When Is It Better to Report That a Source Could Not Be Obtained Rather Than Rely on Secondary Descriptions?

Is an Indirect Description Better Than Admitting You Could Not Get the Original?

You have identified an apparently important source, but repeated attempts to obtain it have failed. Another paper describes its findings. A review summarizes its methods. Perhaps several authors repeat the same quotation or statistic.

It can be tempting to assemble those descriptions and proceed as though you now know what the original source says. Sometimes secondary citation is legitimate, particularly when an original work is genuinely unavailable. But a secondary description always introduces another interpretive layer between you and the evidence.

There are therefore situations in which the more rigorous statement is simply that the original source could not be obtained. That is not an admission of poor scholarship when reasonable retrieval attempts have failed. It is a statement about the boundary of what you were able to verify.

Report Non-Retrieval When Direct Verification Matters

“I Know What Another Author Said About It” Is Not the Same as “I Read It”

Secondary descriptions carry another researcher's interpretation

If Paper B describes Paper A and you can read only Paper B, your direct evidence is Paper B. You know how B represented A. Unless the relevant passage or data can be independently verified, you do not know whether B summarized A completely, interpreted it correctly, omitted an important qualification, used the correct numerical value, or accurately represented the original context.

This is why citation guidance generally encourages researchers to locate and read original sources whenever possible. University guidance implementing APA 7, for example, advises using secondary citations only when the original cannot be found or read and placing in the reference list the source that was actually consulted.

Secondary citation and secondary evidence are not quite the same problem

A secondary citation is a citation practice: you attribute an idea or statement to an original work while making clear that you encountered it through another source. Whether this is acceptable depends partly on the citation style and research context.

A secondary description creates a broader evidential problem. You might know from a review article that an inaccessible study “found no significant difference,” but that sentence may not tell you the sample size, effect estimate, confidence interval, statistical model, missing-data treatment, outcome definition, or even precisely which comparison was tested.

Correct citation syntax does not magically supply missing methodological information. The question of when secondary citation is acceptable should therefore be separated from whether the indirect evidence is sufficient for the analytical task you are attempting.

Report non-retrieval when the missing information is essential

If your purpose requires direct examination of the source, a secondary summary may simply be insufficient. This is particularly important for systematic-review eligibility assessment, data extraction, risk-of-bias assessment, replication, detailed methodological critique, exact quotation verification, and other tasks where the underlying document itself contains necessary evidence.

The Cochrane Handbook explicitly recognizes situations in which potentially relevant study information remains incomplete or unobtainable after efforts to clarify eligibility or obtain missing information. Such cases should be recorded appropriately rather than converted into evidence the reviewer never actually saw.

PRISMA explicitly allows reports to remain unretrieved

PRISMA 2020 includes a “Reports not retrieved” stage in its flow diagram. Published clarification of PRISMA 2020 and PRISMA-S explains that reports may remain unretrieved because, for example, a journal is unavailable locally or through interlibrary loan, authors do not respond, or links are broken.

This is an important methodological point. A systematic review is not required to pretend that every identified report was successfully obtained. Transparent non-retrieval is preferable to quietly treating an abstract, review, or citation as though it were the full report.

Do not convert a secondary author's conclusion into original data

Suppose a review states that an inaccessible 1985 study “reported improved achievement.” That statement may justify saying that the review characterizes the study as reporting improved achievement, provided you cite the review appropriately. It does not give you the study's effect size, exact outcome measure, analytical procedure, uncertainty, or complete interpretation.

Similarly, if a secondary source reproduces a number, you still need to know what the number represents before using it analytically. A percentage without its denominator, an effect estimate without its model, or a quotation without its surrounding qualification can be technically accurate yet unsuitable for your purpose.

Multiple secondary sources do not automatically solve the problem

If five papers repeat the same description of an inaccessible study, you may feel increasingly confident that the description is correct. That confidence can be misleading if all five ultimately copied the same intermediate source.

Citation chains propagate errors. An inaccurate statistic, quotation, or interpretation can acquire apparent authority through repetition without anyone returning to the original. Independent agreement is meaningful only when the secondary sources themselves independently examined the original and report enough information for you to assess their descriptions.

Sometimes the secondary account is itself what matters

There are legitimate situations in which you are interested in how a later author interpreted an earlier work rather than trying to reconstruct the original. Intellectual history, discourse analysis, citation analysis, and studies of how concepts travel through a literature may treat the secondary interpretation as evidence in its own right.

In that situation, cite the secondary work for its interpretation. Do not blur that research question with a claim that the interpretation establishes what the original source objectively contained.

Secondary citation may still be reasonable for genuinely inaccessible originals

Some sources are genuinely difficult or impossible to retrieve: historical works that were never digitized, documents held in inaccessible archives, unpublished material, lost reports, or works unavailable in a language the researcher can adequately evaluate. Citation systems accommodate this reality.

For example, university APA guidance commonly instructs researchers who cannot locate an original source to identify the original in the text using an “as cited in” construction and include only the work actually read in the reference list. The exact formatting varies by citation style, so follow the style required for your work.

Secondary citation is therefore not inherently improper. The problem arises when it disguises evidential uncertainty or substitutes for a reasonably obtainable primary source.

Distinguish “unavailable to me” from “does not exist”

If you cannot retrieve a source, say that it could not be obtained or located through the routes you used. Do not escalate that into a claim that the document does not exist, was never published, or contains no relevant information.

The same principle applies to inaccessible appendices and protocols. Failure to retrieve an object is evidence about your access to it, not necessarily evidence about its historical existence or contents.

Document what you attempted when non-retrieval affects the research

For a casual literature search, a detailed audit trail may be unnecessary. For a systematic review or other reproducible investigation, record enough information to show what happened: citation, retrieval routes, library requests, author contact where relevant, and final status.

If you are unsure how far to pursue the source before declaring it unavailable, apply a proportionate rule for how much effort to spend obtaining a difficult paper. The more consequential the source is to your conclusions, the stronger the case for additional retrieval effort before stopping.

When a Review Describes a Study You Cannot Obtain

The transparent outcome is less satisfying than magically recovering the missing effect size, but it is methodologically stronger. An explicit gap tells readers what remains unknown. A reconstructed result can make uncertainty disappear only on paper.

Common Mistakes When the Original Source Cannot Be Obtained

Choose Transparency When Indirect Evidence Cannot Support the Task

The underlying principle is simple: your prose should never imply a shorter evidential chain than the one you actually have. If you read B's description of A, then B is part of your evidence whether or not mentioning that fact makes the sentence slightly less elegant.

Research writing occasionally becomes more accurate by becoming less impressive. “We could not verify this from the original report” may be a stronger scholarly sentence than a confident paragraph built on evidence nobody in the current analysis actually saw.

Before Relying on a Secondary Description of an Unavailable Source

Questions About Unavailable Sources and Secondary Descriptions

Do Not Hide an Evidence Gap Behind a Secondary Citation

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.