03 · What You Need to Know
The Right Date Range Comes From the Question, Not From a Universal Rule
There Is No Universal Expiration Date for Research
Research papers do not become invalid on their fifth, tenth, or twentieth birthday.
Some evidence becomes less applicable as technologies, practices, institutions, populations, or policies change. Other research remains highly relevant for decades. Foundational theories, seminal experiments, established methodological papers, historical evidence, and studies of relatively stable phenomena may remain important long after publication.
The useful question is therefore not:
“How old is this study?”
It is:
“Has anything about the question, phenomenon, intervention, population, context, or evidence changed enough that this study no longer belongs in the body of evidence I am trying to examine?”
That shift turns publication year from a convenient filter into a substantive methodological decision.
Publication Date Can Be a Legitimate Eligibility Criterion
JBI recognizes publication date as one of the publication characteristics that may form part of a review's inclusion criteria. At the same time, JBI emphasizes that eligibility criteria should be explicit, reasonable, scientifically sound, and justified.
Cochrane similarly permits date restrictions but requires them to be justified. Its current guidance for intervention reviews advises that date restrictions in database searches should be used only when corresponding date restrictions exist in the study eligibility criteria and when there is a defensible reason relevant studies could only have been reported during a particular period.
The key distinction is therefore not between having and not having a date cutoff. It is between a question-driven date restriction and an arbitrary date restriction.
Question-driven date restriction
The boundary corresponds to a meaningful feature of the phenomenon, intervention, policy, technology, historical period, or review objective.
Arbitrary date restriction
The boundary is chosen primarily because a round number such as five or ten years seems convenient or because the researcher wants fewer papers to screen.
A Date Cutoff Makes Sense When the Phenomenon Did Not Yet Exist
The clearest justification occurs when studies before a certain date could not possibly examine the phenomenon defined by your review.
If you are reviewing experiences with a technology introduced in a particular period, searching decades before that technology existed may add little. The same reasoning can apply to a medication before approval, a diagnostic procedure before introduction, a policy before implementation, or an educational platform before deployment.
Cochrane gives a comparable example involving epinephrine autoinjectors: a search beginning with the year the first device was approved can be justified because earlier reports could not meet eligibility criteria concerning that intervention.
Notice the logic. The date itself is not what makes earlier studies irrelevant. The intervention's availability does.
A Major Change in Technology Can Create a Meaningful Boundary
Technology-intensive topics often require careful temporal reasoning because the same label can describe substantially different systems across time.
Consider a review of student experiences with generative AI feedback. Studies of automated feedback from 2005 may concern rule-based systems or other technologies that operate very differently from contemporary generative models.
You have at least two methodological options.
You could define eligibility conceptually, including only technologies that satisfy your definition of generative AI regardless of publication year. Alternatively, if a particular technological development creates a defensible historical starting point for the phenomenon, you might also use a date boundary.
The first approach is often more precise because it excludes studies for what technology they investigated rather than simply when they were published. A date restriction becomes useful when time closely corresponds to the substantive boundary you need.
Policy and Institutional Change Can Also Matter
Suppose you are reviewing an educational practice after implementation of a major national curriculum reform. Research conducted under the previous curriculum may describe a materially different institutional environment.
Or imagine a review of healthcare practice following introduction of a new diagnostic definition. Studies using an older definition may classify participants differently from contemporary research.
In these situations, a temporal boundary can be defensible because the context before and after the change is meaningfully different.
But be precise about what changed. “Research before 2015 is outdated” is much weaker than “the policy implemented in 2015 materially changed the practice being examined, so the review focuses on evidence generated under the post-implementation system.”
Historical Questions May Require Old Studies
Sometimes older evidence is not a nuisance to be filtered out. It is the evidence the question requires.
If you are examining how a concept developed, how professional practice changed, how an educational technology evolved, or how scholarly understanding shifted over time, restricting the review to recent publications could remove the very evidence needed to answer the question.
A review of the historical development of distance education that includes only papers from the last five years would tell you primarily how contemporary authors describe the past, not necessarily how the field itself developed.
The appropriate date range therefore depends on what kind of claim you intend to make.
Foundational Studies Do Not Stop Being Foundational Because They Are Old
Older studies may introduce theories, constructs, instruments, methods, or findings that later research builds upon. Replacing such a source with a recent article merely because the recent article cites it can weaken scholarly accuracy.
If you are discussing the origin of a theory, the original theoretical work may be the most appropriate source. If you are describing development of an instrument, the original validation paper may be necessary. If you are explaining a landmark discovery, a contemporary review cannot always substitute for the original study.
This does not mean old foundational studies should automatically enter the formal evidence synthesis. Their role depends on your review question and eligibility criteria.
An older paper may be indispensable contextual literature while remaining ineligible for the particular empirical synthesis you are conducting. This is the same distinction that matters when an important study falls outside the review's formal criteria.
“Use Sources From the Last Five Years” Is Not a Universal Research Rule
Students frequently encounter instructions to use sources published within the last five years. Such requirements may be imposed by an instructor, institution, assignment, journal, or discipline for a particular purpose. If that requirement applies to your work, you need to follow it.
But it should not be confused with a universal methodological principle governing literature reviews.
Recency may be particularly important in fast-changing areas such as emerging technologies, current policy, rapidly evolving clinical practice, or fields where evidence becomes obsolete quickly. Even then, older evidence may remain necessary for theory, historical context, methods, or comparison.
The methodological question is not whether five years is inherently better than six. It is whether the evidence remains appropriate for the claim you are making.
Do Not Use a Date Cutoff Merely to Reduce the Number of Search Results
A large search result can make a recent-date filter very tempting.
Suppose your search returns 4,000 records. Restricting the search to the last ten years reduces the result to 1,200. That certainly reduces the workload.
It does not, by itself, justify the restriction.
Cochrane specifically discourages arbitrary date limits for intervention reviews because eligible evidence can be lost. Date restrictions can also affect effect estimates in systematic reviews, so the decision is not always methodologically neutral.
Watch Out
Do not choose a publication cutoff solely because the unrestricted search produces more records than you want to screen. A workload problem and an eligibility rationale are different problems.
If Date Is an Eligibility Criterion, Align the Search With It Carefully
If studies before a particular date are genuinely ineligible, restricting database searches by date can improve efficiency. But the implementation deserves care.
Cochrane notes that databases contain different date fields, such as publication date, database entry date, creation date, and update date. Electronic and print versions of the same article can also carry different dates.
As a result, an apparently precise database filter may not correspond perfectly to the publication-date rule in your eligibility criteria. Searching a slightly wider period and applying the final criterion during screening can sometimes reduce the risk of missing eligible records near the boundary.
For an updated systematic review, the relevant date logic may be different again. Search updates are often organized around database entry dates or the date of the previous search rather than imposing a new publication-age criterion on the evidence.
Publication Date and Study Date Are Not the Same Thing
A paper published in 2025 might report data collected in 2016. A paper published in 2018 might analyze a long-running cohort that began decades earlier.
If your substantive concern is whether evidence reflects a particular technological, policy, or historical period, publication year may be only a proxy for the characteristic you actually care about.
Publication date
When the report became publicly available.
Study or data-collection period
When the underlying participants, interventions, exposures, or observations were actually studied.
Suppose a new educational policy took effect in 2020 and you want evidence about experiences under that policy. A paper published in 2022 using data collected in 2018 may not represent the post-policy context, while a publication-year cutoff of 2020 would incorrectly suggest that it does.
When time matters substantively, define the temporal characteristic that actually matters.
Newer Does Not Automatically Mean Better
A recent study may reflect the current context more closely, but publication year does not establish methodological quality.
A rigorous study from 2012 does not become methodologically weak because a poorly designed study appeared in 2025. Likewise, a recent article does not supersede an older study merely by existing.
Methodological quality and relevance require their own assessment. As with distinguishing relevance from study quality, recency should not become a shortcut for judgments it cannot support.
An Older Study May Be Less Applicable Without Being “Bad”
Temporal change can affect applicability even when the study itself was well conducted.
An excellent study of online learning conducted when most students relied on desktop computers and slow internet connections may not transfer perfectly to a contemporary environment dominated by smartphones, ubiquitous broadband, videoconferencing, and learning-management platforms.
The appropriate interpretation is not necessarily “old study equals bad study.” It may be:
“This study provides credible evidence about an earlier technological context, but its applicability to the current environment requires caution.”
That distinction preserves both methodological quality and contextual relevance.
Do Not Invent a Date Restriction After Seeing an Inconvenient Study
Suppose your protocol contains no publication-date restriction. During screening, you find an older study with findings that conflict with more recent evidence. You decide that studies before 2015 are probably too old to matter.
That is no longer simply a question about recency. You are considering a post hoc eligibility change after seeing the evidence.
For systematic reviews, eligibility criteria should generally be prespecified. If screening reveals a genuine need to revise them, the change should have a methodological justification and should not be driven by study findings.
If a temporal boundary genuinely becomes necessary after screening starts, follow the same safeguards involved in changing inclusion criteria during a review. Document the amendment and apply it consistently to all affected studies rather than selectively excluding inconvenient older evidence.
Different Review Types May Need Very Different Time Horizons
There is no single ideal date range because review purposes differ.
| Review situation |
Possible temporal approach |
Reason |
| Emerging technology |
Potentially recent or technology-defined period |
Earlier studies may concern materially different technologies |
| Post-policy implementation review |
Begin around the relevant implementation period |
The policy creates a substantive contextual boundary |
| Historical development of a concept |
Include older and foundational literature |
Temporal development is part of the question |
| Established phenomenon with no obvious temporal break |
A broad date range may be appropriate |
Age alone may not affect relevance |
| Systematic review update |
Search from the previous search using an appropriate update strategy |
The goal is to identify newly available evidence without unnecessarily repeating completed work |
The date range should therefore emerge from the phenomenon and the claims you intend to make, not from a generic expectation that a literature review should look recent.
