF with unchanged content. Please cite the published version.

Keywords
  • COVID-19 vaccination
  • social media
  • perceived vaccine efficacy
  • perceived vaccine safety
  • risk perception
  • intention to vaccinate
  • vaccine hesitancy
  • health communication
  • official sources on social media
  • user-generated content
  • source credibility
  • misinformation
  • public health campaigns
  • social media communication (SMC)
  • source credibility theory
  • Oxford COVID-19 vaccine confidence and complacency scale
  • social media communication scale (Schivinski and Dabrowski)
  • Australia
  • pandemic
  • online survey
  • structural equation modelling (SEM)
  • mediation analysis
  • bootstrapping
  • confirmatory factor analysis
  • lavaan
  • R

Positive views of vaccine information shared on social media go with a stronger intention to get vaccinated, because they make the vaccine seem effective and safe. In a survey of 428 adults in Australia in 2021, this held for both official sources and posts by ordinary users; fear of COVID-19 did not carry the effect.

Key findings

  1. Official vaccine information on social media goes with intention to vaccinate. Among 428 adults in Australia, positive evaluations of official sources’ social media communication about COVID-19 vaccines had a total effect on intention to take the vaccine of β = 0.28 (p = .001); the direct effect, reported as positive in the article, was β = 0.05 (p = 0.10) (Park et al., 2023).
  2. User-generated posts work only through beliefs about the vaccine. User-generated social media communication had no direct effect on intention (β = −0.01, p = .87) but raised it indirectly through perceived vaccine efficacy (indirect β = 0.08, p = .01) and safety (0.02, p = .03): full mediation (Park et al., 2023).
  3. Perceived efficacy is the strongest route. Perceived vaccine efficacy was the strongest predictor of intention (β = 0.57) and carried most of the effect of official sources (indirect β = 0.16, 95% CI 0.09–0.24); perceived safety added a smaller route (0.04) (Park et al., 2023).
  4. Risk perception did not mediate. Official posts raised COVID-19 risk perception (β = 0.20), but risk perception did not predict intention (β = 0.06, p = .14), so neither indirect path through it was significant (Park et al., 2023).

At a glance

Design
Cross-sectional online survey on Qualtrics, late August to early September 2021, shared on Facebook groups and pages (convenience and snowball sampling); structural equation model with mediation, 5,000 bootstrap draws, in R (lavaan)
Participants
428 adults living in Australia (from 554 responses; 126 with six or more missing items excluded), aged 16 to 79 (mean 46.2), 53.2% women, 51.6% from Victoria; 46.9% had two vaccine doses and 27.3% one
Predictors
Evaluations of official source and of user-generated social media communication about the vaccines, 4 items each (adapted from Schivinski & Dabrowski, 2016), α = 0.87 and 0.90
Mediators
Perceived vaccine efficacy (8 items) and safety (5 of 6 items) from the Oxford COVID-19 vaccine confidence and complacency scale (Freeman et al., 2020); COVID-19 risk perception (5 of 6 items, adapted from Yang & Cho, 2017)
Outcome
Intention to take a COVID-19 vaccine, including boosters (3 items, adapted from Huynh & Senger, 2021), α = 0.85
Controls
Gender, age and time spent on social media (none significant)
Model fit
CFA: ML χ²(362) = 800.50, CFI = 0.95, TLI = 0.94, RMSEA = 0.05. Structural model: ML χ²(364) = 1286.66, CFI = 0.91, TLI = 0.90, RMSEA = 0.06; R² for intention = 0.73

The model, with the standardised estimates reported in Table 2

0.29***0.28***0.20***0.14*0.15**0.040.57***0.15***0.06Official sourcesocial mediaUser-generatedsocial mediaVaccine efficacyVaccine safetyCOVID-19 riskperceptionIntention to takea COVID-19 vaccineR² = 0.73
Drawn from Table 2 of the article (the article has no figure). Grey dashed paths are not significant. The direct effects of official source (β = 0.05, p = .10) and user-generated social media communication (β = −0.01, p = .87) on intention are not drawn. * p < .05, ** p < .01, *** p ≤ .001; n = 428.

Indirect effects on intention to vaccinate

Official source social mediaVia perceived efficacy0.16Via perceived safety0.04Total effect0.28User-generated social mediaVia perceived efficacy0.08Via perceived safety0.02Total effect (n.s.)0.09
Standardised indirect effects from Table 2 of the article (n = 428; 5,000 bootstrap draws). Paths through COVID-19 risk perception (0.01) were not significant.

Tables

Table 1. Construct reliability and validity outputs

Construct α CR AVE 1 2 3 4 5 6
1. Official source SMC 0.87 0.88 0.66 0.81
2. User-generated SMC 0.90 0.90 0.69 0.40 0.83
3. Vaccine efficacy 0.88 0.83 0.50 0.33 0.17 0.71
4. Vaccine safety 0.84 0.92 0.61 0.29 0.21 0.72 0.78
5. COVID-19 risk perception 0.82 0.86 0.56 0.23 0.06 0.74 0.59 0.75
6. Intention to take COVID-19 vaccine 0.85 0.89 0.73 0.28 0.14 0.84 0.60 0.67 0.85

SMC = social media communication; α = Cronbach’s alpha; CR = composite reliability; AVE = average variance extracted; square roots of the AVE on the diagonal (in italics). ML χ²(362) = 800.50, CFI = 0.95, TLI = 0.94, RMSEA = 0.05 (90% CI 0.04–0.05); n = 428. Values as printed in the article.

Table 2. Structural results

Path β t / 95% CI p
Direct effects
H1. Official source SMC → intention to take COVID-19 vaccine 0.05 1.63 0.10
H2. User-generated SMC → intention to take COVID-19 vaccine −0.01 −0.16 0.87
a-path 1. Official source SMC → vaccine efficacy 0.29 5.97 0.001
a-path 2. Official source SMC → vaccine safety 0.28 5.89 0.001
a-path 3. Official source SMC → COVID-19 risk perception 0.20 4.15 0.001
a-path 4. User-generated SMC → vaccine efficacy 0.14 2.60 0.01
a-path 5. User-generated SMC → vaccine safety 0.15 2.94 0.003
a-path 6. User-generated SMC → COVID-19 risk perception 0.04 0.08 0.94
b-path 1. Vaccine efficacy → intention to take COVID-19 vaccine 0.57 10.29 0.001
b-path 2. Vaccine safety → intention to take COVID-19 vaccine 0.15 3.08 0.001
b-path 3. COVID-19 risk perception → COVID-19 vaccine intention 0.06 1.46 0.14
Indirect effects (indirect β, 95% CI)
H3. Official source SMC → vaccine efficacy → intention 0.16 0.09–0.24 0.001
H4. Official source SMC → vaccine safety → intention 0.04 0.01–0.08 0.005
H5. Official source SMC → COVID-19 risk perception → intention 0.01 −0.01–0.03 0.17
H6. User-generated SMC → vaccine efficacy → intention 0.08 0.01–0.14 0.01
H7. User-generated SMC → vaccine safety → intention 0.02 0.01–0.04 0.03
H8. User-generated SMC → COVID-19 risk perception → intention 0.01 −0.01–0.01 0.93
Total effects
Official source SMC → intention to take COVID-19 vaccine 0.28 5.77 0.001
User-generated SMC → intention to take COVID-19 vaccine 0.09 1.88 0.06
Control variables
Age → intention to take COVID-19 vaccine −0.03 −1.06 0.28
Gender → intention to take COVID-19 vaccine 0.01 0.27 0.78
Social media usage → intention to take COVID-19 vaccine −0.02 −0.29 0.80

ML χ²(364) = 1286.66, CFI = 0.91, TLI = 0.90, RMSEA = 0.06 (90% CI 0.06–0.07); a-path = independent variable to mediator; b-path = mediator to dependent variable; 5,000 bootstrap draws; n = 428. Values as printed in the article.

Table A. Descriptive statistics for the constructs and measurements
Construct and item λ Mean SD Skew. Kurt.
Official sources of social media communication
I am satisfied with their social media communications about the vaccines 0.90 2.80 1.21 0.04 −0.96
The level of their social media communication about the vaccines meets my expectations 0.83 2.86 1.18 −0.10 −0.82
The level of information they provide about the vaccines on social media is very good 0.83 2.79 1.18 0.05 −0.92
Official social media communications about the vaccines perform well when compared with other types of communications, e.g., TV, radio, newspapers 0.70 2.99 1.19 −0.07 −0.78
User-generated social media communication
I am satisfied with their social media communications about the vaccines 0.90 2.79 1.11 −0.03 −0.60
The level of their social media communication about the vaccines meets my expectations 0.85 2.86 1.05 −0.15 −0.41
The level of information they provide about the vaccines on social media is very good 0.87 2.72 1.13 0.04 −0.66
Social media communications about the vaccines from family and friends perform well when compared with official social media communications 0.71 2.76 1.13 −0.01 −0.66
Vaccine hesitancy: vaccine efficacy
It will be helpful for my community if I get a COVID-19 vaccine 0.92 4.57 1.06 −2.65 5.93
It will save lives if I get a COVID-19 vaccine 0.93 4.39 1.13 −2.04 3.21
It will be dangerous if people over 16 don’t get a COVID-19 vaccine 0.81 4.21 1.24 −1.61 1.45
A COVID-19 vaccine will strengthen my immune system 0.70 3.98 1.32 −1.13 0.08
Taking a COVID-19 vaccine will give me greater freedom 0.72 4.32 1.08 −1.89 3.01
I think I will be infected with COVID-19 over the next 12 months 0.62 2.85 1.08 −0.01 −0.60
A COVID-19 vaccine will work for everyone 0.68 3.33 1.32 −0.36 −1.08
A COVID-19 vaccine will work for me 0.84 4.38 1.06 −1.92 3.08
Vaccine hesitancy: vaccine safety
The speed of developing and testing COVID-19 vaccines means they will be good 0.62 3.19 1.17 −0.21 −0.44
The speed of developing and testing COVID-19 vaccines means they will be safe 0.66 3.33 1.24 −0.27 −0.71
Getting a COVID-19 vaccine is a sign of personal strength 0.72 3.20 1.28 −0.31 −0.74
I expect that receiving a COVID-19 vaccine will be hardly noticeable 0.75 3.59 1.23 −0.49 −0.80
The side effects of a COVID-19 vaccine are mild 0.78 3.80 1.22 −0.90 −0.15
Taking a COVID-19 vaccine will make me feel like a guinea pig* – 4.24 1.22 −1.43 0.73
COVID-19 risk perception
I think I will contract COVID-19 if I come into contact with a COVID-19 patient 0.92 3.93 1.04 −1.08 0.87
I think that I might contract COVID-19 even if I do not come into contact with a COVID-19 patient 0.89 2.33 1.28 0.57 −0.89
My health will be severely damaged if I contract COVID-19 0.60 3.77 1.14 −0.92 0.22
I think COVID-19 is more severe than any other respiratory disease 0.62 3.86 1.28 −0.97 −0.19
Even if I fall ill with other ailments, I will avoid going to hospital because of COVID-19* – 2.09 1.28 0.83 −0.68
COVID-19 will inflict serious damage on my community 0.67 4.18 1.15 −1.49 1.38
Intention to take the COVID-19 vaccine
I intend to take a COVID-19 vaccine 0.88 4.66 0.96 −2.95 7.74
If I had the opportunity to take a COVID-19 vaccine today, I would take it 0.81 4.55 1.13 −2.45 4.60
I plan to take a COVID-19 vaccine 0.88 4.63 0.99 −2.80 6.75

λ = standardised factor loading; * item removed from the analyses for a low loading. All items rated from 1 = strongly disagree to 5 = strongly agree. Items and values as printed in Appendix A of the article.

The question

Social media can spread both accurate vaccine information and misinformation, and studies of its effect on vaccination have been inconsistent, mostly treating all social media content as the same. During the 2021 COVID-19 vaccination campaign in Australia, we asked whether vaccine information from official sources (such as health authorities, governments and news organisations) and from ordinary users relates differently to intention to vaccinate, and whether perceived vaccine efficacy, perceived safety and risk perception explain the link.

Implications

Practical implications

Health agencies should use social media actively: official sources’ credibility carries over to the platform, making it an effective way to reach people looking for vaccine information. Messages should focus on how well and how safely vaccines work rather than on fear of the disease, and communicators can encourage user-generated content that shares accurate information about efficacy and safety, since even posts by ordinary users raise intention when they shape those beliefs.

Theoretical implications

The study separates social media communication by source and shows that the source matters for vaccine intention, which may explain the inconsistent findings of earlier studies that treated social media as one channel. It identifies perceived vaccine efficacy and safety, not risk perception, as the beliefs through which social media communication relates to intention to vaccinate.

Limitations and future research

  • The convenience and snowball sample recruited on Facebook may not represent the Australian population.
  • The data are cross-sectional, so causal links between social media communication and intention cannot be established.
  • Differences between groups (for example by vaccination status or country) were not examined; the availability of official sources on social media may vary by country.

Questions this paper answers

Does social media make people more willing to get vaccinated?

It depends on the source and on what people come to believe. Among 428 adults in Australia, positive views of official vaccine information on social media went with a stronger intention to vaccinate, mostly because they raised perceived vaccine efficacy and safety (Park et al., 2023).

Do posts by ordinary users affect vaccine intentions?

Not directly. User-generated social media communication had no direct effect, but it raised intention indirectly by improving perceptions of how effective and safe the vaccine is (Park et al., 2023).

Does fear of COVID-19 explain vaccination intentions?

Not in this study: COVID-19 risk perception did not predict intention to vaccinate and did not mediate the effects of social media communication (Park et al., 2023).

What should health campaigns on social media focus on?

The authors recommend focusing on vaccine efficacy and safety, using official sources’ credibility on social media, and encouraging accurate user-generated content (Park et al., 2023).

Key concepts

Social media communication (SMC)
Content about a topic shared on social media, here evaluated by how satisfied people are with it, whether it meets their expectations and how good its information is.
Official source vs user-generated SMC
Content from health authorities, governments and news organisations versus content created by ordinary users such as family and friends.
Perceived vaccine efficacy
The belief that a vaccine works to protect oneself and others.
Perceived vaccine safety
The belief that a vaccine is safe, including views on how quickly it was developed and on its side effects.
Risk perception
How likely and how severe a person believes catching a disease would be.
Vaccine hesitancy
Delay in accepting, or refusal of, vaccination despite its availability.

Open science and materials

Study materials

Scale items
All items, with standardised loadings, means, standard deviations, skewness and kurtosis, are in Table A below (all on 5-point agreement scales). Two items were removed for low loadings: ‘Taking a COVID-19 vaccine will make me feel like a guinea pig’ (safety, λ = 0.12) and ‘Even if I fall ill with other ailments, I will avoid going to hospital because of COVID-19’ (risk, λ = 0.05).
Eligibility
Respondents had to live in Australia and be older than 16, the age from which COVID-19 vaccination was available at the time.

How to cite

Park, D., Schivinski, B., Duns-McKay, H., Heeren, B., & McLachlan, E. (2025). The effect of social media communication on intention to vaccinate against COVID-19: A relationship mediated by perceptions of vaccine efficacy and safety. Journal of Marketing Communications, 31(1), 1–20. https://doi.org/10.1080/13527266.2023.2191618

Open access under a Creative Commons CC BY-NC-ND 4.0 licence.

Related publications

All publications