What weighed most heavily on the decision of millions of Brazilians to refuse to go to school was not religion or the lack of access to it: it was . More specifically, loyalty to the then president.
This is what the analysis of data from more than 2,000 Brazilians interviewed in 2021, at the height of the pandemic vaccination campaign, shows.
Voters who declared they would only vote for Bolsonaro, rejecting other candidates, even those aligned with , had a vaccination rate of 65%. Among those willing to consider other options within the conservative camp, the vaccination rate rose to 71%. In other words, the greater the degree of loyalty to the political leader, the lower the vaccination adherence.
The research, published in, was the first to apply machine learning techniques (or machine learning, when the computer learns patterns through data) to investigate the political roots of vaccine hesitancy in Brazil. In this approach, the model is fed with previously classified data (in this case, people who accepted or refused the vaccine) and learns to recognize the patterns that distinguish each group to apply them to new cases.
False belief ecosystem
From this strategy, we identify an interconnected ecosystem of false beliefs. Those who believed in the effectiveness of the so-called (a set of medicines without scientific proof against the coronavirus, promoted by the federal government at the time) also tended to believe that the vaccine altered DNA, implanted microchips or caused serious diseases such as and .
This pattern suggests that people did not reject the vaccine because they independently evaluated scientific evidence. The indication is that they formed attitudes based on intuition and political loyalty and then sought arguments to justify them.
The model also outlined a sociodemographic profile of those who rejected the vaccine. Among the groups with the greatest resistance, the ones that stood out stood out: around 40% declared refusal or hesitation, the highest rate among any religious segment analyzed. It was almost double the rate recorded among (22%) and slightly higher than that of atheists (36%) and agnostics (34%). The youngest age group, 16 to 28 years old, also showed high resistance (36%). Adherence grew consistently with age, reaching the lowest level of refusal (23%) among those over 54 years of age.
The relationship with education and income followed a similar logic, but with an important caveat. Those who only had primary education registered a refusal rate of around 32%, while among university students the rate dropped to 22%. Illiterate people, however, had one of the lowest study refusal rates (21%), which can be attributed to regular visits.
In the economic dimension, refusal was more intense among those with lower incomes (around 30% for those earning up to the minimum wage), but without major variations along the scale: even among those with higher incomes, hesitation was close to 23%. This reinforces that socioeconomic factors, although present, do not alone explain the phenomenon.
The main implication of the study is for vaccination campaigns: if vaccine refusal has political roots, communication strategies need to take this issue into account. This does not imply abandoning information as an argument, but rather recognizing that, in contexts of strong polarization, information alone is not enough. It is necessary to understand the channels through which disinformation circulates and develop approaches that consider political bias as a prominent variable.
Universal methodology
The methodology was made publicly available, allowing researchers from other countries to apply similar techniques in their own contexts, as politically motivated vaccine hesitancy is not a phenomenon exclusive to Brazil.
Finally, it was not possible to point out whether conservatism alone would be a relevant factor if Bolsonaro had not adopted an explicitly anti-vaccine stance. We also cannot determine the direction of causality: did people become anti-vaccine because they supported Bolsonaro, or did they start supporting him precisely because they shared these positions?
Answering this requires further research. But one thing has become clear: in times of health crisis, politics can become as big an obstacle as the lack of information — and perhaps even greater.
This article was co-authored by Leonardo Ribeiro dos Santos, graduated in Statistics from the State University of Campinas (Unicamp)
The article was originally published on The Conversation. Read the original text.
