Which Papers Make the Best Starting Points for Citation Chaining?

What Makes a Paper a Good Starting Point for Citation Chaining?

Citation chaining has to begin somewhere. Perhaps your database search returned hundreds of records, you already know several important papers, or a recent review gives you an obvious entry point. Which paper should become your starting seed?

The tempting answer is to choose the most highly cited paper. Another common strategy is to choose the newest review, the paper in the most prestigious journal, or simply the first clearly relevant study you encounter. Each can work, but none is universally the best choice.

A useful seed paper is one whose citation relationships are likely to expose literature relevant to your particular search objective. That makes seed selection partly a question of relevance and partly a question of where the paper sits within the literature you are trying to explore.

The Best Seed Paper Is Relevant and Well Positioned for What You Need to Find

Choose Seeds According to the Literature You Need Them to Reveal

First, understand what a seed paper actually does

A seed paper, or seed reference, is a known relevant publication used as the starting point for citation searching. In backward citation searching, you examine publications cited by the seed. In forward citation searching, you examine later publications that cite it.

The seed therefore determines where you enter a citation network. This matters because academic literature is not one uniformly connected collection. Papers are embedded in overlapping communities of authors, concepts, methods, journals, disciplines, and historical periods.

Starting from one paper gives you access to the literature connected to that paper. It does not automatically give you access to everything relevant to your question.

Relevance should come before citation count

A good seed should be substantively close to what you are trying to find. If your question concerns a particular intervention, phenomenon, population, theoretical construct, research method, or context, a paper addressing those elements directly is generally a more defensible seed than a famous paper that is only loosely related.

This distinction is easy to overlook because citation counts are visible and convenient. Yet a citation count tells you how often a publication has been cited within the coverage of a particular citation source. It does not tell you how useful its references and citing papers will be for your specific search.

A moderately cited empirical study that closely matches your eligibility criteria may therefore be a better seed than a famous conceptual article with thousands of citations spread across many unrelated applications.

Different seeds can be useful for backward and forward searching

The direction of the search changes what makes a seed useful.

This is one reason it helps to understand the distinction between backward and forward citation searching before choosing seeds.

Seminal papers can be excellent seeds, but they solve only part of the problem

A genuinely foundational publication can be particularly valuable when you want to understand where an idea originated and how subsequent research developed around it. Its forward citations may reveal decades of extensions, applications, criticisms, and modifications.

But being seminal is not the same as being universally connected. A foundational paper may belong to one theoretical tradition while another relevant tradition developed separately. It may also predate methods, populations, technologies, or applications central to your question.

Use seminal papers because their intellectual position makes them useful, not because the label itself guarantees comprehensive retrieval.

Review articles can function as useful maps

A well-conducted review can be an efficient seed for backward searching because its reference list may consolidate a substantial portion of earlier literature. It may also reveal terminology, key authors, theoretical disagreements, and clusters of studies that you did not encounter in your initial search.

However, a review's bibliography reflects its own question, eligibility criteria, search methods, publication date, and limitations. A review should therefore be treated as one map of the literature rather than as the literature itself. If a review is particularly relevant, you can use it to identify studies your database search may have missed while continuing to verify individual records against your own criteria.

Older papers can sometimes outperform newer ones as seeds

For forward citation searching, a very recent paper has an obvious limitation: later researchers have had little time to cite it. An older relevant paper may provide a much richer forward-citation trail.

Older papers can also help when terminology has changed. If contemporary researchers use a term that did not exist twenty years ago, searching backward from a modern paper may eventually reveal the vocabulary used by earlier researchers. Once you identify that vocabulary, you can supplement chaining with a new keyword search designed to retrieve older research described in historical terms.

Methodologically unusual papers may open different citation paths

Suppose most of your initial papers use quantitative surveys, but one relevant study uses ethnography. That qualitative paper may cite a methodological and theoretical literature absent from the survey papers. The same can happen across experimental and observational research, clinical and community settings, or computational and interpretive traditions.

For exploratory searching, such differences can be useful rather than inconvenient. A seed's value may lie precisely in the fact that it is not redundant with your other seeds.

A diverse seed set can be safer than one apparently perfect paper

If you are using a purposive set of seeds rather than all eligible records, consider what each paper adds to the others. Several nearly identical papers from the same research group may reproduce much of the same citation neighborhood. Papers spanning different periods, conceptual vocabularies, research teams, methods, or disciplinary homes may expose more varied connections.

This does not mean mechanically selecting one paper from every imaginable category. Diversity is useful only when the dimensions correspond to plausible divisions in the literature.

Systematic searching changes the seed-selection problem

If you are conducting a systematic search, be cautious about translating advice for exploratory literature searching directly into review methodology. The TARCiS statement recommends that supplementary backward and forward citation searching generally use all records meeting the review's inclusion criteria after full-text screening of the primary search. Its authors note that evidence is currently insufficient to determine an optimal number of seed references or reliably identify which included records will perform best.

TARCiS allows that deviations can sometimes be justified. For example, a very large number of included records may make comprehensive chaining burdensome, while a very small seed set might justify adding other pertinent records. The rationale and selection procedure should then be reported.

The best seed may reveal that you need another seed

Citation searching is often iterative. One seed can reveal a relevant publication from a different period or conceptual tradition, which can then become another seed. TARCiS recommends considering another iteration when citation searching identifies additional eligible records.

That does not mean following citations indefinitely. At some point you need to decide whether additional rounds remain productive. The important point is that your initial seed set does not have to contain every useful entry point you will eventually discover.

Why the Most Cited Paper May Not Be the Most Useful Seed

If this were an exploratory search, using several complementary seeds would likely be more informative than choosing Seed A merely because it has the largest citation count. If this were a systematic search governed by a protocol requiring citation searching of all included records, however, the relevant question would no longer be which one deserves to be the seed. All eligible records specified by the procedure would be used.

Common Mistakes When Choosing Citation-Chaining Seeds

Choose Seeds for Information Value, Not Prestige

Before selecting a seed, specify what you want citation chaining to accomplish. Are you looking for historical predecessors, later empirical studies, alternative terminology, a parallel disciplinary literature, or additional eligible records for a systematic review? The answer changes what makes a paper useful.

Most importantly, do not treat seed selection as a prestige contest. A paper can be obscure yet extremely informative because it sits exactly where your search needs an entry point.

Before Using a Paper as a Citation-Chaining Seed

Questions About Choosing Citation-Chaining Seeds

The Best Seed Is the Paper That Opens the Right Part of the Literature

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.