A Popper–Lakatos Analysis of the CDC's 2025 Vaccine–Autism Statement: Epistemological and Methodological Considerations

A Popper–Lakatos Epistemological Critique of the CDC’s 2025 Wording on Vaccines and Autism Dimas Tadeu Covas University of S˜ao Paulo Abstract In November 2025, the Centers for Disease Control and Prevention (CDC) modified a section of its vaccine safety webpage by suggesting that the statement “vaccines do not cause autism” is not an evidence-based claim. This article offers an epistemological critique of such wording through the frameworks of Karl Popper’s falsificationism and Imre Lakatos’s methodology of scientific research programmes. I argue that the CDC’s formulation is logically flawed because it misunderstands the structure of scientific inference, the epistemic status of negative causal findings, and the demarcation between progressive and degenerative research programmes. The conclusion is that the claim “vaccines do not cause autism” is precisely the kind of statement that is supported by extensive empirical falsification and fits squarely within the methodological norms articulated by Popper and Lakatos. Keywords Philosophy of Science; Popper; Lakatos; Falsificationism; Research Programmes; Scientific Method; Epidemiology; Causation; Autism; Vaccine Safety; Public Health Communication; Evidence-Based Medicine. 1 Introduction Since the late 1990s, the hypothesis that childhood vaccines cause autism spectrum disorder (ASD) has been repeatedly examined in epidemiological, clinical, and mechanistic studies. Large-scale cohort analyses involving millions of children across different countries have consistently failed to demonstrate any causal association. These findings have led most scientific and regulatory bodies to conclude that vaccines do not cause autism. The CDC’s November 2025 website revision introduced significant epistemic ambiguity by stating that this conclusion is “not an evidence-based claim.” This paper scrutinizes this assertion using principles from the philosophy of science, particularly the epistemological frameworks developed by Karl Popper and Imre Lakatos. I aim to show that the CDC’s language contradicts established scientific methodology and undermines the logical basis of causal inference in epidemiology. Previous authors have invoked Lakatos’s notion of degenerative research programmes to characterize vaccine denialism and the shifting vaccine–autism hypothesis. Pagnini (2023) frames the recurring reconfiguration of anti-vaccine arguments—from thimerosal to MMR, aluminum, and “immune overload”—as a paradigmatic case of a Lakatosian degenerative programme. Similarly, Callaghan (2021) notes that the vaccine–autism claim functions as a contemporary illustration in discussions of Lakatos’s model of scientific regress. Other philosophy of science and STS analyses have employed Popperian and Lakatosian tools to understand pseudoscience and misinformation surrounding vaccines. However, to my knowledge, there has been no systematic Popper–Lakatos analysis of the CDC’s own 2025 wording change on autism and vaccines as a case of epistemically regressive institutional communication. This article aims to fill that gap by applying formal criteria from falsificationism and research programme theory directly to a public health agency’s rhetorical framing of causation and evidence. 1 2 Popper’s Falsificationism and the Logic of Negative Causal Claims 2.1 The structure of scientific testing Popper (1959) argues that scientific hypotheses must be falsifiable and that science progresses through the rejection of hypotheses that fail empirical tests. Importantly: Scientific hypotheses, according to Popper (1959), are not confirmed through the accumulation of positive instances but are tested by exposing them to potential refutation. Science does not aim to prove universal negatives, nor does it establish certainty; instead, it examines positive causal hypotheses by deriving testable predictions and subjecting them to rigorous empirical scrutiny. When these predictions repeatedly fail, the associated hypothesis is rejected. In this framework, the relevant question is not whether a claim can be disproven with absolute certainty, but whether the available evidence consistently falsifies the causal assertion under consideration. Thus, the relevant hypothesis in this case is: H1: Vaccines cause autism. If empirical studies systematically fail to corroborate H1, then Popperian methodology requires rejection of the causal claim. 2.2 Negative findings as evidence In epidemiology, a large body of negative evidence—results showing no association—counts as evidence of absence when: 1. The studies are sufficiently powered, 2. The designs minimize bias, 3. Multiple populations are examined, 4. Replication is achieved, 5. Confounding is controlled. These criteria have been met in vaccine-autism research (Taylor et al., 2014; Hviid et al., 2019; Jain et al., 2015). Thus, in Popperian terms, the hypothesis has been thoroughly tested and falsified. 2.3 Why the CDC wording misapplies Popperian logic By suggesting that the statement “vaccines do not cause autism” lacks evidence, the CDC implicitly requires proof of a universal negative—something Popper explicitly stated is impossible and unnecessary. Instead, the correct epistemic formulation is: The causal hypothesis “vaccines cause autism” has been tested and rejected. 3 Lakatos: Research Programmes and the Demarcation Between Progressive and Degenerative Science 3.1 The Lakatosian framework Lakatos (1978) conceptualizes science as a series of research programmes, each structured around a hard core of fundamental assumptions that are not readily abandoned, surrounded by a protective belt of auxiliary hypotheses that can be adjusted in response to empirical pressure. The overall trajectory of a research programme is shaped by its methodological rules and can develop in one of two directions: it may be progressive, generating novel predictions and yielding increasing explanatory power, or it may become degenerative, relying on ad hoc modifications that do not lead to new corroborated findings. 2 3.2 The vaccine–autism hypothesis as a degenerative programme Using Lakatos’s criteria, the original MMR–autism hypothesis (Wakefield et al., 1998) was never replicated, and subsequent reformulations of the claim shifted irregularly from thimerosal, to vaccine schedules, to aluminum, and then to notions of “immune overload,” each appearing only after prior versions had been falsified. None of these successive iterations produced novel predictions or independent corroboration. Taken together, the vaccine–autism line of inquiry exemplifies a degenerative research programme. 3.3 Autism research as a progressive programme In contrast, modern autism research—spanning genetics, neurodevelopmental biology, and environmental factors unrelated to vaccines—has advanced as a clearly progressive programme. It has generated robust and increasingly predictive models of ASD, identified more than one hundred genetic loci associated with neurodevelopmental pathways, and produced mechanistic explanations that cohere with contemporary neurobiological evidence. Recent large-scale genomic and developmental studies, including work identifying more than seventy high-confidence risk genes and detailed mechanistic pathways, further illustrate the expanding explanatory power and empirical refinement characteristic of contemporary autism research (Lord et al., 2020; Fu et al., 2022). 3.4 Why the CDC wording is inconsistent with Lakatos By implying that rejecting the vaccine–autism causal claim lacks evidence, the CDC fails to acknowledge the degenerative status of the vaccine–autism research programme. As noted in the preceding subsections, each successive reformulation of the hypothesis emerged only after prior versions had been empirically refuted, and none has produced novel predictions, independent corroboration, or mechanistic support. From a Lakatosian perspective, this pattern exemplifies a degenerating research programme whose protective belt expands only through ad hoc modifications that do not increase explanatory or predictive power. This makes the CDC’s wording especially problematic: it overlooks the degenerative character of the vaccine–autism hypothesis and obscures the clear methodological grounds on which the entire line of inquiry has already been decisively abandoned. 4 Argument from Ignorance and the Burden of Proof The CDC statement also commits a classical fallacy: the argument from ignorance (argumentum ad ignorantiam). This fallacy arises when a proposition is treated as plausible merely because it has not been disproven with absolute certainty. In scientific reasoning, however, uncertainty is the default, and the absence of complete disproof does not constitute positive evidence in favor of a claim. A statement such as “we cannot rule out Xwith 100% certainty” does not imply that Xis therefore plausible. Were this principle adopted, an unlimited range of unfounded causal hypotheses would remain permanently “open questions.” For instance, it is impossible to prove with absolute certainty that smartphones do not cause autism, that wearing red shirts increases the risk of measles, or that solar eclipses trigger influenza outbreaks. It is equally impossible to eliminate the logical possibility that bananas, shoelaces, or seatbelts contribute to autism. Yet none of these claims warrants scientific consideration, because none meets even the minimal evidentiary thresholds required for causal inference. These examples underscore the fundamental point: scientific reasoning assigns the burden of proof to the party making a positive causal claim. The relevant question is not whether an effect can be conclusively ruled out, but whether there exists sufficient evidence to justify belief in the causal hypothesis in the first place. In epidemiology, a causal assertion (“Xcauses Y”) must be supported by consistent associations, temporal sequencing, biological plausibility, dose–response relationships, replication across populations, and convergence of experimental, clinical, and mechanistic evidence. The vaccine–autism hypothesis satisfies none of these criteria. It has been repeatedly tested and repeatedly falsified. To treat it as “still possible” because universal negatives cannot be proven is to invert the logic of causation. Thus, the CDC’s wording does not merely soften a scientific conclusion; it misplaces the burden of proof and grants unwarranted epistemic standing to a hypothesis that has already been systematically rejected by the available evidence. 3 5 Public Scientific Communication and Epistemic Responsibility Scientific institutions carry a responsibility not only to evaluate evidence, but also to communicate it with clarity, consistency, and epistemic integrity. In fields such as vaccine safety—where public understanding directly affects population-level health outcomes—the precision of official language is itself a form of public health intervention. When authoritative bodies introduce conceptual ambiguity into wellestablished scientific conclusions, they risk generating misinterpretations that propagate far beyond the technical context in which such statements were crafted. The CDC’s 2025 wording on autism and vaccines illustrates this danger. By suggesting that the claim “vaccines do not cause autism” is “not evidence-based,” the agency inadvertently legitimizes a hypothesis that has been repeatedly falsified in high-quality research. Such phrasing not only misrepresents the evidentiary landscape but also undermines decades of work aimed at countering misinformation. In the context of public health, epistemic ambiguity functions as a vector: it lowers trust, shifts individual risk perception, and provides rhetorical support for vaccine hesitancy movements that rely on precisely this type of institutional uncertainty. This dynamic is well documented in the science-communication literature: even subtle ambiguities in institutional messaging have been shown to amplify misinformation, erode public trust, and reduce adherence to evidence-based health recommendations (Lewandowsky et al., 2017; Scheufele and Krause, 2019). From the perspective of science communication, the problem is not merely semantic. It constitutes a breakdown of what may be called an epistemic duty of care—the obligation of scientific agencies to prevent foreseeable harm arising from unclear or misleading statements. Because public health decisions operate under conditions of imperfect information and require high levels of collective trust, even a small departure from established evidentiary consensus can have disproportionate effects, including reduced vaccination rates, resurgence of preventable diseases, and amplification of pseudoscientific narratives. In this sense, the CDC’s phrasing represents not only a conceptual error but an institutional epistemic failure, one that conflicts with the core principles of scientific communication in democratic societies. Effective public health policy requires language that is not only technically correct but also aligned with established evidence hierarchies and sensitive to the dynamics of misinformation. When institutions deviate from this standard, they inadvertently participate in the erosion of the epistemic foundations upon which public health relies. 6 Conclusion Using Popper’s falsificationism and Lakatos’s research programmes, we can conclude: 1. The statement “vaccines do not cause autism” is a legitimate evidence-based statement grounded in repeated empirical falsification. 2. The vaccine–autism hypothesis is a degenerative research programme. 3. The CDC’s 2025 wording miscommunicates the epistemic status of existing evidence. 4. A more accurate scientific statement would be: “The hypothesis that vaccines cause autism has been extensively tested and consistently rejected by high-quality evidence.” Acknowledgments The author used ChatGPT (OpenAI, GPT-5.1) to assist with the drafting and linguistic refinement of portions of this manuscript. All intellectual content, argumentation, revisions, and conclusions are the sole responsibility of the author, who has verified the accuracy of all information presented. References Popper K. The Logic of Scientific Discovery. Hutchinson; 1959. Lakatos I. The Methodology of Scientific Research Programmes. Cambridge University Press; 1978. 4 Feyerabend P. Against Method. Verso; 1975. Taylor LE, Swerdfeger AL, Eslick GD. Vaccines are not associated with autism. Vaccine. 2014;32(29):3623– 3629. Hviid A, Hansen JV, Frisch M, Melbye M. Measles, mumps, rubella vaccination and autism. Ann Intern Med. 2019;170(8):513–520. Jain A, et al. Autism occurrence by MMR vaccine status among US children. JAMA. 2015;313(15):1534– 1540. Wakefield AJ, et al. Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children. Lancet. 1998;351(9103):637–641. Nyhan B, Reifler J. The roles of information deficits and identity threat in science denial. PNAS. 2015;112(52):15837–15842. Institute of Medicine. Adverse Effects of Vaccines: Evidence and Causality. National Academies Press; 2012. Lord C, Brugha TS, Charman T, et al. Autism spectrum disorder. Nat Rev Dis Primers. 2020;6(1):5. doi:10.1038/s41572-019-0138-4. Grove J, Ripke S, Als TD, et al. Identification of common genetic risk variants for autism. Nat Genet. 2019;51(3):431–444. doi:10.1038/s41588-019-0344-8. Fu JM, Satterstrom FK, Peng M, et al. Rare coding variation provides insight into the genetic architecture and phenotypic context of autism. Nat Genet. 2022;54:1283–1293. doi:10.1038/s41588-022-01146-2. Lewandowsky S, Cook J, Fay N, Gignac GE. Rational denial of scientific evidence. Psychol Sci. 2017;28(6):691–699. Scheufele DA, Krause NM. Science audiences, misinformation, and evidence-based communication. Proc Natl Acad Sci USA. 2019;116(16):7662–7669. Oreskes N, Conway EM. Merchants of Doubt. Bloomsbury Press; 2010. Centers for Disease Control and Prevention. Autism and Vaccines — Vaccine Safety. Updated November 2025. U.S. Department of Health and Human Services; 2025. Available from: https://www.cdc.gov/ vaccinesafety/autism. Pagnini C. Vaccine Hesitancy and Degenerative Research Programmes: A Lakatosian Analysis. PhilPharm Blog; 2023. Available from: https://philpharmblog.wordpress.com. Callaghan M. Lakatos Revisited: Innovation and ‘Novel Facts’. EconStor; 2021. Available from: https://www.econstor.eu. Author(s) unknown. Popper, Pseudoscience, and the Logic of Demarcation in Vaccine Debates. PhilSci Archive; 2020. Available from: https://philsci-archive.pitt.edu. 5