Autonomy-Frustrating Memories Increased COVID Conspiracy Beliefs

TL;DR: A 2025 Journal of Personality study found that reactivating autonomy-frustrating memories made COVID conspiracy claims more believable, enraging, and shareable in two Quebec experiments.

Source: Journal of Personality (2025) | Leonard et al.

Conspiracy theories usually get treated as information failures. Someone saw the wrong post, trusted the wrong source, did not know the right facts.

This Journal of Personality paper argues for a more personal layer underneath that.

A false claim becomes emotionally sticky when it gives someone an explanation for why they felt controlled before — and why they feel controlled now.

COVID Mask Vignette Reactivated Autonomy-Frustrating Memories

The first experiment looked deceptively simple. Quebec adults read a short story about Alex entering a grocery store without a mask during COVID.

Depending on condition, Alex either faced no restriction, a moderate refusal of entry, or a harsher denial of access.

The mask rule was not the actual variable being manipulated. Researchers wanted to know what kind of personal memory the scenario pulled up.

After reading, participants described the first memory that came to mind and rated how much it frustrated their basic psychological needs — especially their need for autonomy.

A rule can be annoying in the moment. A rule that reactivates an old memory of being controlled lands differently. It stops feeling like a one-off policy and starts feeling like the return of a familiar loss of agency.

Restriction Plus Autonomy-Frustrating Memory Was the Combination That Mattered

1 week later, participants reported their endorsement of 7 COVID-related conspiracy beliefs. The pattern was specific.

It was not that negative memories always predicted conspiracy thinking. It was that autonomy-frustrating memories predicted stronger conspiracy belief only when the current scenario also involved restriction.

• No restriction: autonomy-frustrating memories did not meaningfully predict endorsement.
• Moderate or harsh restriction: those memories became predictive.
• Personal memory prime: the effect was strongest when participants were cued with their own autonomy-frustrating memory.

Under the no-restriction story, that link disappeared. The same kind of memory became politically and psychologically active only when the present situation echoed it.

This is more specific than a generic “angry people believe conspiracies” finding. The study identifies a fit between memory and environment. A freedom-restrictive cue makes old autonomy frustration easier to recruit as an interpretive frame.

Personality variables filled in the backdrop. Psychological reactance — the pushback people feel when they sense their freedom is being threatened — was elevated in stronger conspiracy endorsers.

So was general autonomy frustration in everyday life. The conspiracy effect was riding on real personality terrain, not on a mood-of-the-moment artifact.

A 60-Millisecond Prime Closed the Causal Loop

Study 1 could still be read as correlational. Maybe people who already liked conspiracy theories happened to describe different memories.

Maybe the memory ratings were just another mood index. Study 2 pushed harder.

Participants again described memories after a freedom-restrictive vignette. A week later, they returned for a visual task in which memory-related keywords appeared for only 60 milliseconds — fast enough to prime a memory representation, too fast to consciously read.

The clean part of the design was the control. Some participants got their own memory keywords.

Others got someone else’s memory, or autonomy-satisfying material. The yoked-control structure asks whether the effect comes specifically from one’s own autonomy-frustrating memory — not just from any emotionally charged words flashed at threshold.

Then everyone saw a fake Twitter-style post presenting a bogus COVID conspiracy claim.

The autonomy-frustrating prime moved the response. Compared with the control conditions, it raised endorsement of the false claim, increased anger toward it, and increased willingness to share it.

The subliminal memory cue made the false claim seem more believable and more spreadable — without participants knowing what had been activated.

Conspiracy Belief as a Crude Repair of Lost Agency

The theoretical frame is self-determination theory, which argues that humans have core psychological needs for autonomy, competence, and relatedness. Autonomy here means feeling that one’s actions are self-endorsed rather than imposed by something or someone external.

When autonomy gets frustrated, people seek ways to restore it. A conspiracy theory offers one.

It converts a confusing restriction into an explanation complete with villains, motives, and hidden control. The believer is no longer a passive recipient of restriction — they are someone who has seen through it.

That does not make the belief true, rational, or harmless. It helps explain why correction can be difficult.

If a belief is doing emotional work, facts alone do not erase the appeal. The appeal is not really to the claim’s accuracy. It is to the agency the claim restores.

The same mechanism explains why some public-health messages backfire. A rule communicated as pure command collides with a person’s history of being controlled. The same rule framed with choice, transparency, and respect is less likely to activate the defensive pathway.

Study Limits: Quebec Samples, Short-Term Manipulations, Self-Reported Sharing

Both studies ran in Quebec during the COVID-19 period, and the conspiracy measures were COVID-specific.

The result should not be auto-generalized to election conspiracies, climate denial, or every other domain — those carry different identity and group dynamics.

Selection is another open question. People with very high mistrust of science may have been less likely to participate at all, leaving the most hardened conspiracy communities underrepresented in the sample.

And Study 2’s outcome was willingness to disseminate, not actual behavior on a live platform.

Real-world sharing also depends on audience, identity, platform incentives, and social risk. Still, willingness-to-share is one step closer to behavior than private belief, and the prime moved both.

COVID Fact Checks Can Backfire When They Recreate Autonomy Threat

The narrower lesson is not that every conspiracy believer is secretly processing childhood memories.

It is that old experiences of being controlled shape which explanations feel satisfying when a new restriction appears — and that the activation can happen below conscious awareness.

A fact check that says “you are wrong” can accidentally recreate the same autonomy threat that made the false claim seem useful in the first place.

A more effective intervention combines evidence with language that preserves choice: here is what is known, here is why, here is what remains uncertain, here is what you can decide for yourself.

The paper suggests conspiracy belief can be less a purely intellectual error than a defensive interpretation of constraint.

That does not excuse harmful misinformation. It gives communicators a sharper target — not only the false claim, but the psychological need the false claim is pretending to meet.