Trust May Be the Scoreboard, Not the Lever
Experiments can move reported trust without moving health intentions — while showing people invisible mechanisms reliably changes what they understand. Public health may be polishing the gauge instead of fixing the engine.
By MyAudioBooks.ai ·
Listen free: Trust May Be the Scoreboard, Not the Lever
Every public-health official on Earth has heard the diagnosis: society faces a crisis of trust. Trust in science, trust in institutions, trust in experts — declining, fragile, weaponized, pick your adjective. And from that diagnosis flows an enormous industry of remedy: campaigns to restore trust, initiatives to rebuild trust, task forces to understand why trust collapsed and how to win it back. The logic seems unarguable. People who trust science behave in healthier ways — the correlation is one of the most robust findings in behavioral research. If trust predicts protective behavior, then raising trust should raise protection. What could be simpler?
A twenty twenty-five study in the journal Communications Psychology put that logic to a direct experimental test, across four preregistered studies with a combined five thousand three hundred eleven participants. First, the researchers confirmed the familiar correlation: people who reported more trust in science also reported stronger intentions to distance, wash hands, wear masks, and vaccinate. Then they did the crucial thing that separates science from intuition: they experimentally manipulated trust, nudging participants' reported trust up or down, and measured whether health intentions moved with it. The intentions did not move. Equivalence tests — analyses designed to detect whether any remaining effect is meaningfully small rather than merely undetected — supported the null. The distinction deserves a moment, because most readers have been trained to hear 'no significant effect' as 'maybe the study was too weak to find it.' Equivalence testing flips the logic: instead of asking whether an effect can be distinguished from zero, it asks whether the effect, whatever it is, is smaller than a threshold anyone would care about. A result can be non-significant and uninformative, or non-significant and highly informative. Here the analysis suggested the informative kind: if manipulating trust does anything to these intentions, it does less than a pre-specified meaningful amount. That is a much stronger form of nothing. An internal meta-analysis of nearly four thousand participants and a machine-learning analysis pointed the same way. Trust moved; intentions shrugged.
Meanwhile, on the other side of the world, a different team was testing a different lever. A twenty twenty-six study compared how people learned about aerosol disease transmission from four formats: plain text, explainer comics, animation, and no media at all, among three thousand two hundred seventeen educators, health workers, and university students across Indonesia, Malaysia, the Philippines, and Taiwan. All three media improved knowledge of how aerosols spread; the comics and animation generally outperformed text. The effects on willingness to take protective action were weaker, culturally variable, and measured as stated willingness rather than observed behavior — the study's own authors were careful about that. But the direction was unmistakable: showing people the invisible mechanism changed what they understood, reliably, across four countries, in a way that asking them to trust did not.
Put the two results side by side and a heretical possibility comes into view. What if public health has misdiagnosed the problem? What if trust is not the lever that moves behavior but the scoreboard that records whether the levers worked — a lagging indicator of social alignment rather than a button communicators can press? The thesis of this article, stated plainly, is that short-term trust may often be the outcome of functioning communication — competence shown, mechanisms explained, incentives aligned — rather than the most actionable intervention target, and that mechanistic comprehension and people's belief in their own ability to act are more direct levers. This is a hypothesis to be tested, not a settled conclusion, and the strongest counterevidence will get a full hearing below. It does not claim trust is unimportant. It asks whether the field has the causal arrows pointing the right way.
At My Audio Books dot A I, you can create your own audiobooks from prompts, turn your documents into audio, all with one subscription, and store your items in your own personal library.
To see why the arrow matters, it helps to be precise about what correlation licenses. When two measures move together — trust and intention, in this case — statisticians say they covary, and covariance by itself is silent about which one drives the other, or whether a third factor drives both. Trust and healthy intention might correlate because trust causes intention. But they might correlate because understanding causes both: people who grasp how a virus moves through air feel more capable of protecting themselves and, independently, more trusting of the institutions that explained it. Or because aligned communities produce both: where institutions have performed well, people trust them and comply more, without the trust doing any causal work at all. Only an experiment that manipulates one variable while holding the rest constant can distinguish these stories. That is what makes the five-thousand-person null so important: it is the rare case where the field actually ran the experiment instead of assuming the arrow.
The limits of that experiment matter, and honest treatment requires stating them before the interpretation. The studies manipulated short-term, self-reported trust in scientists — state trust, not the deep institutional trust built over years. The outcomes were intentions, not observed behavior. The samples drew heavily from German pandemic-era populations. A failed short-term manipulation cannot prove that durable, hard-won institutional trust has no causal force; it can only show that the quick version — the trust a campaign can plausibly change — did not move the needle on the intentions measured. Skeptics of the null are entitled to all of these caveats, and the article will return to them.
There is also a vocabulary problem hiding inside the word trust itself, and the debate cannot be understood without unpacking it. Psychologists distinguish trait trust — a person's stable, long-running disposition to trust others — from state trust, the fluctuating, situation-specific kind a single message can nudge. Both differ again from institutional trust: the settled confidence, or lack of it, that people hold toward particular agencies, journals, and professions, built over years of watching them perform. And all three differ from trust in methods — confidence that the scientific process itself, with its error corrections and replications, tends toward truth. A campaign can plausibly move state trust in an afternoon. Trait trust barely moves at all. Institutional trust moves on the timescale of years and mostly in response to observed performance. When headlines say trust is declining, they rarely specify which ledger is falling, and when campaigns promise to restore trust, they rarely specify which ledger they intend to raise. The five-thousand-person experiment moved one ledger — state trust — and found the intentions did not follow. The other ledgers were never touched, and their causal power remains an open question rather than a confirmed one.
Now consider the other experiment's contribution, because it supplies the missing piece: a mechanism. Aerosol transmission is, from a communication standpoint, a nightmare concept. The particles are invisible, they linger, they accumulate in ways that defy ordinary intuition about colds and coughs. You cannot see why opening a window helps; you cannot watch six feet of distance fail in a crowded, unventilated room. The public-health message asks people to change behavior based on a mechanism they cannot observe. The four-country study tested whether making the mechanism visible — through animation that shows particles drifting, pooling, dispersing — changes comprehension, and it did, consistently, with comics close behind. A twenty twenty-five systematic review of eighty-eight trials found health animations frequently improved knowledge, though effects on behavior and attitudes were less uniform.
The concept at work here has a name from psychology: self-efficacy, a person's belief that they can successfully perform an action. Decades of behavioral research identify self-efficacy as one of the strongest proximal predictors of whether people act on what they know. A mechanism you understand is a mechanism you can use: once you see aerosols pooling in a sealed room, opening the window stops being obedience and becomes obviousness. The animation studies suggest comprehension and perceived capability may be the levers that trust campaigns keep reaching past. The intervention does not ask for faith. It hands over the model.
There is a darker finding that rounds out the picture, and it complicates any simple trust-is-good story. Experiments have shown that participants with high trust in science can become more receptive to false claims dressed in scientific language — white lab coats and confident jargon work precisely because trust generalizes. Reminding people to evaluate evidence critically reduces that vulnerability. Trust, in other words, is not a shield against misinformation; it can be the carrier. The finding has an uncomfortable implication for institutional strategy. A campaign that successfully raises generalized trust raises the value of the costume as much as the value of the genuine article — every pseudoscientific pitch delivered in confident technical language inherits the credibility the campaign built. The only durable defense is a public trained to ask how someone knows what they claim to know, which is precisely the skill that mechanism-focused communication exercises and trust-focused communication does not. A population trained to trust without being trained to evaluate is a population whose trust can be hijacked by anyone wearing the right costume. This is not an argument against trust. It is an argument that the capacity being trained matters more than the attitude being measured.
At My Audio Books dot A I, you can listen to this story and thousands of others that explore the hidden science and mechanics behind the headlines.
The strongest case against the scoreboard thesis deserves to be stated at full strength, because much of it rests on solid ground. Brief laboratory manipulations may simply fail to model institutional trust, which develops over years through competence, honesty, shared identity, and lived experience; showing that a five-minute nudge does nothing is like showing that a five-minute conversation does not repair a marriage — true, and beside the point. Large cross-national panel studies have associated trust in scientists with compliance, and some experimental work has reported causal support for trust, so the null literature is not the only word. The animation study used a quasi-experimental design, custom knowledge scales, relatively permissive significance thresholds for willingness, and stated intentions rather than observed behavior. And most importantly: animation may work precisely because it increases trust — the mechanism explanation and the trust it engenders may be mediation, one factor acting through another, rather than rivals competing for the same role. If comprehension builds trust and trust builds action, the scoreboard thesis has the causal order exactly backwards, and the trust campaign was right all along.
That last possibility is the serious one, and the honest response is not dismissal but design. The competing models make different predictions that future experiments can test. If trust is the lever, then interventions that raise durable trust should produce durable behavior change after comprehension is statistically controlled. If comprehension and self-efficacy are the levers, then mechanistic interventions should outperform trust-messaging even when trust is held constant.
And if the relationship is mediation — trust doing its work as one link in a chain rather than as the engine — then the strongest interventions should move both measures, in sequence, and careful analysis should reveal the order in which they move. These are three different studies, all runnable, none requiring new theory. The field's tragedy is that it has spent a decade funding trust campaigns without funding the adjudication.
The epistemic question underneath — the question about knowledge itself, what justifies believing something — is older than any of these studies. Science communication has been here before. For decades, the field operated on what became known as the deficit model: the assumption that public resistance to science reflects a deficit of information, fixable by pouring in more facts. The model failed, and its failure was thoroughly documented — people are not empty vessels, and information does not flow downhill into behavior. The trust paradigm replaced the deficit model with a different single variable: not knowledge but confidence. The scoreboard thesis suggests the field may have swapped one monomania for another. Facts alone do not move people; but faith alone does not either, and the interventions that show mechanisms may be the ones that finally escape the single-variable trap by building capability rather than filling any deficit at all. Institutions asking for trust are asking for a conclusion before showing the derivation. Institutions showing mechanisms are offering the derivation and letting the conclusion follow. The second is slower, harder to mass-produce, and impossible to sloganize; the first fits on a poster. But the pandemic years supplied a natural comparison of the two postures at global scale, and the posture that survived contact with the public was not "believe us." The communicators who weathered the crisis best were the ones who showed their work — who explained what was known, what was uncertain, and why the guidance changed when it changed. Their credibility did not come from asserting authority; it came from displaying the machinery of their reasoning in public, errors and revisions included. The audiences that watched someone think out loud, honestly, through an evolving emergency, emerged with something no trust campaign can manufacture on demand: a working model of why the guidance deserved following, and a calibrated sense of how much confidence each claim had earned. That is not trust as a gift to institutions. It is trust as a residue of comprehension — which is exactly the causal ordering the scoreboard thesis predicts.
So what would prove the scoreboard thesis wrong? Four findings would disprove it, each concrete. First, if longitudinal or field experiments that durably increase trust produce substantial changes in observed health behavior after other pathways are controlled, then trust is a lever after all and this article's framing collapses. Second, if mediation analysis shows visual or mechanistic explanations work primarily by increasing trust rather than by increasing understanding or perceived capability, then the animation evidence actually supports the trust camp. Third, if the animation effects disappear in preregistered replications using real behavior and conventional significance thresholds, the mechanism side of the contrast weakens to a knowledge-only effect with no behavioral teeth. Fourth, if generic trust-building messages outperform mechanism-focused explanations across cultures, institutions, and health decisions, then the poster wins and the derivation loses. Each of these is a study someone could run; the thesis stands only until one of them returns.
It is worth saying plainly what this article has not claimed. It has not claimed trust does not matter — the correlation between trust and protective behavior is real, replicated, and important. It has not claimed distrust is rational or that institutions should abandon credibility and accountability; an institution that lies loses the right to be believed, whatever the causal ordering. It has not claimed animation is a miracle format; the willingness effects were weaker and culturally variable, and behavior is harder to move than knowledge. And it has not claimed the trust literature is wrong — it has claimed the causal ordering within it is unsettled, which is a different and more useful thing to say.
The policy stakes are not academic. Health agencies worldwide are funding trust-restoration campaigns whose theory of change is: raise trust, and behavior follows. If the scoreboard thesis is even partly right, that money is being spent polishing the gauge instead of fixing the engine — measuring alignment while underinvesting in the comprehension and capability that produce alignment. The alternative investment is less glamorous: clear mechanistic explanations, tested across cultures, designed to leave people more capable rather than more convinced. It does not photograph as well as a trust campaign. It may simply work better, and the only way to know is to run the adjudicating experiments neither side has yet funded.
Which returns the story to the diagnosis everyone has heard. A society that cannot be bothered to explain itself will always find its trust declining, and will always mistake the decline for the disease. The scoreboard is not the game. The institutions that learn to show their work — patiently, visually, in language that assumes intelligence rather than demanding faith — may find that the trust they stopped chasing was waiting at the end of the explanation all along. And the institutions that keep campaigning for belief without showing a mechanism may discover that you cannot advertise your way to a relationship, in public health or anywhere else. The experiments that would settle which future we are in have not been run. Running them is the next study, and it has not been scheduled.
At My Audio Books dot A I, you can create fiction, non-fiction, and turn your documents into audio, all stored in one place with a single subscription — plus get instant access to thousands of audiobooks and deep-dive investigations. Learn more today at My Audio Books dot A I.