Although there’s evidence that the anti-science beliefs surrounding vaccines cross a broad political spectrum, I’ve always wondered if rich white liberal women were the center of the anti-vaccine universe – this is based on my own personal anecdotal evidence, so let’s just consider that a belief than a fact. A recent analysis of anti-vaccine tweets may or may not confirm my beliefs about these rich white liberals.
There has been a dramatic increase, over the past few years, in the volume of tweets that claim that life-saving vaccines are linked to autism. Anyone who reads this blog knows that that claim is demonstrably and scientifically false. Despite the science, the belief that vaccines cause autism remains. And this view is promulgated on various locations on the internet.
Like with a lot of other controversial topics, the Twitter outrage about the danger of vaccines doesn’t actually reflect a sudden surge in anti-vaccine beliefs amongst the general population. According to a recently published peer-reviewed article, most of increase in these anti-vaccine tweets represent a very specific demographic. Individuals from affluent, populated areas in five states – California, Connecticut, Massachusetts, New York, and Pennsylvania – seem to be the backbone of this sudden increase in anti-vaccine tweets.
Let’s take a look at this new paper. It could provide us with some information about the who is pushing the anti-vaccine narrative.
The anti-vaccine tweets article
The article, by TS Tomeny et al. and published in Social Science and Medicine, examined anti-vaccine tweets about autism spectrum disorder (ASD) and attempted to correlate these tweets with geographical areas. For the study, Tomeny et al. collected around 550,000 tweets from January 2009 to August 2015 using keywords related to both ASD and vaccines. The keywords included:
vaccine, vaccinated, immunization, mmr vaccine, mmrvaccine, #b1less, #hearus, heavy metals, leaky gut, mercury, ethylmercury, methylmercury, thimerosal, preservative, dpt, diphtheria-pertussistetanus, pharmaceutical companies, big pharma, autism, autistic, asperger. We also included slang and misspellings of search keywords (i.e., vacinne, vacine, antivax, anti vax, aspie, asberger, assberger, asd). Finally, we included hashtags that journalists described in their coverage of anti-vaccination beliefs on Twitter (i.e., cdcwhistleblower, cdc whistleblower, sb277).
I’ve got to believe that dozens of my tweets from that period of time were included in this initial collection. Of course, none of mine are anti-vaccine.
Then the researchers used a machine learning algorithm, which uses pattern recognition and computational learning theory in artificial intelligence, to sort the tweets into pro-vaccine, neutral, and anti-vaccine groups. Overall, the researchers found 272,546 that they considered to be anti-vaccine tweets. Of those tweets, about 108,000 from 47,000 users were geo-coded—based mostly on self-disclosed locations in public profiles—to 732 census Metropolitan and Micropolitan Statistical Areas.
The researchers used complex analyses to statistically correlate anti-vaccine tweets with various metropolitan and micropolitan areas in the USA. The authors then examined rates of anti-vaccine beliefs on Twitter over time, usually at the state level, because data per metropolitan or micropolitan areas tended to be sparse. The researchers used a Granger causality test to identify which state’s anti-vaccine tweets deviated from the national trend.
Anti-vaccine tweets – state by state
The researchers found that the magnitude of anti-vaccine tweets was mostly stable, and, interestingly, quite low between 2009 and 2014. However, by July 2014, the number of anti-vaccine tweets began to spike in five states – California, Connecticut, Massachusetts, New York, and Pennsylvania.
For example, New York’s volume of anti-vaccine tweets jumped by 5X. By January 2015, California’s volume jumped by 10X, probably as a function of the early rumblings to eliminate personal belief exemptions that eventually resulted in SB277, which did just that.
But the most interesting thing was that the bulk of these anti-vaccine tweets came from areas in those states that had common demographic variables.
- These areas were disposed to have a relatively higher population size and higher numbers of women who had recently given birth.
- These areas also tended to have higher numbers of households with annual incomes equal to or more than $200,000.
- The tweets correlated with areas that had higher numbers of men between 40 and 44, and men who were not college graduates.
- Lastly, the tweets were linked to areas with lower numbers of women aged 15 to 17.
Overall, the data suggest that the anti-vaccine movement has taken root with affluent, married white women with higher educations and older birth age. And geographically, many of the metropolitan and micropolitan areas that are the source of these anti-vaccine tweets appear to be traditionally liberal areas in fairly liberal voting states.
Despite this analysis, other research may indicate there are other centers of anti-vaccine nonsense. According to the research from the CDC on non-medical vaccine exemptions, several states saw substantial increases in personal belief exemptions, which can be a good indicator of anti-vaccine sentiment. Between 2009 and 2016, researcher has shown that Connecticut, Florida, Iowa, Kentucky, Maryland, New York, North Carolina, North Dakota, Ohio, Oklahoma, and Virginia experienced increases in these personal belief exemptions.
Of course, Tomeny et al. have stated that their study has some limitations. Because of the limitations of their search algorithm, they didn’t examine all anti-vaccine tweets. Furthermore, Twitter users only represent about 23% of all internet users, so it may not be a representative sample. An analysis of Facebook pages and groups may show a different demographic grouping, although many of us notice that the anti-vaccine message on Facebook is skewed to a few areas of the country (mainly California).
From this analysis, Tomeny et al. concluded:
Monitoring social media for up-to-date anti-vaccine beliefs allows public health professionals to address such beliefs by targeting geographic areas where these beliefs are most prevalent and tailoring the approach to demographic characteristics of populations most correlated with these beliefs.
The researchers recommended that pediatricians take note of online trends and—if the tweets are coming from inside their areas—press the issue with patients. They also wrote that:
…monitoring social media for up-to-date anti-vaccine beliefs allows public health professionals to address such beliefs by targeting geographic areas where these beliefs are most prevalent and tailoring the approach to demographic characteristics of populations most correlated with these beliefs. Future research should further understand the qualitative variation in anti-vaccine beliefs by geographic distribution and demographic correlates. Further, future research should aim to better identify and track the wide spectrum of negative beliefs about vaccines (e.g., condemnation versus hesitancy) that are expressed on social media.
I guess those of us on the pro-science/pro-vaccine fact-based side of the discussion will have to watch out for postings from California, Connecticut, Massachusetts, New York, and Pennsylvania. We’ll be watching your anti-vaccine tweets.
- Bazzano A, Zeldin A, Schuster E, Barrett C, Lehrer D. Vaccine-related beliefs and practices of parents of children with autism spectrum disorders. Am J Intellect Dev Disabil. 2012 May;117(3):233-42. doi: 10.1352/1944-7558-117.3.233. PubMed PMID: 22716265.
- Kata A. Anti-vaccine activists, Web 2.0, and the postmodern paradigm–an overview of tactics and tropes used online by the anti-vaccination movement. Vaccine. 2012 May 28;30(25):3778-89. doi: 10.1016/j.vaccine.2011.11.112. Epub 2011 Dec 13. PubMed PMID: 22172504.
- Tomeny TS, Vargo CJ, El-Toukhy S. Geographic and demographic correlates of autism-related anti-vaccine beliefs on Twitter, 2009-15. Soc Sci Med. 2017 Oct;191:168-175. doi: 10.1016/j.socscimed.2017.08.041. Epub 2017 Sep 4. PubMed PMID: 28926775.
Please help me out by sharing this article. Also, please comment below, whether it's positive or negative. Of course, if you find spelling errors, tell me!
There are two ways you can help me out. First, you can make a monthly (or even one-time) contribution through Patreon:Become a Patron!
Buy ANYTHING from Amazon.