Actions Speak Louder Than Words: Sentiment and Topic Analysis of COVID-19 Vaccination on Twitter and Vaccine Uptake.

Murooj Yousef, Timo Dietrich, Sharyn Rundle-Thiele
Author Information
  1. Murooj Yousef: Social Marketing @ Griffith, Department of Marketing, Griffith University, Nathan, Australia. ORCID
  2. Timo Dietrich: Social Marketing @ Griffith, Department of Marketing, Griffith University, Nathan, Australia. ORCID
  3. Sharyn Rundle-Thiele: Social Marketing @ Griffith, Department of Marketing, Griffith University, Nathan, Australia. ORCID

Abstract

BACKGROUND: The lack of trust in vaccines is a major contributor to vaccine hesitancy. To overcome vaccine hesitancy for the COVID-19 vaccine, the Australian government launched multiple public health campaigns to encourage vaccine uptake. This sentiment analysis examines the effect of public health campaigns and COVID-19-related events on sentiment and vaccine uptake.
OBJECTIVE: This study aims to examine the relationship between sentiment and COVID-19 vaccine uptake and government actions that impacted public sentiment about the vaccine.
METHODS: Using machine learning methods, we collected 137,523 publicly available English language tweets published in Australia between February and October 2021 that contained COVID-19 vaccine-related keywords. Machine learning methods were used to extract topics and sentiments relating to COVID-19 vaccination. The relationship between public vaccination sentiment on Twitter and vaccine uptake was examined.
RESULTS: The majority of collected tweets expressed negative (n=91,052, 66%) rather than positive (n=21,686, 16%) or neutral (n=24,785, 18%) sentiments. Topics discussed within the study time frame included the role of the government in the vaccination rollout, availability and accessibility of the vaccine, and vaccine efficacy. There was a significant positive correlation between negative sentiment and the number of vaccine doses administered daily (r=.15, P<.05), with positive sentiment showing the inverse effect. Public health campaigns, lockdowns, and antivaccination protests were associated with increased negative sentiment, while vaccination mandates had no significant effect on sentiment.
CONCLUSIONS: The study findings demonstrate that negative sentiment was more prevalent on Twitter during the Australian vaccination rollout but vaccine uptake remained high. Australians expressed anger at the slow rollout and limited availability of the vaccine during the study period. Public health campaigns, lockdowns, and antivaccination rallies increased negative sentiment. In contrast, news of increased vaccine availability for the public and government acquisition of more doses were key government actions that reduced negative sentiment. These findings can be used to inform government communication planning.

Keywords

References

  1. J Thorac Dis. 2016 Sep;8(9):E928-E931 [PMID: 27747028]
  2. Int J Environ Res Public Health. 2021 Jun 01;18(11): [PMID: 34206119]
  3. Pediatr Ann. 2010 Aug;39(8):476-82 [PMID: 20704143]
  4. Lancet Reg Health West Pac. 2021 Sep;14:100250 [PMID: 34423325]
  5. Intern Med J. 2021 Dec;51(12):1987-1989 [PMID: 34713544]
  6. J Paediatr Child Health. 2021 Apr;57(4):472-473 [PMID: 33817870]
  7. Aust Fam Physician. 2004 Jul;33(7):568-71 [PMID: 15301182]
  8. PLoS One. 2021 Jan 20;16(1):e0245907 [PMID: 33471821]
  9. Expert Rev Vaccines. 2015 Jan;14(1):99-117 [PMID: 25373435]
  10. BMC Public Health. 2021 Apr 28;21(1):818 [PMID: 33910558]
  11. Clin Infect Dis. 2022 Apr 9;74(7):1166-1173 [PMID: 34292319]
  12. Biostatistics. 2000 Sep;1(3):279-92 [PMID: 12933509]
  13. PLoS One. 2021 May 12;16(5):e0251605 [PMID: 33979412]
  14. J Med Internet Res. 2021 May 19;23(5):e26953 [PMID: 33886492]
  15. Aust J Rural Health. 2020 Apr;28(2):149-158 [PMID: 31970832]
  16. Cureus. 2018 Jul 3;10(7):e2919 [PMID: 30186724]
  17. PLoS One. 2021 Mar 24;16(3):e0248892 [PMID: 33760836]
  18. Med Mal Infect. 2016 May;46(3):117-22 [PMID: 26987960]
  19. Health Commun. 2022 May;37(5):628-636 [PMID: 33349050]
  20. Hum Vaccin Immunother. 2014;10(9):2600-2 [PMID: 25483479]
  21. Hum Vaccin Immunother. 2018 Jul 3;14(7):1654-1659 [PMID: 29553872]
  22. Vaccine. 2015 Aug 14;33(34):4161-4 [PMID: 25896383]
  23. Hum Vaccin Immunother. 2020 May 3;16(5):1062-1069 [PMID: 32118519]
  24. PLoS One. 2012;7(10):e45653 [PMID: 23049682]
  25. World Dev. 2021 Jan;137:105168 [PMID: 32895596]
  26. Vaccine. 2021 Oct 28;: [PMID: 34823912]
  27. Indian J Psychol Med. 2020 Oct 13;42(6):575-576 [PMID: 33354086]
  28. Int J Nurs Stud. 2011 Oct;48(10):1281-9 [PMID: 21474136]
  29. Vaccines (Basel). 2022 Apr 25;10(5): [PMID: 35632437]
  30. J Med Internet Res. 2021 Jun 29;23(6):e24435 [PMID: 34115608]
  31. Vaccines (Basel). 2021 Nov 25;9(12): [PMID: 34960143]

Word Cloud

Created with Highcharts 10.0.0vaccinesentimentCOVID-19governmentpublicuptakevaccinationnegativehealthcampaignsstudyTwittereffectpositiverolloutavailabilityincreasedvaccineshesitancyAustraliananalysisrelationshipactionslearningmethodscollectedtweetsusedsentimentsexpressedsignificantdosesPubliclockdownsantivaccinationfindingsBACKGROUND:lacktrustmajorcontributorovercomelaunchedmultipleencourageexaminesCOVID-19-relatedeventsOBJECTIVE:aimsexamineimpactedMETHODS:Usingmachine137523publiclyavailableEnglishlanguagepublishedAustraliaFebruaryOctober2021containedvaccine-relatedkeywordsMachineextracttopicsrelatingexaminedRESULTS:majorityn=9105266%rathern=2168616%neutraln=2478518%Topicsdiscussedwithintimeframeincludedroleaccessibilityefficacycorrelationnumberadministereddailyr=15P<05showinginverseprotestsassociatedmandatesCONCLUSIONS:demonstrateprevalentremainedhighAustraliansangerslowlimitedperiodralliescontrastnewsacquisitionkeyreducedcaninformcommunicationplanningActionsSpeakLouderWords:SentimentTopicAnalysisVaccinationVaccineUptakesocialmedia

Similar Articles

Cited By