03 · What You Need to Know
First Determine Whether the Problem Is Missing Evidence or Fragmented Evidence
Primary Research and Evidence Synthesis Solve Different Problems
Primary research generates new observations or data. Depending on the question, this might involve experiments, surveys, interviews, observations, measurements, administrative records, fieldwork, or other forms of original data collection.
Evidence synthesis works with existing studies. Systematic reviews use explicit methods to identify and appraise research relevant to a defined question and synthesize what the studies collectively show. Depending on the question and evidence, synthesis may be quantitative, qualitative, narrative, mixed, or use another established review methodology.
Primary study
Generates new empirical observations to answer a research question.
Evidence synthesis
Systematically identifies and integrates existing research to determine what the body of evidence shows and where uncertainty remains.
Neither is inherently more important. The appropriate choice depends on what information is currently missing.
Finding Many Studies Should Make You Consider Synthesis Before More Data
If your search reveals a substantial literature directly relevant to the question, another primary study should not be the automatic next step.
The studies may already contain enough information to answer the question collectively even when no individual paper does. Differences among their results may reflect sampling variation, methodological quality, populations, interventions, measurements, settings, or other sources of heterogeneity that become clearer when the evidence is examined systematically.
In clinical research, recent REVEAL guidance explicitly recommends beginning the planning of a new trial by identifying existing systematic reviews. When necessary, researchers then search for published, unpublished, and ongoing trials to determine whether a genuine evidence gap remains and to inform the design of any new trial.
The exact workflow differs among disciplines, but the underlying principle travels well: before generating more evidence, understand the evidence you already have.
A Literature Review in a Thesis Is Not Automatically an Evidence Synthesis
Nearly every research project includes some form of literature review. That does not mean the project has systematically synthesized the evidence.
A conventional narrative literature review may provide conceptual background, describe influential studies, identify theories, and position a research problem. A systematic review answers a defined question using explicit, reproducible methods for searching, selecting, appraising, and synthesizing relevant evidence.
Other synthesis approaches serve different purposes. A scoping review may map a broad field, characterize concepts or research activity, and identify areas requiring further investigation. Qualitative evidence synthesis can integrate findings about experiences, perceptions, or processes. Meta-analysis is a statistical synthesis technique that may be used within a systematic review when quantitative combination is appropriate.
Choose the synthesis method according to the question rather than treating every review as interchangeable.
Use a New Primary Study When the Required Evidence Does Not Yet Exist
Sometimes the answer really does require new data.
Perhaps an important outcome has never been measured. Existing studies may not include the population needed for a consequential decision. The available evidence may rely on designs incapable of addressing an important causal question. A new technology or policy may have created conditions not represented in earlier research.
In such cases, synthesis can document the gap but cannot manufacture evidence that does not exist.
The primary study is justified when it is capable of providing information that the existing evidence base genuinely lacks.
Use Synthesis When Individual Studies Cannot Tell You What the Evidence Shows Collectively
Suppose 20 studies estimate the same relationship, but their effect estimates vary.
Reading them one at a time may leave you with an impression that the literature is “mixed.” A systematic review can examine study quality, compare methods, investigate sources of heterogeneity, and, when appropriate, statistically combine estimates.
The resulting synthesis may reveal that most apparent disagreement is compatible with sampling variation. Alternatively, it may show meaningful heterogeneity requiring explanation.
Either result provides information that another isolated study may not.
Conflicting Studies Do Not Automatically Mean More Primary Research Is Needed
When studies disagree, researchers often conclude that “further research is necessary.” That may be correct, but it skips an important step.
First ask why the evidence conflicts.
Do higher-quality studies differ from weaker ones? Are outcomes defined differently? Do effects vary systematically by population, setting, intervention intensity, follow-up period, or another characteristic? Are some estimates simply imprecise?
A synthesis can help distinguish these possibilities. Only then can researchers determine what kind of additional primary evidence would actually resolve the uncertainty.
Synthesis Can Reveal That the Evidence Is Already Good Enough
A review does not need to end by identifying another research gap.
Sometimes the responsible conclusion is that the existing evidence already supports a sufficiently credible and precise answer for the question at hand.
This matters because determining whether existing evidence is already sufficient is part of research prioritization. If synthesis shows that the question is adequately answered, another similar primary study may have little value.
In clinical research, failure to consider prior evidence has been associated with unnecessary or poorly designed trials, avoidable participant risk, and wasted resources. Evidence synthesis can therefore serve not only as a research product but as a safeguard against unnecessary data collection.
Synthesis Can Also Reveal Exactly What Primary Study Is Needed
Evidence synthesis and primary research should not be treated as competitors.
A good synthesis can reveal that most studies are underpowered, use weak measurements, exclude an important population, lack long-term outcomes, or rely on designs vulnerable to the same source of bias.
That information can make the next primary study substantially better targeted.
The REVEAL framework makes this relationship explicit for trials. Prior evidence can help determine whether a new trial is needed while also informing sample-size calculations, outcome selection, and other design decisions.
In this sense, synthesis can be the bridge to better primary research rather than an alternative that prevents it.
A Systematic Review Is Particularly Useful When the Question Is Focused and Multiple Relevant Studies Exist
A systematic review is well suited to questions for which relevant primary studies already exist and can be identified using explicit eligibility criteria.
The review can assess the available evidence systematically and, where appropriate, estimate an overall effect or examine variation across studies.
If this describes your research problem, consider more specifically whether a systematic review would be more useful than another primary study.
The answer depends partly on whether a recent, rigorous review already exists. Conducting another review of the same evidence can itself become redundant.
A Scoping Review May Be More Appropriate When You Need to Map the Evidence
Sometimes you do not yet have a sufficiently narrow question for a conventional systematic review.
You may need to understand how a concept has been studied, which populations and methods appear in the literature, what types of evidence exist, or how research activity is distributed across a broad field.
A scoping review can be useful for mapping such evidence. It should not be selected merely because the topic is broad or because meta-analysis seems difficult. The review question and intended use should determine the methodology.
Mapping evidence can also reveal whether a more focused systematic review or new primary study is warranted next.
Meta-Analysis Is Not Synonymous With Systematic Review
A systematic review does not necessarily contain a meta-analysis, and a statistical pooling exercise is not automatically a rigorous systematic review.
Meta-analysis is appropriate when quantitative results are sufficiently compatible and statistical combination answers a meaningful question. Important differences among studies may sometimes make pooling inappropriate or require more sophisticated modeling and cautious interpretation.
Even without meta-analysis, a systematic review can provide valuable evidence by systematically identifying studies, evaluating their credibility, and synthesizing patterns and limitations.
Do Not Conduct a New Review Merely Because You Prefer Reviewing to Collecting Data
Evidence synthesis also needs justification.
If a recent, high-quality systematic review already answers the same question and no important new evidence has appeared, another nearly identical review may contribute little.
Recent guidance on planning clinical trials makes recency explicitly topic-dependent. In rapidly changing areas, a review with a relatively recent search date can become outdated quickly; in stable fields, an older review may remain useful when the question has been settled.
Watch Out
“There is no systematic review published this year” is no stronger a research justification than “there is no primary study from our institution.” Before conducting another review, determine whether the existing synthesis is methodologically adequate, current enough for the field, and sufficiently aligned with your question.
Primary Research Is Not Automatically More Original Than Evidence Synthesis
Researchers sometimes treat primary data collection as the more substantive form of research and evidence synthesis as secondary in importance.
That distinction is misleading. A rigorous synthesis can answer questions no individual primary study can answer, reveal patterns across studies, quantify uncertainty, identify sources of heterogeneity, assess the credibility of an evidence base, and prevent unnecessary research.
Conversely, a small primary study that reproduces an already well-established result may add less knowledge than a carefully designed synthesis of the entire literature.
Originality should therefore be judged by contribution, not by whether the researcher personally collected every observation.
Feasibility Should Influence the Choice but Not Determine It
Resources matter. A primary study may require recruitment, laboratory facilities, fieldwork, longitudinal follow-up, or substantial funding. A rigorous systematic review also requires methodological expertise, searching, screening, appraisal, extraction, synthesis, and careful reporting.
Choosing the easier project regardless of the evidence need is poor research prioritization.
Instead, identify the research question that matters, determine which form of evidence can answer it, and then evaluate whether the project is feasible. If the ideal study cannot be conducted, narrow or reformulate the question rather than pretending that a convenient design answers something it cannot.
Sometimes You Need Both
A mature research program may appropriately include evidence synthesis followed by primary research.
The synthesis establishes what is already known and identifies the uncertainty that remains. The primary study is then designed specifically around that uncertainty. When the new study is completed, its findings can be interpreted in relation to the earlier evidence rather than as an isolated result.
This cumulative sequence is preferable to repeatedly conducting primary studies first and asking afterward how they fit together.
