Understanding Research Gaps in Research
Most accounts of research gaps describe topics that nobody has studied. The methodological framework describes something different. It characterises where the evidence falls short, and then why it falls short, in four defined categories. That shift matters. A gap stated as a missing topic cannot tell an investigator what the next study must do, while a gap stated as a reason can.
- The four reasons a body of evidence can fall short of a conclusion
- Which apparent gaps are artefacts of reporting rather than real absences
- Why a gap is not by itself a justification for a new study
- “Read the future research sections of recent papers to find the gap.” Applied to two evidence reports, a formal framework identified a mean of 14.75 gaps each, and those gaps were more detailed than the reports' own recommendations.
- “A gap in the literature justifies a new study.” The ethical guidelines locate the justification in scientific and social value. An absence that nobody needs filled is not a reason to fill it.
What this article covers
This article sets out the methodological account of research gaps rather than the informal one. The informal account treats a gap as a subject nobody has written about. The formal account treats it as a property of an evidence base, described in defined elements and attributed to one of four reasons. The second account is harder to produce and considerably more useful, because it specifies what a subsequent study would have to do.
Three further matters follow from that account and are covered here. The first is the distinction between an absence of studies and an absence of synthesis, which call for different remedies. The second is the set of apparent gaps that are artefacts of reporting rather than real absences. The third is the step beyond identification, where a gap has to be shown to matter before it justifies anything.
Choosing a design once a gap has fixed the question is covered in how to select the right study design. The elements a gap statement is written in are covered in why use the PICO framework. Presenting the synthesis that a gap analysis emerges from is covered in tables for a systematic review.
The definition in the methodological framework
Where the evidence falls short, and why
The framework most widely used for this purpose was developed for evidence synthesis centres. It has two components rather than one. A gap is characterised using defined elements, and the reason the gap exists is identified separately.1 Its authors summarise the structure in a single line. The framework “characterizes where the evidence falls short and how or why the evidence falls short”.1
Both halves are necessary. A statement of where the evidence falls short locates the gap in a population, an intervention, a comparison, an outcome or a setting. A statement of why it falls short tells an investigator what kind of study would close it. A gap described only as a missing topic supplies neither, which is why the informal account leaves authors no better placed than before.
Absence of studies and absence of synthesis
A second distinction cuts across the first and is easier to apply. Guidance on evidence gap maps defines a gap as “an absence of evidence on a given intervention and outcome intersection”.4 It then separates two kinds of absence that look identical at first glance.4
The first is an evidence gap. Cells on such a map with “few or no impact evaluations” mark places where new primary studies are needed.4 The second is a synthesis gap. Areas carrying substantial primary evidence but “few or no medium- or high-confidence systematic reviews” mark places where a review is needed instead.4 The remedies differ entirely, and so does the cost.
A synthesis gap can often be closed by a review that costs a fraction of a new study. Treating it as an evidence gap commissions primary research that the existing literature may already have answered. The question of which kind of absence is present should therefore be settled before any design work begins.
The structure of such maps
An evidence gap map is a matrix. Interventions form the rows and outcomes form the columns.4 Bubbles at the intersections show that studies exist for that pair, with size reflecting the volume of evidence and colour indicating study type.4 Systematic reviews carry a confidence rating, and the guidance directs users to place “the greatest emphasis on evidence from medium- and high-confidence systematic reviews”.4
The structure is worth borrowing even where no formal map is produced. Listing interventions against outcomes exposes an asymmetry that prose conceals. A literature can look substantial in aggregate while leaving particular pairs entirely unstudied, and those pairs are where a gap analysis should concentrate.
Figures: identifying a research gap
Figure 1 sets out four reasons why existing evidence can fall short, each pointing to a different kind of study. Figure 2 distinguishes three situations that look identical in a literature search. Authors can use them to state the gap precisely when framing the research question.
Each reason points at a different kind of next study. Naming the reason is what converts an observation into a research question.
What to look for. Row B is the one most often misread. Where existing studies are biased, adding another study of the same design does not close the gap; it enlarges a biased evidence base. The framework asks reviewers to name the reason that most prevents a conclusion from being drawn. That forces a choice between these four, rather than a general statement that more research is needed.
Schematic. The four categories and their descriptions follow the framework cited in this article; the right-hand column states their practical implication.
A search of the published literature returns the same result in all three cases. The appropriate response differs in each.
What to look for. The third row is invisible to any method that searches journals alone. A question may have been investigated several times, with the results never reaching publication. A literature search reports that absence exactly as it reports an absence of studies. Distinguishing the two requires looking outside the published record, which is the step most gap analyses omit.
Schematic. The registry figure is the one recorded in the methodological guidance cited here, and is dated there to May 2024.
Characterising a gap
The framework asks two questions about every gap, and both have defined answers. The first concerns the elements in which the gap sits. The second concerns the reason the evidence is inadequate.
The elements
Gaps are described using the same elements that structure a review question, with setting added. The framework pairs its reasons with “the characterization of the gap using PICOS elements (also including setting)”.2 Only the elements that the evidence inadequately addresses need describing.2 A gap concerning older adults does not require every other element to be restated.
Not every gap fits the elements, and the framework's own pilot found this. About two gaps per report, some 14 per cent, could not be characterised in PICOS at all.1 These concerned prevalence, incidence and the factors affecting them, and were recorded as free text instead.1 The elements are a useful default rather than a universal container.
The four reasons
Four categories cover the reasons a body of evidence falls short, and they are not interchangeable. The first is insufficient or imprecise information, arising from few studies or small samples.2 Imprecision has a working definition here: a confidence interval “wide enough to include both superiority and inferiority”.2 The remedy is a larger or better powered study.
The second is biased information, arising from significant methodological limitations or suboptimal designs.2 This category is the one most often mishandled. Adding a further study of the same flawed design enlarges the biased evidence base rather than closing the gap. The remedy is a better design rather than more of the same.
The third is inconsistency, or unknown consistency. Effect estimates may point in different directions, or a single study may leave consistency untested altogether.2 Replication addresses the second case. Investigating the source of disagreement addresses the first, and that is a different study from a simple repetition.
The fourth is that the available information is not the right information. Results may not apply to the population or setting of interest, the outcomes measured may not be the ones that matter, or follow-up may be too short.2 The remedy is a study run where the answer is actually needed, measuring what decision makers use.
The framework asks reviewers to identify the reason or reasons that “most preclude conclusions from being made”.2 That instruction forces a judgement. A literature can be simultaneously small, biased and inconsistent, and a gap statement listing all three specifies no particular study. Naming the binding constraint is what makes the analysis actionable.
Where in the process this happens
The framework was designed for use at a specific point. Its authors envisage investigators completing the worksheet “soon after the data synthesis phase, while in the process of writing the results section”.2 They recommend that “the same investigative team that synthesizes the evidence apply the framework while writing the results”.2
Applying it afterwards proved harder. The pilot test did exactly that, working retrospectively from completed reports, and the authors describe the exercise as challenging.2 The implication for an author starting from outside is worth stating. Gap analysis is a by-product of synthesis rather than a substitute for it, and an analysis conducted without the synthesis is working from a much weaker position.
One shortcut exists for reviews that graded their evidence. Gaps can be located wherever strength of evidence was rated moderate, low or insufficient, since those ratings record the same inadequacies the framework classifies.2 The four reasons map onto the domains used in the major certainty systems.2
Weak signals and strong ones
Several methods for finding gaps circulate together as though they were equivalent. They are not. The evidence on their relative yield is limited but it points consistently in one direction.
Future research sections
Reading the closing recommendations of recent papers is the most commonly suggested method and among the weakest. The framework's pilot provides a direct comparison. Two team members applied the framework independently to two evidence reports, then compared their results against each report's own future-research section.1
The framework found far more, and found it in usable form. It identified a mean of 14.75 research gaps per evidence report, at an average of three and a half hours each.1 Those gaps were more detailed than the general recommendations the reports themselves offered.1 A future-research paragraph is a summary written at the end of a long project. It reflects what its authors found interesting rather than what the evidence base lacks.
Narrative impressions
A second method is to read widely and form an impression. The difficulty is that an impression is not auditable, and a reader has no way to distinguish a genuine absence from an incomplete search. The framework's authors found that explicit frameworks were rare in practice. Most organisations handled gaps informally, and the PICO format was the most common structure where any was used.1
The relevant contrast is not between a formal method and no method. It is between a statement a reader can check and one they cannot. A gap located in named elements, attributed to a named reason, and derived from a documented search is checkable. An assertion that little work exists on a subject is not.
Systematic assessment
The strongest position holds that systematic assessment should precede any new study at all. The recommendation in the research-priorities literature is unqualified: “investment in additional research should always be preceded by systematic assessment of existing evidence”.3
Practice falls a long way short of that. In reports of controlled trials, only one of 29 in 2009 and one of 35 in 2012 cited an updated systematic review used to design the trial.3 Among 446 research ethics protocols, only four, amounting to one per cent, used meta-analyses to plan sample sizes.3 Fewer than a quarter of previous relevant trials were cited in new trial reports.3
Those figures describe studies designed without a systematic account of what was already known. The verdict offered in the same paper is not hedged: “the present situation is ethically, scientifically, and economically indefensible”.3 The gap analysis that precedes a study is therefore not a presentational requirement. It is the step that determines whether the study should exist.
Apparent gaps that are not gaps
A search that returns nothing is ambiguous. It may mean the studies were never done. It may equally mean they were done and never reached the published record. The two are indistinguishable from inside a literature search, and only one of them is a research gap.
Results that exist but were never reported
The evidence that results go missing is substantial. Methodological guidance describes “convincing evidence for several types of non-reporting bias” and characterises the overall evidence as overwhelming.5 One cited example found that only 26 per cent of thirteen trials of a particular antidepressant had been published.5
Registry data give the clearest measure of the scale. As of May 2024, one major registry held almost half a million studies. About 13 per cent of them had summary results posted.5 The guidance also records that unfavourable results, for benefits and for adverse effects alike, are substantially under-reported in journal articles relative to regulatory documents.5
The consequence for gap analysis is direct. A search confined to journals will overstate how much is genuinely unknown, and will do so in a patterned way. The questions most likely to appear unstudied are those where the completed studies found nothing. An apparent gap is therefore at its least trustworthy precisely where it looks most promising.
Where to look instead
Two routes extend the search beyond the published record. The first is to search trials registers, manufacturer data, clinical study reports and regulatory approval packages.5 The second is to restrict attention to studies known to have been initiated, such as prospectively registered trials. Every study can then be accounted for, rather than only those that surfaced.5
Registration serves this purpose because it is required in advance. Research must be registered in a publicly accessible database before the first participant is recruited.58 The register therefore records studies that were started, including those whose results never appeared. Checking it converts an unanswerable question about absence into an answerable one.
Work already under way
A third category of apparent gap is simply premature. A question may be unanswered today and already the subject of an ongoing study, in which case a new protocol duplicates effort rather than closing anything. Registers address this case as well, since they list studies in progress alongside completed ones.
The same check applies to evidence synthesis. A question that looks unsynthesised may have a review in progress, registered and not yet published. Searching the registers for both primary studies and reviews is a short task, and it is the difference between identifying a gap and identifying a delay.
From gap to justification
Identifying a gap is the first half of the work. The second half establishes that the gap is worth closing, and the two are regularly conflated. An absence that nobody needs filled remains an absence, and filling it produces a study that nothing required.
Where the justification is located
The ethical guidelines place the justification for research in two properties held together. The first guideline states that “The ethical justification for undertaking health-related research involving humans is its scientific and social value”.6 Social value depends on the importance of the information a study would produce, such as its relevance to a significant health problem.6
Scientific value is necessary but insufficient. The guidance states that “scientific value is a necessary but not a sufficient condition for the social value of health research”.6 A methodologically sound study of a question nobody needs answered satisfies one condition and fails the other. The gap statement has to carry both.
Components of a justification
Three components make a gap statement defensible, and each corresponds to something already discussed. The first names the elements in which the gap sits, which locates it. The second names the reason the evidence falls short, which specifies the study required. The third establishes who needs the answer and what turns on it, which supplies the social value.
The third component is the one most often missing from a thesis introduction or a grant application. A sentence reporting that little research exists on a subject does not establish that anyone is waiting for the result. Research priorities should reflect disease burden, the needs of potential users such as patients and clinicians, and what is already known or already being researched.3
Scope, once the gap is fixed
A gap determines a question, and the question still has to be scoped. The methodological guidance treats breadth as a deliberate choice rather than a default. A broad review addresses a wide question, while a narrow one addresses a specific question.7 Equal precision across elements is neither expected nor required.7
That applies directly to a gap statement. A gap characterised in one element can be stated precisely in that element while remaining open in the others. A gap concerning outcomes in a particular setting does not require the intervention to be narrowed as well. Narrowing every element at once produces a question too small to matter.
Three things should be recoverable from the statement alone. The first is where the evidence falls short, named in elements. The second is why it falls short, among the four reasons. The third is who needs the answer, and what decision turns on it. A statement leaving any of the three unrecoverable has identified something, but not yet a research gap in the sense the methodological literature uses.
Common mistakes
| Fault | Stage | What it costs | The remedy |
|---|---|---|---|
| The gap is stated as a missing topic | Identification | No study is specified, since no reason is named | State the elements and the reason the evidence falls short |
| A synthesis gap treated as an evidence gap | Identification | Primary research commissioned where a review would answer it | Check whether studies exist but have never been pooled |
| Only the published literature searched | Identification | Unreported results appear as absent evidence | Search registers, regulatory documents and clinical study reports |
| Future research sections used as the method | Identification | A formal framework found far more gaps, in usable detail | Apply a framework to the evidence rather than to its summaries |
| Several reasons listed at once | Characterisation | No particular study follows from the statement | Name the reason that most prevents a conclusion |
| Bias answered with another study of the same design | Response | The biased evidence base grows rather than improves | Change the design, since volume does not correct bias |
| Ongoing work never checked | Identification | A registered study already addresses the question | Search registers for studies and reviews in progress |
| Every element narrowed at once | Scoping | A question too specific to interest anyone | Tighten only the elements the gap actually concerns |
| The gap identified without a synthesis behind it | Process | An unauditable claim about what the literature lacks | Derive the gap from a documented search and synthesis |
| Nobody named who needs the answer | Justification | Scientific value without social value, which the guidelines require | State who is waiting for the result and what turns on it |
The third row is the one with no visible symptom. A literature search that misses unreported results returns exactly what a search of a genuinely unstudied question returns. Nothing in the output distinguishes them, and the error is therefore invisible to peer review as well as to the author. The tenth row is the one most often caught, usually by a reviewer asking why the question matters.
A research gap is a property of an evidence base rather than a subject nobody has written about. The framework developed for evidence synthesis centres describes it in two parts: the elements in which the evidence falls short, and the reason it falls short. Four reasons are recognised, and each points at a different kind of study. Three further distinctions do most of the remaining work. An absence of studies is not an absence of synthesis. An absence in the published record is not an absence of results. An unanswered question is not necessarily one anybody needs answered. The ethical guidelines place the justification in scientific and social value together, and a gap statement that establishes only the first has done half the job.
How to apply this to your work
- Derive the gap from a documented synthesis, rather than from an impression of the literature.
- Establish whether studies exist but have never been pooled, since that calls for a review instead.
- Search registers and regulatory sources before concluding that studies were never done.
- Check for work in progress, both primary studies and registered reviews.
- Characterise the gap in elements, describing only those the evidence inadequately addresses.
- Name the single reason that most prevents a conclusion being drawn.
- Match the response to the reason, and change the design where the existing evidence is biased.
- State who needs the answer and what decision turns on it.
Common questions
What distinguishes a research gap from a topic nobody has studied?
A gap is a property of an evidence base, described in two parts. The framework “characterizes where the evidence falls short and how or why the evidence falls short”.1 It pairs a description in PICOS elements, including setting, with one of four reasons the evidence is inadequate.2 A missing topic names neither, and therefore specifies no study.
What are the four reasons a gap exists?
Insufficient or imprecise information, such as few studies or small samples; biased information, from methodological limitations or suboptimal designs; inconsistency, or unknown consistency where a single study leaves it untested; and not the right information, where results do not apply to the population, setting or outcomes of interest.2 Reviewers are asked to name the reason that “most preclude conclusions from being made”.2
Is reading future research sections an adequate method?
It is considerably weaker than applying a framework. In a pilot comparison, the framework identified a mean of 14.75 gaps per evidence report at about three and a half hours each.1 Those gaps were more detailed than the general future-research recommendations the same reports contained.1
What is the difference between an evidence gap and a synthesis gap?
An evidence gap is a place with “few or no impact evaluations”, where new primary studies are needed.4 A synthesis gap carries substantial primary evidence but “few or no medium- or high-confidence systematic reviews”, where a review is needed instead.4 The two look alike on a map and call for entirely different responses.
Can a literature search alone establish that something is unstudied?
No, because unreported results are indistinguishable from absent ones. Guidance on missing results describes “convincing evidence for several types of non-reporting bias”.5 As of May 2024, one major registry held almost half a million studies, and only about 13 per cent had summary results posted.5 Registers and regulatory documents have to be searched as well.
When should gap analysis be carried out?
During synthesis rather than after it. The framework's authors envisage the worksheet being completed “soon after the data synthesis phase, while in the process of writing the results section”, by “the same investigative team that synthesizes the evidence”.2 Applying it retrospectively, as the pilot did, proved challenging.2
Does a gap by itself justify a new study?
No. The ethical guidelines locate the justification in “its scientific and social value” together, and state that “scientific value is a necessary but not a sufficient condition for the social value of health research”.6 A gap statement must therefore establish who needs the answer, not only that the answer is missing.
How often is new research preceded by a systematic assessment of what is known?
Rarely, against a recommendation that it always should be. The research-priorities literature holds that “investment in additional research should always be preceded by systematic assessment of existing evidence”.3 In practice, one of 29 controlled trial reports in 2009 and one of 35 in 2012 cited an updated systematic review used to design the trial.3 Among 446 ethics protocols, four used meta-analyses to plan sample sizes.3
Declared interest: Pubrica provides research and medical writing support commercially. Nothing on this page should be read as suggesting that a gap statement can be written persuasively enough to compensate for a question nobody is waiting on. Where the search establishes that the evidence already answers the question, the useful finding is that the study is unnecessary, and saying so is part of the work.
References and sources
- Robinson KA, Saldanha IJ, McKoy NA. Development of a framework to identify research gaps from systematic reviews. Journal of Clinical Epidemiology. 2011;64(12):1325–1330. doi:10.1016/j.jclinepi.2011.06.009 read 21 October 2026, in full, from a copy hosted by the Alliance for Health Policy and Systems Research at the World Health Organization.View the full text
- Robinson KA, Saldanha IJ, Mckoy NA. Frameworks for Determining Research Gaps During Systematic Reviews. Methods Future Research Needs Report No. 2. AHRQ Publication No. 11-EHC043-EF. Rockville, MD: Agency for Healthcare Research and Quality; June 2011. Prepared by the Johns Hopkins University Evidence-based Practice Center. read 21 October 2026; the report runs to roughly 119,000 characters and the sections cited here are the framework, the reasons for a gap, and the recommendations on timing.View the report
- Chalmers I, Bracken MB, Djulbegovic B, Garattini S, Grant J, Gülmezoglu AM, Howells DW, Ioannidis JPA, Oliver S. How to increase value and reduce waste when research priorities are set. The Lancet. Published online 8 January 2014. doi:10.1016/S0140-6736(13)62229-1 read 21 October 2026, in full, from a copy hosted by Testing Treatments; the version consulted does not state volume or page numbers.View the full text
- International Initiative for Impact Evaluation. How to use evidence gap maps. read 21 October 2026; the document consulted carries no publication date, and the copy is dated February 2024 in its file name.View the guidance
- Chapter 13: assessing risk of bias due to missing results in a synthesis. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editors. Cochrane Handbook for Systematic Reviews of Interventions, current version. Cochrane. read 21 October 2026; the registry figure cited here is dated May 2024 in the chapter itself.View the chapter
- Council for International Organizations of Medical Sciences, in collaboration with the World Health Organization. International ethical guidelines for health-related research involving humans. Fourth edition. Geneva: CIOMS; 2016. read 21 October 2026; the guideline cited here is 1 and its commentary.View the full text
- Chapter 2: determining the scope of the review and the questions it will address. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editors. Cochrane Handbook for Systematic Reviews of Interventions, current version. Cochrane. read 21 October 2026, in full.View the chapter
- World Medical Association. WMA Declaration of Helsinki: ethical principles for medical research involving human participants. Adopted June 1964; current revision adopted by the 75th WMA General Assembly, Helsinki, October 2024. read 21 October 2026, in full.View the Declaration
Two further sources were read and are deliberately not cited. A 2026 scoping review of gap-identification methodology carries a formal peer review of “not approved”, with the reviewer citing conflicting inclusion criteria and unclear derivation of its categories. A large study of how trial reports cite prior research could not be opened beyond a bibliographic record, so the figures attributed to it elsewhere are not reproduced here.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
Work through these before the gap statement is written.
BEFORE CONCLUDING THAT A GAP EXISTS [ ] The search was documented and is reproducible [ ] Trial and review registers searched, not only journals [ ] Regulatory documents or clinical study reports consulted where they exist [ ] Studies in progress checked, both primary studies and registered reviews [ ] Considered whether studies exist but have never been pooled [ ] Considered whether results exist but were never published CHARACTERISING THE GAP [ ] Population described, where the evidence falls short on it [ ] Intervention or exposure described, where relevant [ ] Comparison described, where relevant [ ] Outcome described, where relevant [ ] Setting described, where relevant [ ] Elements the evidence addresses adequately are NOT restated [ ] Gaps that do not fit these elements recorded as free text instead THE REASON THE EVIDENCE FALLS SHORT [ ] One reason named as the binding constraint, not a list of several [ ] A. insufficient or imprecise: few studies, small samples, wide intervals [ ] B. biased: methodological limitations or suboptimal designs [ ] C. inconsistent, or consistency untested because only one study exists [ ] D. not the right information: wrong population, setting, outcome or follow-up [ ] The reason stated is the one that most prevents a conclusion MATCHING THE RESPONSE TO THE REASON [ ] For A, the proposed study is larger or better powered [ ] For B, the proposed study changes the design rather than adding volume [ ] For C, the proposal replicates or investigates the source of disagreement [ ] For D, the proposal runs where the answer is actually needed [ ] No proposal answers bias by adding another study of the same design JUSTIFICATION [ ] Who needs the answer is named [ ] The decision that turns on the answer is stated [ ] Relevance to a significant problem established, not assumed [ ] Scope set deliberately, tightening only the elements the gap concerns [ ] The possibility that no study is needed has been considered and recorded
We help you establish what the evidence already settles before a protocol is written. We characterise the gap in the terms a reviewer will test, and say plainly when the study is not needed.