Do You Need a Literature Review Before You Have a Research Question?

Which Comes First: the Literature Review or the Research Question?

Many diagrams of the research process suggest a reassuring sequence: choose a topic, review the literature, identify a gap, formulate the research question, and conduct the study.

Other advice seems to reverse the order. You are told to develop a clear research question first because you cannot search efficiently until you know exactly what you are looking for.

Both contain some truth, which is precisely why the issue becomes confusing.

You need enough direction to decide what literature is relevant. Yet the literature may reveal terminology, theories, previous findings, methodological problems, or assumptions that substantially change the question you thought you wanted to ask. In many forms of research, the relationship is therefore iterative rather than strictly sequential.

You Usually Need Some Literature Before the Final Question, Not a Finished Review

Research Questions and Literature Usually Develop in Conversation With Each Other

The apparent chicken-and-egg problem disappears once you distinguish between an initial research direction and a finalized research question.

You do not need a perfectly formulated question to begin learning from the literature. You do, however, need some starting point. That might be a broad topic, an observed problem, a theoretical puzzle, a practical concern, an initial question, or a phenomenon you want to understand.

That starting point gives your early searching direction. What you discover then helps you decide whether the question you initially imagined is actually worth asking and, if so, how it should be framed.

You Need a Starting Point Before You Search

Searching without any conceptual direction quickly becomes unmanageable.

Imagine saying, “I want to research artificial intelligence in education.” That identifies an area of interest, but not yet a researchable problem. The relevant literature could include intelligent tutoring systems, automated assessment, learning analytics, generative AI, academic integrity, teacher adoption, student attitudes, algorithmic bias, feedback, accessibility, policy, and many other strands.

You need some provisional boundaries simply to decide what to search.

Those boundaries do not have to constitute your final research question. At the beginning, they may be deliberately tentative. You might know that you are interested in university students' use of generative AI for learning, for example, without yet knowing which behavior, outcome, population, relationship, or process deserves investigation.

Early searching can help you make those decisions.

Preliminary Searching Helps You Learn What Question Is Worth Asking

A preliminary search is not merely a hunt for references that will eventually appear in your manuscript. It can function as an investigation of the research territory.

You may discover that researchers use different terminology from yours. A concept that appeared straightforward may have several competing definitions. A proposed relationship may already have been examined repeatedly. An apparent research gap may disappear once you search under a different term.

The opposite can happen as well. You may discover contradictory findings, a population that has received little attention, a methodological weakness repeated across studies, or a theoretical assumption that has rarely been tested.

These discoveries can change the question itself.

Guidance from the University of Galway describes development of a literature-review research question as iterative and recommends an initial survey of the literature to assess the amount and nature of available evidence. A preliminary search may show that a proposed question is too broad, too narrow, or otherwise needs adjustment.

This is one of the main reasons searching the literature can contribute directly to research development rather than simply documenting what others have already done.

You Do Not Need to Complete the Entire Review Before Drafting a Question

The phrase “review the literature first” can create an impossible standard for beginning researchers.

How would you know that the review was complete if you had not yet decided what question or problem the review was meant to inform?

Literature is potentially limitless. A clearer question helps determine which concepts, populations, contexts, outcomes, theories, and bodies of scholarship deserve attention. Without those boundaries, researchers can spend months reading interesting material without becoming noticeably closer to a researchable problem.

It is therefore more useful to think in terms of progressive focusing. Begin with enough direction to search intelligently. Use the literature to improve that direction. Then search again with greater precision.

The arrows in this process are not one-way. You may return to earlier stages several times. Research, regrettably for anyone fond of perfectly linear methodology diagrams, often behaves that way.

Your Early Question Can Be a Working Question

A useful distinction is between a working research question and the question that eventually governs the study.

A working question gives your inquiry enough structure to move forward. It tells you roughly what phenomenon, relationship, population, context, or problem deserves attention. But it remains open to revision as you learn more.

This allows you to avoid two extremes.

At one extreme, you search aimlessly because you refuse to formulate any question until you have “read everything.” At the other, you become so attached to your first question that the literature is reduced to finding citations that support a decision you have already made.

A working question gives the literature something to interrogate while allowing the literature to interrogate the question in return.

The Literature Helps You Test the Scope of the Question

One of the first things preliminary searching reveals is whether your proposed scope is workable.

If a search produces an enormous and conceptually diverse literature, the question may be too broad. You may need to specify a population, context, construct, relationship, outcome, theoretical perspective, or time frame.

If almost nothing relevant appears, the question may be too narrow. But sparse results do not automatically mean that you have discovered an untouched research gap. Your terminology may be inappropriate, your search may be too restrictive, or the relevant research may exist in another disciplinary literature.

This makes early searching diagnostic. It tells you something not only about the literature but also about the question you are using to interrogate that literature.

Whether you should search broadly first or begin from a more focused question depends partly on how well developed your understanding already is.

The Literature Helps You Test Whether the Question Is Meaningful

A research question can be clear and answerable yet still make a weak contribution.

Suppose you formulate a precise question and subsequently discover several recent high-quality studies that have already answered essentially the same question in comparable populations and settings. That does not automatically prohibit another study. Replication, changed conditions, methodological improvements, or unresolved inconsistencies may justify further research.

But you now need a better rationale than “this has not been studied.”

Conversely, the literature may show that the question is more important than you initially realized. Several studies might report inconsistent results. Existing evidence may rely heavily on cross-sectional self-report data when the phenomenon calls for longitudinal observation. A theory may make a prediction that has rarely been examined in your context.

In either case, literature engagement improves the intellectual justification of the question.

The Literature Can Change the Concepts Inside Your Question

Sometimes the problem is not the overall topic but the language used to formulate it.

You may begin with a familiar everyday term and discover that researchers distinguish several related constructs. “Student participation,” for example, may overlap with but not be equivalent to behavioral engagement. “AI literacy” may be operationalized differently across emerging frameworks. “Academic performance” may refer to course grades, standardized measures, task performance, retention, or another outcome depending on the study.

Those distinctions matter because a research question implicitly commits you to concepts that eventually need to be defined, operationalized, observed, or interpreted.

Preliminary literature engagement can therefore prevent you from building a study around conceptual distinctions that the relevant field has already spent years untangling.

The Literature Can Inform Feasibility and Methodological Fit

Question development is not only about originality.

Previous studies can reveal whether the phenomenon has been measured successfully, whether suitable instruments exist, whether particular populations are difficult to recruit, whether important confounders need consideration, or whether a proposed design has repeatedly encountered limitations.

This may lead you to reformulate a question even when its substantive idea remains interesting.

A question that requires data you cannot realistically obtain is not rescued by its originality. Likewise, discovering that the available methods cannot adequately capture the phenomenon may suggest that a different kind of question would be more defensible.

For a Review Study, You Need a Question but Preliminary Searching Still Comes First

When the literature review itself is the research method, the relationship becomes somewhat more structured.

A review needs a defined purpose and question because these guide eligibility criteria, searching, screening, extraction, and synthesis. Methodological guidance commonly places question formulation before the formal search used to identify the evidence set.

That does not mean you should formulate the review question in complete ignorance of the literature.

Preliminary searching can establish whether a similar review already exists, indicate how much evidence may be available, expose useful terminology, and help determine whether the proposed scope is viable. Guidance for knowledge synthesis consequently recommends preliminary searching during planning and recognizes that scanning the literature and refining the review question can be iterative.

Confusing these two stages can cause unnecessary anxiety. Refining a question after exploratory searching is not the same as secretly changing a systematic review question halfway through evidence selection because the results are inconvenient.

Methodology Can Change How and When You Engage With Prior Literature

The general advice to engage with literature early requires an important qualification: research traditions do not all treat prior literature in exactly the same way.

Grounded theory is the classic example.

Historically, some grounded theory traditions cautioned against conducting an extensive review of literature directly related to the substantive area before data collection and analysis. The concern was that strong prior conceptual frameworks could encourage researchers to impose existing categories on data rather than allowing theoretical categories to emerge through analysis.

The issue is more nuanced than the familiar instruction to “do not read the literature.” Methodological scholarship documents substantial debate over when, why, and how existing literature should be used in grounded theory. Different grounded theory traditions take different positions, and later methodological discussions have proposed ways of engaging with literature while maintaining reflexivity about how prior knowledge may shape interpretation.

The broader lesson is not that literature should normally be avoided. It is that the timing, purpose, and depth of engagement should fit the epistemological and methodological logic of the study.

You Should Usually Finalize the Question Before Major Methodological Commitments

Iteration does not mean the research question should remain indefinitely unstable.

At some point, you need sufficient conceptual clarity to decide what evidence is required to answer the question. That question then informs research design, sampling, measurement or data-generation procedures, analysis, and other methodological choices.

Changes may still become necessary, particularly in research traditions where emergent design is expected. But changing a question after substantial data collection or after seeing results raises different methodological issues from refining it during initial research development.

The literature-review stage is therefore an especially useful period for allowing the research question to change for defensible reasons before later decisions make revision more consequential.

How a Broad Interest Can Become a Researchable Question

The student did not complete an exhaustive literature review and then suddenly invent the research question. Nor did the student formulate the final question in isolation and use the literature merely to decorate its rationale with citations.

The question became more precise because searching and reading progressively improved the student's understanding of what was actually worth investigating.

Common Mistakes About What Should Come First

Begin With Direction, Then Let the Literature Improve the Question

If you are at the beginning of a research project, you do not need to choose between “question first” and “literature first” as though they were mutually exclusive procedures.

Begin with the clearest statement you can currently make about the phenomenon or problem that interests you. Turn it into a provisional question if possible. Then use preliminary searching to test that thinking.

As you search, pay attention to evidence that could force you to reconsider the question, not merely evidence that makes your original idea look promising. Ask whether your concepts mean what you thought they meant, whether the problem has already been addressed, whether contradictory evidence exists, whether your proposed population or context matters theoretically, and whether the question can realistically be answered.

You also do not need the same depth of knowledge at every stage. What constitutes enough topic knowledge to begin searching is different from what you should understand before defending a final research question.

Most importantly, preserve some record of why substantial refinements occurred. Your final proposal or manuscript does not need a diary of every abandoned idea, but you should be able to explain the intellectual rationale for the question you ultimately chose.

Before You Finalize Your Research Question

Questions About Literature Reviewing Before a Research Question

You Need Enough Literature to Develop the Question, Not Every Answer Before You Begin

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.