How Do You Know Whether You Are Searching for Evidence or Searching for Support?

Are You Investigating a Question or Building a Case?

You begin a literature search with a research question. A few papers later, a plausible answer starts to emerge. Then something subtle can happen: instead of asking, “What does the evidence show?” you begin asking, “What else supports this?”

The search still looks scholarly. You are using academic databases, reading peer-reviewed papers, following citations, and taking notes. Yet the purpose of the search may have changed. You are no longer giving competing answers a fair opportunity to survive. You are assembling a case.

That distinction matters because a literature review should do more than document that evidence exists for your position. It should help you determine how well that position survives contact with the relevant evidence, including evidence you would rather not find.

The Difference Is Whether the Search Can Change Your Conclusion

What Separates Evidence Seeking from Support Seeking?

A hypothesis gives you something to investigate, not something to protect

Researchers rarely approach a topic with completely empty minds. Previous studies, theory, professional experience, preliminary observations, and sometimes simple intuition may lead you to expect a particular result. That is not inherently problematic. Research questions and hypotheses often arise precisely because researchers have reasons to expect something.

The methodological issue is what happens next.

Confirmation bias has long been used in the psychological literature to describe tendencies to seek or interpret evidence in ways that favor existing beliefs, expectations, or hypotheses. In a literature search, this can affect not only how you interpret papers but also which papers you ever encounter.

The distinction is not simply whether you find supportive studies. A fair search may ultimately uncover overwhelming support for one conclusion. What matters is whether alternatives had a reasonable opportunity to be found and considered.

Your search terms can reveal what question you are really asking

Suppose you expect social media use to harm academic performance. Compare searches built around terms such as “social media negative effects academic performance” with searches combining broader concepts such as “social media use” and “academic achievement.”

The first query is not automatically illegitimate. It may be useful when specifically investigating adverse effects. But if it is effectively your entire search strategy for a broader question about the relationship between social media use and achievement, the wording already privileges one possible answer.

Search terminology deserves particular attention because relevant studies may describe similar phenomena using different language. Cochrane's guidance for systematic searching recommends using a variety of search terms within relevant concepts and designing searches to identify relevant evidence as comprehensively as practicable. The exact requirements depend on the review type, but the underlying lesson travels well beyond intervention reviews: do not let the wording of the expected answer define the boundaries of the evidence.

Look at what you do after finding a supportive paper

Search bias is not confined to database queries. Citation chasing can become directional too.

You find a paper that strongly supports your argument. You examine its references, search for papers citing it, look up the authors' other work, and follow several related studies. Then you encounter a paper reporting the opposite result. You record the citation and move on.

Nothing about either action seems unreasonable in isolation. Across dozens of papers, however, that asymmetry can create two very different bodies of evidence: a richly explored supportive literature and a thinly sampled contradictory one.

A useful corrective is to search deliberately for evidence that could contradict your expected conclusion, especially once an initial interpretation begins to feel convincing.

Notice whether your quality standards change with the result

Evidence seeking requires critical appraisal. But critical appraisal itself can become asymmetric.

A supportive study with a modest sample might seem “promising.” A contradictory study with a comparable sample suddenly seems “underpowered.” A favorable cross-sectional association becomes “important evidence,” while an unfavorable one cannot establish causality. Both methodological criticisms may be valid, but they should not appear only when the findings are inconvenient.

This is closely related to giving disproportionate attention to studies that agree with you. One practical test is to temporarily ignore the direction of each result and ask how you would evaluate the design if the finding were reversed.

Ask what evidence would actually make you reconsider

This is perhaps the most revealing diagnostic question: What could you find that would make you revise your current interpretation?

If you can describe such evidence, your current position remains testable in a meaningful sense. Perhaps several well-designed longitudinal studies showing no association would weaken your argument. Perhaps a competing explanation would become more plausible if an effect disappeared after controlling for a particular variable.

If the answer is effectively “nothing,” the literature search has a different function. Evidence that agrees with the conclusion confirms it, while evidence that disagrees is always explainable away. A claim protected in that manner is difficult for any search to challenge.

A search can evolve without becoming dishonest

Literature searching is often iterative. Researchers discover unfamiliar terminology, additional databases, influential authors, alternative theoretical frameworks, and new ways of conceptualizing a topic. Search strategies may reasonably change as understanding improves.

The problem is not adaptation itself. The problem arises when adaptation consistently makes favorable evidence easier to find while making contradictory evidence easier to exclude.

For systematic reviews, transparent reporting standards such as PRISMA 2020 require authors to report information sources and full search strategies, including filters and limits. For less formal reviews, you may not need the same level of documentation, but keeping a record of important changes can still help you distinguish genuine refinement from retrospective steering.

Two Researchers Can Search the Same Topic Very Differently

The difference is not that one researcher believes something and the other does not. Both began with the same expectation. The difference lies in what they permit the evidence to do to that expectation.

Common Mistakes When Searching for Evidence

Build a Search That Can Surprise You

You do not need to pretend that you have no expectations. Instead, design small checks that make it harder for those expectations to determine what evidence you encounter and retain.

Keep in mind that review types differ. A systematic review normally requires a much more explicit and reproducible search process than a traditional narrative review. You should not casually claim systematic comprehensiveness when your methods cannot support it.

Whatever the review type, the principle remains useful: your search should be designed to inform the answer, not merely decorate an answer you have already chosen.

Check What Your Literature Search Is Actually Doing

Frequently Asked Questions About Evidence Seeking

Let the Search Put Your Preferred Answer at Risk

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.