Can Research Be Worth Doing Mainly Because It Rules Out an Explanation?

Can Finding Out What Is Wrong Be a Meaningful Research Contribution?

Researchers often imagine a valuable study as one that discovers the correct explanation for something. You identify a problem, test possible causes, and eventually explain what is happening. But research does not always progress so neatly.

Sometimes a study cannot tell us what the correct explanation is. What it can do is show that one apparently plausible explanation is inconsistent with the evidence, too weak to account for the phenomenon, or unlikely under the conditions examined.

That may sound like an incomplete result. After all, if you have ruled out one explanation but still cannot say what the answer is, what exactly have you contributed?

Yes, Eliminating a Plausible Explanation Can Advance Knowledge

Eliminating Explanations Is Part of Building Better Knowledge

Research does not advance only by accumulating confirmed explanations

Suppose researchers have several credible explanations for the same phenomenon. If a well-designed study substantially weakens one of them, researchers now face a smaller and better-defined set of possibilities.

They may not yet know which remaining explanation is correct. Nevertheless, their state of knowledge has changed.

This illustrates why the purpose of research extends beyond producing a completely new finding. Research can also test existing claims, correct mistaken assumptions, discriminate among explanations, establish boundaries, and refine what can reasonably be believed.

Elimination can therefore be informative. Knowing that an attractive explanation is inadequate may redirect theoretical development, prevent later studies from repeatedly pursuing the same account, or force researchers to consider mechanisms that previously received less attention.

Ruling something out is stronger than failing to find it

This is the most important distinction in the entire discussion.

A conventional non-significant result does not, by itself, establish the second conclusion. A study can fail to detect an effect because the effect is absent, but it can also fail because the sample is too small, measurements are noisy, implementation is weak, assumptions are violated, or the true effect is smaller than the study was capable of detecting.

This is the familiar distinction between absence of evidence and evidence of absence. Recent methodological work on replication of null findings reinforces this point: statistically non-significant results may remain inconclusive unless the study and analysis are designed to evaluate evidence for the absence of a meaningful effect.

What would count as evidence against an explanation?

There is no single test that applies to every discipline or research design. In general, however, the study must generate observations that would be difficult to reconcile with the explanation being evaluated while addressing credible competing reasons for obtaining those observations.

For quantitative studies concerned with whether an effect is absent or too small to matter, researchers may need approaches specifically suited to that question. Depending on the research problem, these can include confidence intervals interpreted relative to a meaningful effect size, equivalence testing, Bayesian approaches such as Bayes factors, or other model-comparison procedures.

For qualitative, historical, theoretical, or case-based research, ruling out an explanation may instead depend on evidence such as incompatible observations, failed predictions, process evidence, contradictory records, comparison among cases, or the inability of an explanation to account for critical features of the phenomenon.

The methodological details differ, but the underlying principle is similar: the evidence must actually bear on the explanation you claim to have weakened.

The explanation needs to be plausible enough to be worth testing

Not every rejected idea constitutes an important contribution. Researchers could invent countless implausible explanations and design studies showing that they are wrong. Eliminating them would add little.

The value of ruling out an explanation depends partly on its prior plausibility and importance. Was it supported by theory? Had previous studies suggested it? Was it commonly assumed in professional practice? Would researchers reasonably have pursued it without stronger contrary evidence?

A study that challenges a serious contender contributes more than one that rejects an explanation nobody had good reason to believe.

Ruling out an explanation can redirect subsequent research

Consider a field in which three mechanisms might explain an observed pattern. Researchers may repeatedly design studies around all three because none has been adequately tested.

If credible evidence makes one mechanism substantially less plausible, subsequent research can concentrate on the remaining possibilities or develop alternatives. This is one way research can be valuable because it makes better future studies possible.

The benefit is not simply that researchers now possess one more published finding. Their search space has changed.

Negative evidence can also correct a distorted literature

Scientific literatures do not necessarily contain a representative sample of all studies that have been conducted. Research reporting positive or statistically significant findings has historically been more likely to appear in the published record in many contexts, contributing to publication bias.

This matters because a literature containing mostly supportive findings can make an explanation appear more secure than the complete evidence warrants. Reporting rigorous negative or null findings can help produce a more accurate evidential record and may improve later evidence synthesis.

The value of such findings depends on their quality. A poorly designed negative study does not become informative merely because the literature needs more negative results. Methodological rigor remains the relevant standard.

Some explanations can only be narrowed, not completely eliminated

Researchers should also be cautious with the phrase rule out. Scientific explanations often contain assumptions, boundary conditions, auxiliary hypotheses, and predictions that can be revised when contrary evidence appears.

A study may therefore justify a more precise conclusion: perhaps an explanation does not account for the phenomenon under these conditions, cannot produce an effect of a practically meaningful magnitude, or is inconsistent with a particular prediction.

That is not rhetorical weakness. It is often a more accurate description of what the evidence establishes.

When Eliminating One Explanation Changes the Research Question

Common Mistakes When Research Appears to Reject an Explanation

Ask Whether Your Study Can Actually Discriminate Among Explanations

If your research is intended to challenge or eliminate an explanation, design it around that objective from the beginning. Do not conduct a conventional study designed to detect an effect and then reinterpret an inconclusive result as proof that the explanation is wrong.

This also provides a useful way to think about research that produces no immediate change in practice, policy, or behavior. Correcting an explanation may initially affect only what researchers believe. That can still matter if subsequent research would otherwise have proceeded from a mistaken premise.

Most importantly, match your language to the strength of the evidence. “We found no statistically significant association,” “the data were inconclusive,” “the results are inconsistent with the predicted pattern,” and “effects larger than the prespecified threshold are unlikely” are not interchangeable claims.

Before Claiming That Research Has Ruled Out an Explanation

Questions About Research That Rules Out Explanations

Knowing Which Explanation Does Not Work Can Move Research Forward

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.