When Should You Retrieve the Full Text of a Paper?

When Is the Abstract No Longer Enough?

You have screened the title and abstract of a paper, but the decision is still not obvious. Perhaps the participants are described only as “students,” even though your review is limited to higher education. The abstract mentions an intervention that sounds eligible but does not explain exactly what it involved. Or everything appears to fit, and you now need to confirm eligibility from the complete report.

Should you retrieve the full text?

Usually, yes. Full-text retrieval becomes necessary when a record survives initial screening as potentially eligible or when the information available in the title and abstract cannot support a confident inclusion or exclusion decision.

The important point is that full-text retrieval is not simply “reading more.” In a structured review, it marks a transition from preliminary screening to a more complete eligibility assessment. Knowing when to make that transition can save considerable time while reducing the risk of excluding relevant evidence too early.

Retrieve the Full Text When a Paper Could Still Be Eligible

Full-Text Retrieval Begins When Preliminary Screening Can No Longer Answer the Eligibility Question

Title and Abstract Screening Is Only the First Filter

A literature search can retrieve hundreds or thousands of records. Obtaining and reading every complete article would usually be inefficient, which is why structured study selection proceeds in stages.

Cochrane describes a typical process in which reviewers first examine titles and abstracts to remove obviously irrelevant reports, while generally being over-inclusive at that stage. They then retrieve the full texts of potentially relevant reports and assess those reports against the review's eligibility criteria.

This distinction gives title and abstract screening a deliberately limited purpose. You are not necessarily establishing final eligibility. You are deciding whether the record can confidently be excluded or should remain under consideration.

That is why screening titles and abstracts should focus on the information required for that preliminary decision rather than trying to reconstruct the entire study from a few hundred words.

Retrieve the Full Text When the Record Appears Eligible

The clearest reason to retrieve a full text is that the title and abstract suggest the study may satisfy your criteria.

Suppose your review concerns university students' experiences of receiving generative AI-generated feedback. An abstract reports interviews with undergraduates who received ChatGPT-generated comments on their academic writing. Nothing in the abstract establishes ineligibility.

You should not treat that as final inclusion merely because the abstract looks ideal. The full report may reveal information that changes the decision. Perhaps only a small subgroup actually received AI-generated feedback, the participants were enrolled in a non-university setting, or the article reports the same underlying study as another publication already identified.

In Cochrane intervention reviews, inclusion decisions for potentially eligible studies should be based on full texts when possible. At least two people working independently are required to make the final determination of whether such studies meet the eligibility criteria.

Retrieve the Full Text When an Essential Detail Is Unclear

A second common reason is uncertainty.

An abstract may not tell you:

• the precise characteristics of the participants;
• whether an eligible subgroup can be separated from a mixed sample;
• exactly how an intervention or exposure was defined;
• whether the setting satisfies your criteria;
• what study design was actually used;
• whether an outcome relevant to eligibility was measured;
• whether the article contains original empirical evidence; or
• whether another eligibility requirement was satisfied.

The correct response depends on whether the missing information matters to eligibility. If it does, and the record could still qualify, the absence of information is usually a reason to investigate further rather than exclude.

Those situations are not equivalent. Treating them as though they were is one of the easiest ways to lose potentially relevant evidence during screening.

You Do Not Need Full Text for Every Clearly Irrelevant Record

Full-text retrieval should not become a ritual performed for every database result.

If your review includes only university students and the title and abstract clearly describe preschool children, retrieving the complete article simply to verify that preschool children are not university students adds no meaningful evidence to the decision.

This is the efficiency gained from not treating every search result as a paper that must be read in full. Search results are candidates. Screening determines which candidates require progressively closer examination.

Full-Text Assessment Is More Than Reading the Introduction

Once you retrieve the paper, the purpose is to assess the eligibility criteria that could not be established during preliminary screening. The relevant information may be scattered across the report.

Population characteristics may appear in the Methods section. Details of an intervention may be distributed across the Methods, supplementary material, and appendices. Study dates or setting may appear only in a participant description. A paper's relationship to another publication may become apparent from recruitment dates, sample sizes, institutional sites, author lists, or intervention details.

Full-text screening is therefore targeted reading. You are looking for evidence that establishes whether the underlying study meets the review's criteria.

Cochrane also emphasizes that the study, rather than an individual report, is the primary unit of interest. A single study may have several reports, each containing different information needed to establish eligibility or understand the study.

Do Not Assume One Paper Equals One Study

Retrieving a full text sometimes reveals that the paper is not an independent study at all.

A trial may produce a protocol, primary outcomes article, follow-up report, secondary analysis, conference abstract, and process evaluation. Conversely, one article may contain information about several related studies.

If two reports appear to describe the same participants, recruitment period, sites, interventions, and study characteristics, they may need to be linked before you decide how they contribute to the review. Cochrane requires multiple reports of the same study to be collated so that they are not mistakenly treated as independent studies.

This is why you eventually need to identify multiple papers originating from the same study or dataset. Full-text retrieval can provide the details needed to recognize those relationships.

What if You Cannot Obtain the Full Text?

A potentially eligible study should not automatically become an excluded study simply because the article is difficult to obtain.

First, distinguish between ineligibility and inability to obtain enough information to determine eligibility. These are different methodological situations.

Depending on the review and the material available, you might look for another report of the same study, check trial or study registrations, examine institutional or subject repositories, use library services, or contact the study authors when appropriate.

Cochrane guidance explicitly allows correspondence with investigators when eligibility remains unclear. If the required information remains unobtainable, the study may need to be recorded as awaiting classification rather than being assigned an unsupported exclusion reason.

Retrieving the Full Text Does Not Mean the Study Will Be Included

A record reaches full-text assessment because it survived preliminary screening, not because inclusion is guaranteed.

The complete report may reveal that:

• the population falls outside the intended scope;
• the intervention or phenomenon differs from what the abstract suggested;
• the study design is ineligible;
• the publication is not the type of evidence required;
• the setting does not meet a justified contextual criterion; or
• another essential eligibility requirement is not satisfied.

This is normal. The point of retrieving the full text is precisely to resolve those questions.

For systematic reviews, records that appeared potentially eligible but are excluded after closer examination need appropriate documentation. Cochrane distinguishes these from records discarded during initial title and abstract screening and requires sufficient documentation of study selection to support the review flow diagram and excluded-study reporting.

Keeping track of why studies were excluded after closer assessment therefore becomes increasingly important once you reach the full-text stage.

Full-Text Retrieval and Full-Text Screening Are Not Quite the Same Thing

These terms are sometimes used interchangeably, but distinguishing them is useful.

PRISMA 2020 preserves a similar distinction in its study-selection flow by separating reports sought for retrieval from reports assessed for eligibility. This matters because some reports sought may not actually be retrieved, and those should not be silently counted as though they underwent full-text assessment.

Different Review Types May Use the Full Text Differently

The logic described above is particularly important for systematic and other structured evidence syntheses. JBI guidance for scoping reviews likewise calls for reporting the process by which sources of evidence were screened and assessed for eligibility, with the numbers screened, assessed, and included reported transparently.

A traditional narrative review may not have a formal two-stage eligibility process. Researchers often retrieve complete articles because they are theoretically important, historically influential, methodologically informative, or useful for developing an argument.

Even there, however, the practical distinction remains helpful. You do not need to obtain and deeply read every record merely because it appeared in a search. The depth of reading should follow the purpose for which the source is being considered.

Which of These Papers Actually Needs the Full Text?

Notice that Records B and C reach full-text assessment for different reasons. B looks eligible and needs confirmation. C is uncertain and needs clarification. Both justify retrieval because neither can yet be defensibly excluded.

Common Mistakes About Full-Text Retrieval

Retrieve the Paper When More Information Can Change the Decision

You can make the retrieval decision by asking a practical question: Would information in the complete report be necessary to establish whether this study belongs in the review?

If the answer is no because the record is already clearly ineligible, full-text retrieval is usually unnecessary for selection. If the answer is yes because the study looks eligible or an important criterion remains uncertain, retrieve the report.

Once you retrieve the paper, apply the eligibility criteria consistently. A single clearly failed essential criterion can be sufficient for exclusion in a structured review, so there is no methodological prize for continuing through every remaining criterion once the decisive reason has been established. Cochrane explicitly notes that criteria can be assessed in an order that allows the first failed criterion to serve as the primary exclusion reason.

The harder cases arise when the complete paper still does not answer the question. In that situation, the next problem is how to handle insufficient information without turning uncertainty into an arbitrary decision.

Before You Retrieve a Full-Text Paper

Questions About Retrieving and Screening Full-Text Papers

Retrieve the Full Text When the Paper Still Has a Real Chance of Belonging in the Review

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.