How Does Research Help Us Make Decisions Under Uncertainty?

How Can We Decide When the Evidence Is Incomplete?

Many important decisions cannot wait until everything is known.

A physician may need to choose among treatments whose benefits and harms cannot be predicted perfectly for an individual patient. A school may need to decide whether to adopt an intervention before years of local evidence accumulate. A government may have to respond to an emerging problem while information is still incomplete. An organization may need to choose between competing strategies whose future outcomes remain uncertain.

Research is often expected to help. Yet research itself contains uncertainty. Samples do not perfectly represent every population. Measurements have limitations. Estimates have margins of error. Findings may depend on assumptions and context. Different studies may disagree.

So what exactly does research contribute if it cannot tell us with certainty what will happen?

Research Makes Uncertainty More Informative and More Manageable

Research Improves Decisions Without Promising Certainty

Uncertainty is normal, not evidence that research has failed

A decision is made under uncertainty whenever its relevant facts, causal relationships, future conditions, or outcomes are not completely known.

That describes a great many real decisions.

Research can reduce some uncertainty, but expecting it to eliminate uncertainty altogether sets an unrealistic standard. Even a strong study observes a particular set of people, conditions, measurements, and outcomes. Applying its findings elsewhere requires judgment about how well that evidence fits the new situation.

The purpose is therefore not necessarily to make uncertainty disappear. It is to improve what we know about the choices before us.

Research can estimate what is likely to happen

Instead of relying solely on intuition, research can provide empirical estimates of frequencies, associations, effects, risks, benefits, harms, preferences, experiences, or other relevant outcomes.

These estimates do not predict every individual case. They provide evidence about patterns observed under specified conditions.

A decision-maker can then ask a more useful question than "Will this definitely work?" The question becomes something closer to "Given the available evidence, what outcomes appear plausible, how large might they be, and how uncertain are those estimates?"

Research can compare alternatives

Many decisions are not choices between "evidence" and "no evidence." They are choices among alternatives.

Should an organization retain its current practice, modify it, or replace it? Which of several interventions appears to offer the most favorable balance of benefits and harms? Does one approach work better for particular populations or conditions?

Comparative research can provide evidence about these alternatives. Depending on the question and design, researchers may estimate differences in outcomes, identify conditions under which options perform differently, or document trade-offs that would otherwise remain hidden.

Research can tell us how uncertain an estimate is

A numerical result without information about its uncertainty can create false precision.

Statistical analyses often provide ways of expressing uncertainty around estimates, such as confidence intervals. Other approaches assess the certainty or quality of a broader body of evidence. Qualitative research can also make uncertainty visible by identifying variation in experiences, contextual differences, or competing interpretations that simple averages might obscure.

What matters is not merely obtaining a result. Decision-makers need some basis for judging how much confidence to place in it.

Research can reveal what we still do not know

One of the most useful outcomes of research is sometimes a clearer map of ignorance.

A study may show that evidence is strong for one population but weak for another. It may reveal that an intervention's average effect is reasonably understood while its long-term harms remain uncertain. A systematic review may find that apparently confident recommendations rest on surprisingly limited evidence.

This is useful information. Decisions can be designed differently when uncertainty is known rather than hidden. Decision-makers may proceed cautiously, collect additional data, limit implementation, establish monitoring, or commission further research.

Evidence-informed does not mean evidence-determined

Research evidence is often only one input into a decision.

The World Health Organization describes evidence-informed decision-making as identifying, appraising, and mobilizing the best available evidence for policy and programmes. Its guidance also recognizes that decisions can involve multiple forms of evidence and practical considerations.

For example, an intervention might appear effective in research but be prohibitively expensive in a particular setting. Another may offer benefits while raising important equity concerns. Stakeholders may consider certain outcomes more important than others. Implementation may require infrastructure that does not exist locally.

Evidence therefore informs judgment; it does not abolish judgment.

Research can make the reasoning behind a decision more transparent

Without systematic evidence, decisions may still be made, but the assumptions behind them can remain implicit. Research can make some of those assumptions testable.

A decision-maker can identify which evidence supports a choice, where that evidence is weak, which considerations lie outside the research evidence, and what assumptions must hold for the decision to make sense.

This does not guarantee the correct decision. It does make the reasoning more open to scrutiny and revision.

That is part of why research offers something different from relying only on experience, tradition, or authority. Its methods and evidence can be examined rather than accepted solely because a particular person or institution endorses the conclusion.

The best decision can still lead to a bad outcome

This is one of the most important consequences of uncertainty.

Suppose the available evidence indicates that Option A has a better expected balance of benefits and harms than Option B. Choosing A can be reasonable even though an unfavorable outcome remains possible. Conversely, choosing poorly can occasionally produce a fortunate result.

We should therefore distinguish the quality of the decision process from the outcome of a single decision. Under uncertainty, good decisions are based on the information reasonably available at the time, not judged solely by whether events happened to turn out well afterward.

Research Can Change a Decision Without Providing a Certain Answer

Research has not eliminated uncertainty. It has changed the decision from an unsupported guess into a more explicit judgment informed by evidence, local conditions, and acknowledged uncertainty.

Misunderstandings About Evidence, Decisions, and Certainty

Use Research to Improve the Decision, Not to Pretend Uncertainty Has Disappeared

Whether you are conducting research or using it, begin by identifying the uncertainty that actually matters to the decision.

Do you need to know whether an intervention has any meaningful effect? How large that effect might be? Whether it creates harms? Whether findings transfer to your population? Whether one option is better than another? The evidence required depends on the decision.

This last point is particularly important. Evidence-informed decision-making is not a one-time transfer of findings from a paper into practice. Evidence changes, contexts change, and implementation itself can produce new information.

Research is therefore part of a learning process. Decisions can generate further questions, and those questions can motivate new research. This relationship is one reason research serves purposes beyond simply producing publications.

Before Using Research to Support a Decision

Questions About Research and Decisions Under Uncertainty

Good Decisions Do Not Require Pretending We Know Everything

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.