Honesty About Uncertainty Builds Trust

In a world full of overconfident charlatans, it’s easy for scientists to think that admitting uncertainty will make people trust them less. Better to sound confident. Better to smooth over the doubt.

The evidence actually points to the opposite.

A 2026 literature review found that transparency about scientific uncertainty generally builds trust rather than eroding it. Downplaying or concealing uncertainty tends to foster distrust, especially over the long term. The fear that admitting uncertainty backfires is mostly unsupported.

That doesn’t mean uncertainty communication is simple or risk-free. It means the risks are different than many people assume, and they depend heavily on how you do it. Let’s dive into some specifics:

First off, not all uncertainty language lands the same way, and the words you use to describe uncertainty can make a big difference.

A series of studies investigated how different types of uncertainty language affects trust. The clearest finding to me was that numeric ranges (”between 1,413,000 and 1,555,000”) communicate uncertainty while barely reducing trust. Vague verbal hedges (”it could be somewhat higher or lower”) register as uncertainty too, but they cost more trust. And weak words like “estimated” often don’t even register as uncertainty to most people at all.

Concrete numbers frame uncertainty as a property of measurement, something precise and scientific. Vague words read as evasiveness, something closer to hedging your bets.

Practical takeaway: if you have a range, give the range. Don’t round it off into a mushy qualifier or vague hedging.

Here’s a useful finding for anyone communicating science, but especially if they do science that can be rather fast-moving. This study examined how communicating uncertainty about vaccines affected trust. The researchers found that when evidence is later reversed, trust drops less if uncertainty was flagged up front. Simply stating “this may change” is enough to buffer the loss.

Disclosing uncertainty early makes future updates feel like science working as intended, not science failing.

This is the opposite of how a lot of institutions communicate during a crisis. The instinct is to sound sure of things right away, because certainty feels reassuring in the moment. But that certainty is like a loan against future trust, and the interest rate is high.

Uncertainty communication doesn’t hit everyone the same way, as prior beliefs do matter. A 2025 study found when evidence matches what someone already believes, adding uncertainty slightly lowers trust. However, when evidence contradicts people’s prior beliefs, communicating uncertainty actually increases trust. The researchers focused on vaccination and climate change because these issues have well-defined groups of believers and skeptics, making it easier to see how prior beliefs shaped people’s responses.

What I found most interesting was the overall pattern. Even on highly polarized issues, there was a kind of balancing effect. Communicating uncertainty caused trust to converge across the belief spectrum.

In other words, honest uncertainty doesn’t just preserve trust on average. With polarized topics, it can actually narrow the gap between people who started out skeptical and those who started out convinced.

It’s not enough to simply acknowledge uncertainty. Research shows that public health communication is more trustworthy when uncertainty is explained with compassion, its causes are made clear, it is presented as a normal part of science, and people are told when to expect updates.

But there’s a ceiling on what any single message can do. This study also found that prior trust in institutions was the strongest predictor of how a public health message landed. No single message undoes years of eroded trust.

This is an important caveat, and I’ll share a few more at the end of this post. Good communication technique matters, but it’s operating within limits set by history. A perfectly worded message from a distrusted source will still struggle.

So far these communication strategies have focused on how to build trust, but this study found how you can quickly undermine it. Uncertainty language applied to settled science can manufacture doubt and damage trust, rather than build it.

The 2025 CDC vaccine-autism webpage revision provided us with a real-world case study that could be evaluated. The new CDC under RFK Jr. changed language from “there is no link” between vaccine and autism to “studies have not ruled out the possibility.” This study recently published in Science tested how people felt about that change in communication, and found that it raised perceived vaccine risk, lowered vaccination intentions, reduced CDC trust, and increased science-denial reasoning (across the political spectrum).

Notably, the damage was institution-specific. It hit trust in the CDC, not a blanket loss of trust in scientists generally.

While disclosing genuine uncertainty can build trust, manufacturing doubt about settled science does the opposite. The damage doesn’t stay contained to the specific claim as it can bleed into trust in the institution itself.

Of course, none of this works in a vacuum, and it’s worth being honest about the limits of simply improving our communication.

Better communication is helpful, but not sufficient. No single message undoes years of eroded trust or a hostile information environment. Science communicators need real support, but right now we ask researchers and clinicians to address falsehoods with no training, no time, and little institutional backing.

The incentives also run the wrong way. Algorithms reward certainty and outrage. Careful, hedged, accurate communication is at a systemic disadvantage before it even reaches an audience.

And AI and disinformation can’t be ignored. Generative AI produces confident falsehoods at scale, and it exploits the same identity dynamics that make uncertainty hard to communicate in the first place.

The fix has to be systemic, not just individual. We need to support the messengers of science and also change the structures.

Transparency about genuine uncertainty, communicated thoughtfully and with attention to audience, can build trust.

For specific tips: Use concrete numbers/ranges when appropriate over vague hedges. Disclose early. Explain why the uncertainty exists. Normalize it as part of how science works. Consider who is delivering the message and what the audience already believes.

Feel free to disclose what you don’t know. Just don’t manufacture doubt about what you do.