What Date Range Should You Use for a Literature Search?

Should You Search the Last Five Years, Ten Years, or Everything?

You open a research database and encounter one of the easiest filters to understand: publication year.

Perhaps you have also heard familiar advice such as “use studies from the last five years” or “references should not be more than ten years old.” A recent cutoff certainly makes a search smaller, and newer literature may be especially important for rapidly changing topics.

But a convenient number of years is not automatically a defensible research decision.

A five-year window could be entirely appropriate for a question about a technology introduced four years ago. The same restriction could be difficult to defend in a study examining a theory developed decades earlier, a long-standing educational practice, the historical development of a field, or an intervention whose foundational evidence predates the cutoff.

The appropriate date range therefore cannot be chosen from a universal rule. It has to follow from what you are trying to find.

Use the Time Period Your Research Question Can Actually Justify

A Date Range Is an Eligibility Decision, Not Just a Database Convenience

There is no general rule that research must come from the last five years

The familiar “last five years” rule is often treated as though it were a standard principle of research methodology. It is not.

Different research questions require different temporal coverage. Some questions concern rapidly developing technologies or policies. Others depend on theories, methods, or evidence accumulated over decades. Still others explicitly ask how something has changed over time.

A publication does not become irrelevant merely because it crosses an arbitrary age threshold.

The two can overlap, but they are not equivalent. A recent paper can be only marginally relevant, while an older publication may contain the original theory, instrument, method, or empirical finding that later research still relies on.

Start with the research question, not the calendar

The first question should not be “How many years should I search?” Ask instead:

When could evidence relevant to this question reasonably have begun to exist?

For some topics, the answer provides a natural lower boundary.

If you are studying the educational use of a particular technology, evidence cannot predate the technology itself. If you are evaluating the effects of a policy introduced in a known year, literature examining that policy could not meaningfully precede its introduction. If you are studying a diagnostic procedure that became available after a particular regulatory approval or technical development, that event may provide a defensible starting point.

Cochrane gives precisely this type of justification for date restrictions. Its current Handbook allows date limits when they follow from the eligibility criteria and it is known that relevant studies could only have been reported during a particular period. It gives the example of an intervention involving epinephrine autoinjectors, where a starting date can be tied to approval of the first device rather than an arbitrary preference for recent studies.

Ask what happened at the proposed starting date

A useful test for any lower date boundary is surprisingly simple:

Why this year?

If your answer is merely “because ten years seems reasonable,” the boundary may need reconsideration.

Stronger rationales are connected to the phenomenon itself.

This reasoning also helps distinguish the question of how far back you should search the literature from the mechanical act of selecting years in a database interface.

Rapidly changing topics may justify greater emphasis on recent literature

Recency can matter considerably when the object of study changes quickly.

Artificial intelligence provides an obvious example. Research concerning a particular generation of AI systems, regulatory environment, platform capability, or educational practice may become dated more quickly than research on a long-established psychological construct.

But even in a fast-moving field, “recent” should not become shorthand for “relevant.”

Older work may still be needed to explain the theoretical construct being measured, establish prior technological developments, identify earlier forms of the phenomenon, or understand what changed when the newer technology appeared.

A sensible strategy may therefore use recent evidence heavily for the current state of the topic while retaining older literature where it performs a different intellectual function.

Foundational literature does not expire on schedule

Suppose a theoretical framework central to your study was introduced 25 years ago. A five-year publication restriction would exclude the original source while leaving recent articles that cite or reinterpret it.

That may be inappropriate if you need to explain the theory itself.

Similarly, an original measurement instrument, methodological paper, landmark trial, or defining conceptual publication may remain directly relevant long after newer research appears.

This is one reason a literature review should not be assembled by applying the same recency rule to every source. Different publications serve different purposes.

Systematic reviews should be especially cautious about arbitrary date limits

For a systematic review, a publication-date limit affects which evidence can enter the review. It therefore has methodological consequences.

Cochrane's current guidance states that searches should capture as many studies as possible that meet the eligibility criteria and that publication-date restrictions should be justified. Its mandatory methodological expectation is that date restrictions in the search should be used only when corresponding date restrictions exist in the study eligibility criteria. Arbitrary limits are not recommended for Cochrane intervention reviews because eligible data may be lost.

JBI guidance takes a similarly cautious approach. Date limits may be appropriate for a recent intervention or innovation, but setting the limit too recently can omit relevant and seminal early studies. JBI therefore recommends making such decisions from informed knowledge of the topic and relevant key papers rather than applying an arbitrary period.

Your date criteria and database search limits should correspond

If your review protocol specifies that studies published before a particular date are ineligible, applying a corresponding database restriction may be reasonable.

The reverse is more problematic. If studies from all years are theoretically eligible but the database search is restricted to recent years merely for convenience, the search cannot identify part of the evidence base that the eligibility criteria say should count.

This is why the decision about whether to establish inclusion and exclusion criteria before the formal search matters. For systematic evidence synthesis, temporal eligibility should normally be decided as part of the review design rather than improvised after seeing the size of the result set.

A lower date boundary and an upper date boundary solve different problems

Researchers often speak of a “date range” as though both ends perform the same function.

They do not.

The lower boundary answers how far back the search should extend. The upper boundary usually identifies how current the search is.

For a completed review, the upper boundary should normally reflect the actual final search date rather than an arbitrary earlier year.

Cochrane emphasizes that published systematic reviews should be as current as possible. Its current methodological standards require searches to be rerun or updated within 12 months before publication, with six months preferred when feasible, if the original search is older.

An old search can therefore create a different problem from an overly early starting date: the review may simply be out of date.

Publication date is not the only date a database may store

This technical issue is easy to overlook.

Bibliographic databases may contain several date fields, including publication date, electronic publication date, database entry date, indexing date, creation date, and update date. These fields do not necessarily represent the same event.

Cochrane cautions that date restrictions need careful implementation because database providers offer different date fields and publication metadata may be inconsistent. An electronic publication date, for example, may differ from the print publication date. Its technical supplement also shows that superficially similar date fields in PubMed and Ovid MEDLINE are not always equivalent.

If the temporal boundary is methodologically important, check what the database's date filter actually searches.

A slightly wider database range may sometimes be safer than the eligibility period

Suppose your eligibility criteria begin on 1 January 2020. It might seem obvious to restrict the database to records labelled 2020 onward.

Metadata can complicate that logic. Online-first publication, print publication, database entry, and indexing may occur at different times. Cochrane consequently advises caution and notes that a search may need a somewhat wider date range than the period of interest to account for inconsistent publication dates in database records.

You can retrieve slightly beyond the eligibility boundary and apply the definitive rule during screening if that reduces the risk of accidental omission.

This reflects a broader principle already encountered elsewhere in literature searching: eligibility can be more precise than retrieval.

Updates are a special case

If you are updating an existing systematic review, you do not necessarily need to repeat the entire historical search from the beginning.

Cochrane notes that an update may rerun the previous strategy with date limits covering the period after the most recent search, provided the strategy remains appropriate. For updates, the relevant database field may be the date of entry into the database rather than publication date, because the objective is to identify records newly available since the previous search.

This is a reminder that “date range” is not one universal technical setting. The correct date field depends on what the search is trying to accomplish.

Different parts of your literature review may need different temporal coverage

Not every literature search is a systematic review, and not every part of a thesis or research paper has the same information need.

Suppose you are studying student adoption of generative AI for academic writing. You might need:

• recent studies describing current generative AI use;
• older theoretical literature on technology adoption;
• the original development and validation literature for an instrument;
• earlier research on automated writing technologies that provides historical context.

A single five-year filter applied to every search could distort that evidence base.

It may be more defensible to conduct searches with different temporal purposes while remaining transparent about how and why each body of literature was identified.

A date restriction should not be a substitute for fixing an overly broad search

This is one of the most common practical temptations.

Your search retrieves 12,000 results. Selecting “2022–2026” reduces it to 3,000. Problem solved?

Not necessarily.

If those 12,000 records appeared because one search term is ambiguous, the date restriction has not corrected the search strategy. It has merely removed older relevant and irrelevant records together.

Before using publication year to control volume, determine why the search is too broad. Conceptual refinement is generally preferable to an arbitrary chronological cutoff.

Nor should a very recent date range be used to manufacture a research gap

Suppose you search only the last three years and find four studies. It would be misleading to conclude that the topic has barely been researched if a substantial literature exists outside your chosen period.

A date-limited search answers a date-limited question.

If you claim that research is scarce, the scope of that claim should match the scope of the search. Better still, challenge the temporal restriction before interpreting scarcity.

A small result set might indicate a genuinely emerging topic, but it can also mean that your search has become too narrow.

Report the date range and the date you actually searched

For reproducible evidence synthesis, readers need to know both the temporal coverage of the search and when it was run.

Cochrane reporting guidance recommends specifying the dates of coverage for databases and the date of the last search whose studies were incorporated. It also requires authors to declare limits such as publication-date restrictions and provide the exact database strategies, including limits and filters, so the search can be evaluated and replicated.

These are different pieces of information:

Both may matter when someone later evaluates how comprehensive and current your search was.

Choosing a Date Range for an Emerging Research Topic

The important point is that the date range emerges from the research question and the history of the phenomenon. The researcher can explain why relevant evidence should begin at that point.

If no defensible lower boundary exists, searching further back may be necessary. That becomes the more specific question of how far back the literature search should extend.

Common Mistakes When Choosing a Literature Search Date Range

Choose a Starting Date You Can Explain Without Saying “Because Five Years”

A defensible date range should survive a simple question from another researcher: Why did you start there?

Your answer should connect the boundary to the research problem, not merely to convention or convenience.

You may ultimately choose five years, ten years, twenty years, a historically defined starting point, or no lower publication-date restriction at all. Any of those can be defensible in the right study.

The number itself is not the methodology. The rationale is.

Before Applying a Date Limit to Your Literature Search

Questions About Date Ranges for Literature Searches

There Is No Magic Number of Years

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.