Jan 1 / Atul Sinha

How do you decide whom to trust when experts disagree?

You are considering a decision that could change the next few years of your life.

One experienced person tells you to act now. Another, equally confident, tells you to wait. Both sound informed. Both can explain their positions. Both seem to have your interests at heart.

Perhaps the question is whether to buy a home, how to weigh a health recommendation, or what to believe about a new technology. You search for a third opinion and discover five more.

At this point, a tempting question is: Which expert is right?

But there is a more useful question to ask first:

What, exactly, are they disagreeing about?

Sometimes experts disagree about facts. Sometimes they accept the same facts but interpret the evidence differently. Sometimes their advice diverges because they are thinking about different people, outcomes, risks or priorities. And sometimes a confident-sounding claim reaches much further than the underlying source allows.

Learning to tell those situations apart does not require becoming an expert in every subject. It requires a more disciplined way of listening.

First, separate the person from the claim

A credential is a reason to pay attention. It is not a replacement for the reasoning you need to examine.

A physician may know a great deal about one medical specialty and much less about another. A financial professional may understand borrowing costs but know little about the housing supply in a particular neighborhood. A highly respected researcher may be speaking beyond the question their research studied.

The U.S. National Institutes of Health (NIH) recommends checking whether someone's expertise is actually relevant to the subject; whether the claim reflects a body of research or one study; whether the source fairly represents uncertainty; and whether professional or financial interests might shape the presentation. [1]

So instead of asking only, “Is this an expert?”, ask:

  1. Expert in what, precisely?
  2. What is the exact claim—not just the headline or recommendation?
  3. What evidence would I be able to inspect if I wanted to check it?

This matters because a conclusion may travel through several hands before it reaches you: an original study, an expert's interpretation, a journalist's summary, a social-media post, and finally your own inference.

Each step may add something useful. Each step may also change the claim.

A useful rule is to keep what the source shows, what someone infers from it, and what you decide to do visible as separate things. This is an editorial application of the source-discipline principle taught in Whiz.Life's From Crisis to Clear Seeing: whose claim is this, and what supports it?

First, separate the person from the claim

Not every expert disagreement is the same

Before choosing sides, identify the kind of disagreement in front of you.

What differs? What it might sound like The question to ask
The facts “That happened.” / “It didn't.” What record, measurement or observation could settle this?
The evidence or interpretation “This study shows a real effect.” / “That conclusion is too strong.” Do the methods and total body of evidence support the claimed conclusion?
The forecast “Prices will rise.” / “Prices could fall.” What assumptions drive each prediction, and what happens if they fail?
The recommendation “Do it.” / “Wait.” Are they weighing the same costs, risks, goals and circumstances?

These are different intellectual problems. An additional opinion may help with some of them, but another voice alone cannot settle them all.

The National Academies of Sciences, Engineering, and Medicine emphasizes that decisions draw on evidence and factors such as cost, convenience and values. It also cautions against relying on a single study, particularly for consequential choices. [2]

The GRADE approach, used in developing clinical recommendations, makes a similarly important distinction: certainty about evidence is not identical to the strength or direction of a recommendation. A recommendation also reflects benefits, harms, values, feasibility and other considerations. That is a domain-specific example of a distinction worth remembering more generally; it is not a universal formula for every life decision. [3]

A worked example: should Maya buy a home or keep renting?

The people and conversations below are illustrative, not quotations from real advisers. The underlying housing considerations come from the U.S. Consumer Financial Protection Bureau (CFPB).

Maya is 28. She has been saving steadily, likes the idea of having her own place, and has started looking at apartments and houses.

She asks two people for advice.

Adviser A says: “Buying is a way to build equity. If you can afford it, you should get started.”
Adviser B says: “Keep renting. You might move for a new opportunity, and you don't want to tie up your savings.”

Who should Maya trust?

Try inspecting their claims before answering.

1. Do they actually disagree on the facts?

Adviser A is pointing to a real possibility: homeownership can help someone build equity over time.

Adviser B is pointing to real risks: buying and selling involve transaction costs; owners bear maintenance expenses and the risk of falling property values; renting usually offers more flexibility.

The CFPB explicitly discusses both sides. It also says that the length of time someone expects to stay, income stability, repair costs and assumptions in rent-versus-buy calculators can materially affect the result. [4]

The advisers may not be disputing the same fact. They may be emphasizing different parts of a larger picture.

2. Where does interpretation enter?

“Buying can build equity” is different from “Buying is your best financial decision now.”

The second statement requires additional assumptions about the price, financing, ownership costs, future home value, alternative uses of savings and Maya's time horizon.

Likewise, “Renting preserves flexibility” does not prove that renting will always be cheaper or better.

When an expert moves from a fact to advice, ask them to make the bridge visible:

“What would need to be true for your recommendation to be right?”

3. What does Maya's situation change?

Suppose Maya is likely to relocate in eighteen months. That strengthens the case for caution about transaction costs and flexibility. If instead she expects to stay for many years and can comfortably cover ownership expenses and unexpected repairs, the balance may change.

Notice what happened: she did not find a more prestigious expert. She found a more precise question.

Her next task is not to count opinions. It is to compare realistic scenarios: likely years in the property, full monthly and upfront costs, maintenance, available cash after purchase, and plausible alternative outcomes. The CFPB recommends trying different assumptions when comparing renting and buying because forecasts can substantially change the result. [4]

4. What can Maya reasonably conclude?

Not “Expert A is right” or “Expert B is right.”

A better provisional conclusion might be:
“Both advisers identify genuine considerations, but neither recommendation is complete without my expected time in the home and a comparison of total costs. Because I may relocate soon, I will not treat the general argument for building equity as sufficient reason to buy now.”
That is not indecision. It is a conclusion sized to the information available.

And it is not individualized financial advice: an actual purchase decision still needs real figures and, when appropriate, competent independent advice.

How to test an expert's claim without reading fifty papers

You are not expected to recreate an entire field of research every time you face disagreement. In practice, four checks can reveal a great deal.

Check 1 — Is the claim specific enough to examine?

“Technology is ruining attention” is too broad to test as stated.

Which technology? Used by whom? Compared with what? Attention measured how? Over what period?

Ask the expert to narrow the claim until you could imagine what evidence would support or challenge it. This also exposes when two experts are answering different questions.

Check 2 — What is the source trail?

An attractive chart is not an original dataset. A news article is not necessarily the underlying study. A recommendation is not the same as a measurement.

Follow the claim backward where you can: Who gathered the information? What method did they use? Who interpreted it? What was omitted in the summary?

Researchers at Stanford found that professional fact-checkers were particularly effective at evaluating online sources because they used lateral reading: leaving the original site to investigate it through other sources rather than trusting its appearance or its own self-description. [5]

For an ordinary reader, a practical version is to open two independent, credible sources about the same claim—not simply two websites repeating the same press release.

Check 3 — How much weight can the evidence carry?

A vivid personal story can alert you to a possibility. It cannot by itself show how common that outcome is.

A single study can be informative. It may not overturn a larger, consistent body of research. And a study that documents a relationship does not automatically establish that one factor caused the other.

The NIH urges readers to look at study design, the population studied, the size and age of the evidence, whether findings have been replicated, and whether conclusions are being overstated. [1] The National Academies similarly encourages looking across studies rather than treating the newest one as decisive. [2]

There is an important consequence: two experts expressing opposite opinions does not mean the evidence is evenly divided.

A genuine minority criticism may uncover a flaw in the prevailing view. It deserves evaluation on its merits. But it should not receive equal evidentiary weight merely because it can be placed opposite a consensus on a screen.

Check 4 — What remains uncertain, and does it matter here?

A trustworthy expert can say, “We know this reasonably well, but this part is less certain.”

That is not a weakness. The National Academies' 2026 guidance on research trustworthiness explicitly emphasizes acknowledging limits, correcting errors and distinguishing scientific evidence from wider policy or value judgments. [6]

Ask three follow-ups:

  1. How confident should I be in this conclusion—and why?
  2. What fact or finding would most change your advice?
  3. Does your conclusion apply to people or situations like mine?

A specialist who can answer those questions may be more useful than one who insists that every case is obvious.

Three traps to avoid

Trap 1: Choosing the most confident person. Fluent explanations and prestigious titles can be impressive, but neither makes an unsupported claim true. Judge the claim's support, not only its delivery.

Trap 2: Treating every disagreement as proof that nobody knows. Scientific disagreement can reflect real uncertainty, but it can also be a productive process of testing methods and refining explanations. Conversely, a carefully reached consensus is meaningful evidence of agreement without implying infallibility. [6]

Trap 3: Outsourcing your values along with the facts. An expert can help estimate a risk or clarify a tradeoff. That person does not automatically know whether you value flexibility over stability, convenience over cost, or speed over caution. Those priorities are part of the decision and should be stated openly.

None of this means rejecting expertise. It means making better use of it.

Try it on the next disagreement you encounter

Think of two credible people who have recently recommended different courses of action. Before consulting a third person, write a brief response to the following:

  1. Their claims: What, exactly, is each person saying? Separate factual statements from recommendations.
  2. Their support: What evidence or reasoning does each rely on? Is any part merely an assertion?
  3. Their scope: Under what conditions might each be right?
  4. Your uncertainty: What one missing fact—or one better-qualified source—would most help you decide?

Then complete this sentence:
“For now, I lean toward __ because __. My confidence is __, and I would reconsider if __.”

You may discover that the disagreement is smaller than it sounded. Or that the evidence is too weak to settle the matter. Or that the experts agree on the facts and disagree only on the priorities you must choose between.

Any of those discoveries is progress.

You do not need certainty about everything before acting. But you do need to know which parts of your decision are supported, which are interpretive, which are personal, and which remain open.

The goal is not to find one person you can always trust. It is to become someone who can trust expertise intelligently.

Sources and further reading

External sources checked 19 October 2026. These links support specific claims above; they do not collectively endorse every Whiz.Life teaching example or synthesis.

  1. National Institutes of Health (NIH), “How to Evaluate Trustworthiness in Science” (last reviewed June 26, 2025). Source relevance, expertise, study design, context, conflicts and limits.
  2. National Academies of Sciences, Engineering, and Medicine, “Can science help people make decisions?” (updated April 16, 2021). Body of evidence, uncertainty, and the role of context and values in decisions.
  3. American Academy of Pediatrics, “The GRADE Approach” (accessed October 19, 2026). Domain-specific example distinguishing evidence certainty from recommendation judgments.
  4. Consumer Financial Protection Bureau, “Consider whether it's the right time for you to buy” (accessed October 19, 2026). Ownership costs, flexibility, transaction risks and the importance of changing assumptions; also see “What are some of the financial considerations of buying a home?” (last reviewed June 27, 2024).
  5. Stanford Graduate School of Education / Stanford Report, “Fact checkers outperform historians when evaluating online information” (October 24, 2017). Lateral reading to investigate source credibility.
  6. National Academies of Sciences, Engineering, and Medicine, On Being a Scientist: A Guide to the Responsible and Ethical Conduct of Research, Fourth Edition — “Earning and Maintaining Trust” (September 2026). Expertise, integrity, transparency, substantive disagreement and calibrated communication. DOI: 10.17226/29488/3.1.0.
Created with