Political Affiliation Drives Beliefs on Tylenol-Autism Link Over Wording
Survey experiment shows partisan divides outweigh linguistic nuances in public health perception.
1Study Design and Methodology
Researchers recruited 1,957 U.S. adults via the CloudResearch Connect platform to test how survey phrasing impacts health beliefs. Participants were randomly assigned to read claims using 'causes,' 'increases the risk,' or 'link' regarding Tylenol and autism. The study was preregistered to ensure analytical rigor and prevent bias during data collection.
2Minimal Impact of Phrasing
The experiment revealed that specific wording had a surprisingly modest effect on overall agreement rates. While participants were slightly less likely to endorse the claim when framed as 'causes,' the difference between the three phrasings was not statistically significant. Most respondents maintained their positions regardless of the specific linguistic framing used.
3Partisan Divide Remains Dominant
Political affiliation emerged as the strongest predictor of belief, far outweighing the influence of survey wording. Republicans were nearly ten times more likely to endorse the link between Tylenol and autism compared to Democrats. Independents also showed significantly higher odds of agreement than Democrats, highlighting a deep, persistent partisan polarization.
4Education and Causal Reasoning
Educational attainment influenced how participants interpreted causal language. Respondents with a four-year college degree were 34 percent less likely to endorse the claim overall. Specifically, these graduates were more sensitive to the distinction between 'causes' and 'increases the risk,' whereas participants without a degree showed almost no difference across the three phrasings.
5Context of the Controversy
The study followed a September 2025 press conference where President Donald Trump and HHS Secretary Robert F. Kennedy Jr. discussed acetaminophen use during pregnancy. This event triggered widespread public debate and a subsequent national poll by the Kaiser Family Foundation, which reported high levels of belief in a Tylenol-autism connection.
6Scientific Evidence vs Public Perception
While some observational studies have noted statistical associations, the broader scientific consensus does not support a cause-and-effect relationship. A 2024 study of 2.5 million children in Sweden accounted for genetic factors and found no increased risk, contradicting the narrative that prenatal acetaminophen exposure directly causes autism in developing children.
7Limitations of Survey Data
The researchers acknowledged that their sample was not nationally representative, as participants were recruited from an online platform. Additionally, the study examined a single, highly polarized health claim. These factors suggest that wording effects might function differently in less politically charged health topics, requiring further investigation into broader public health communication.
8Distinguishing Correlation and Causation
Scientific communication often struggles to convey the difference between correlation and causation to the general public. Roughly one-third of university press releases have been found to use causal language for correlational findings. This study suggests that current survey methods fail to capture whether respondents truly believe in direct causation.
9Future Research Directions
The authors propose developing new survey formats that explicitly separate awareness of statistical associations from beliefs in direct cause-and-effect. By refining these tools, policymakers and health communicators can better understand public sentiment. Such improvements are essential for addressing misinformation and ensuring that scientific findings are accurately interpreted by the general population.
10Implications for Health Policy
The findings underscore that changing survey wording is insufficient to alter deeply held beliefs on polarized issues. Because political identity and education level drive perception more than linguistic precision, health authorities face significant challenges in correcting misinformation. Effective communication strategies must account for these underlying social factors rather than relying on phrasing.
Original article: PsyPost — Political Affiliation Drives Beliefs on Tylenol-Autism Link Over Wording