The Effects of Anti-Vaccine Conspiracy Theories on Vaccination Intentions

If you want to understand why measles crept back into the United Kingdom by 2008, you can't ignore a ghost from 1998. That was the year Andrew Wakefield published a Lancet paper suggesting a link between the measles, mumps, and rubella vaccine and autism. The paper was retracted and debunked, but the doubt it cast lingered. As uptake slipped below the roughly 95 percent coverage public health relies on for herd immunity, the door opened to outbreaks. That history sets up a sharper question than "Are people misinformed?" It asks: when people encounter anti-vaccine conspiracy theories, does that nudge them away from vaccinating, and if so, through what psychology? Conspiracy theories, in this context, are stories that recast vaccination as the hidden work of malevolent actors—pharmaceutical companies and governments—keeping truths under wraps. Jolley and Douglas focus on a simple but potent idea: these narratives don't just swap facts; they change how people feel and whom they trust. They propose four pathways that could carry the effect from conspiracy thinking to lower vaccination intentions. First, perceived dangers of vaccines—if you think shots are risky, you hesitate. Second, feelings of powerlessness—if you feel you can't control vaccine outcomes, why engage? Third, disillusionment with the people involved in immunization—disappointment can curdle into disengagement. And fourth, trust in authorities—lower trust means lower willingness to follow their advice. The question is whether those four actually transmit the effect. To test that, they ran two studies. The first, correlational, asked whether people who endorse anti-vaccine conspiracy beliefs also say they'd be less likely to vaccinate. The second, experimental, asked whether a brief exposure to conspiracy content can cause a drop in intentions—and whether the same four psychological channels do the carrying. This two-step—first the association, then the causal probe—allows them to move from a map of the terrain to a test drive on the road. Study 1 recruited 89 British parents—mostly women, average age about 38—and walked them through a tightly controlled scenario. Everyone read about a fictitious disease, dysomeria, described as a droplet-spread "DS virus" that causes fever and vomiting. They were told a vaccine was recommended, and then asked, on a seven-point scale, whether they'd vaccinate their eight-month-old next week. Before that decision, participants completed reliable scales tapping anti-vaccine conspiracy beliefs and those four proposed mediators. The measures performed well statistically—internal consistencies were high, with, for example, perceived dangers of vaccines at about 0.86 and powerlessness around 0.82—so the team could treat each as a distinct construct. What happened when belief met intention? Parents who endorsed anti-vaccine conspiracy beliefs were less inclined to vaccinate the fictional child. In a regression, that belief–intention link was statistically robust, explaining about 16 percent of the variance in intentions—F of 1 and 87 equals 15.97, with a p-value below 0.001, and the slope pointing down. That's not everything, but it's not trivial either. Then they looked at the proposed channels. Higher conspiracy belief aligned with stronger perceptions that vaccines are dangerous, more powerlessness, more disillusionment, and less trust. Some of those links were hefty—perceived dangers alone accounted for roughly 58 percent of the variance in its separate model—while the others, like powerlessness at about 32 percent and disillusionment near 46 percent, still held significant weight. Trust was lower too, with about 20 percent of its variance tied to conspiracy belief. Because these four mediators were correlated with one another, Jolley and Douglas checked that they really hang together as separate factors. Pooling items across both studies, they ran a factor analysis and found the data were excellent for this kind of modeling: the Kaiser–Meyer–Olkin index was 0.93, well above the usual 0.6 benchmark, and Bartlett's test indicated plenty of shared variance to work with. A four-component solution emerged cleanly, with each mediator loading on its own component. That's the structural scaffolding. The mechanistic test came next. Using a 5,000-sample bootstrap mediation model—the statistical equivalent of stress-testing the indirect paths—they asked whether each mediator carried unique explanatory power from conspiracy belief to intention when all four were in the model together. Each one did. Perceived dangers, powerlessness, disillusionment, and lower trust all significantly mediated the link. One more breath before we speed ahead: that's still correlational. It tells you the pathways exist in the data, not that a dose of conspiracy content causes movement along them. So, Study 2 tightened the screws. This time, 188 participants recruited online were randomly assigned to read one of three brief passages: a pro-conspiracy text arguing that vaccines are harmful and that authorities are hiding the truth; an anti-conspiracy refutation; or a neutral control. Everyone then faced a similar fictitious disease vaccination decision. First check: did the manipulation shift conspiracy endorsement? Yes. Across conditions, endorsement differed significantly—F of 2 and 185 equals 13.79, with a p-value below 0.001, and a partial eta squared around 0.15. It was highest after the pro-conspiracy passage, averaging just over 4.1 on a seven-point scale, lower in the control at about 3.6, and lowest after the refutation at around 2.9. So the texts moved belief in the expected directions. Did intentions move too? Here the effect was smaller but clear. Exposure condition affected vaccination intentions—F of 2 and 185 equals 4.81, with a p-value of 0.009, and a modest effect size, with partial eta squared approximately 0.05. People who read the pro-conspiracy piece were least likely to vaccinate the fictional child, averaging about 4.87 on the seven-point scale. Those who read the refutation averaged higher, around 5.69, and the control sat in between at roughly 5.47. Pairwise tests told a clean story: pro-conspiracy differed from both the refutation and the control, but the refutation didn't reliably outperform the control. Translation: a single dose of conspiracy content knocked intentions down; a single dose of counter-argument reduced conspiracy belief but didn't lift intentions above baseline. Now the mechanism check. Just as in Study 1, participants completed the four mediator scales, which again showed good reliability—alphas around 0.90 for perceived dangers, 0.88 for powerlessness, 0.93 for disillusionment, and a respectable split-half reliability for trust near 0.73. The authors reconfirmed the four-factor structure across both studies, as the KMO stayed at 0.93. Then they ran bootstrapped multiple-mediator models with 5,000 resamples to see which pathways carried the causal effect of condition to intention. The answer depended on the comparison you make, but one pattern was rock solid. Perceived vaccine dangers consistently transmitted the effect. When you compare pro-conspiracy to the refutation, perceived dangers, powerlessness, and disillusionment each carried significant indirect effects. When you compare pro-conspiracy to the control, perceived dangers and reduced trust in authorities did the mediating. You can feel the psychology shifting: the pro-conspiracy message raised the sense that vaccines are dangerous—means around 4.00 versus roughly 2.97 after the refutation and 2.39 in the control—and, depending on the contrast, it either sapped a sense of agency and deepened disillusionment or undermined trust. This is a lot of statistics. What does it amount to in human terms? Brief exposure to anti-vaccine conspiracy narratives nudged people off the fence, and it did so by changing how risky vaccines felt, how much control people thought they had, how let down they felt by institutions, and how much they trusted the authorities who recommend shots. The causal chain that was only implied in Study 1 came into focus in Study 2: condition changes mediators; mediators shift intentions. Let's linger on two caution lights. First, the effect sizes were small. That partial eta squared of 0.05 for intentions is meaningful but not a tidal wave. In the real world, people are bombarded with a mix of messages over months and years, so a single brief text is a minimal dose. Second, both studies used a fictional disease to strip away prior knowledge and ideology. That's a strength for internal validity, but it also means we're looking at intentions in a laboratory-like frame, not actual clinic visits. Jolley and Douglas are clear on this point: intentions don't always translate to behavior, and follow-up work should track real decisions. There's also a subtle asymmetry worth noting. The refutation lowered conspiracy endorsement below control, which is good news for correcting false beliefs, but it didn't lift vaccination intentions above control in this one-shot setting. That tells you correcting a belief isn't the same as cultivating a decision. In practical terms, you might need to pair counter-arguments with messages that directly reduce perceived danger and build a sense of agency and trust. Step back to the bigger picture we started with. The return of measles in the U.K. after coverage dipped below the 95 percent mark isn't just a case of "people got the facts wrong." It's a case of stories about hidden motives reshaping how vaccines feel, who seems credible, and whether your actions matter. Jolley and Douglas provide that process with names we can measure—perceived danger, powerlessness, disillusionment, and trust—and then show those measures are not window dressing. They sit on the path from belief to intention, and at least in a controlled setting, they carry causal weight. What should we do with that? The authors suggest two broad moves. One is direct refutation—don't let conspiracy claims go unanswered—and their manipulation check shows that can reduce endorsement. The other is trust-building and efficacy-building. If perceived danger is the most consistent mediator, then safety communication has to be vivid, specific, and repeated. If powerlessness and disillusionment play roles, then giving people clear, controllable actions and making the decision-making process transparent can matter too. None of this is a silver bullet; the observed effects here are modest, the samples were a British parent group and a U.S. online panel, and the disease was fictional. But they point the spotlight where it belongs. Future work, as Jolley and Douglas argue, should get closer to the clinic: larger and more diverse samples, real vaccination behavior, and careful tests of how the order and source of information shift both beliefs and decisions. That's not speculative hype; it's the straightforward next step if we want to turn a clean lab result into a public health win. In the meantime, when you hear a snappy story about shadowy forces and vaccines, remember what these studies showed. The story doesn't just change what people think. It changes how it feels to act. And that feeling is where intentions live.