01Why one good paper beats a better search.
A database matches words. A reference list was assembled by somebody who knows the field, read the literature, and decided which parts of it mattered enough to cite. Those are not comparable instruments, and the second one is doing work the first cannot.
That is the whole idea behind chaining. Rather than trying to describe your topic well enough for a machine to recognise it, you find one paper that is unmistakably about it and then travel outwards along the connections other researchers have already made.
It is also how people in the field actually work, which is worth saying because chaining sometimes feels like a shortcut. It is not. It is the primary method, and keyword searching is the thing used to find the first paper.
02Backwards, through the references.
Open the reference list of your good paper and read it as a document rather than as an appendix. Several things are visible there that are visible nowhere else.
- Anything cited in the introduction is usually foundational to the topic rather than incidental, because introductions establish the ground.
- Anything cited in the method is where the tools came from: the instrument, the framework, the analytic approach. If you are doing something similar, these are your sources too.
- Names that appear more than once across a single reference list are people working continuously in the area, which is a name worth searching directly.
- Older items cited by several recent papers are the field's shared furniture, and citing one signals that you know where the conversation started.
- And the dates tell you when the area was active, which sometimes reveals that a topic you assumed was current had its argument twenty years ago.
03Forwards, through who cited it.
This is the half students almost never use and it is the more valuable one. Most databases and most publisher pages show you the papers that have since cited the one you are reading, and that list is the future of the conversation rather than its past.
It is how you find out whether a finding held up, whether somebody contradicted it, and whether a whole subfield grew out of it. It is also the most reliable way to find recent work on a narrow topic, because recency filters in a database are blunt while a citing list is precise.
And it is how you discover that the paper you were about to build an argument on was corrected or withdrawn, which is why this step and the retraction check are really the same two minutes of work done for two reasons.
04Knowing when to stop.
Chaining has a natural stopping point and recognising it is part of the technique. After a few rounds in both directions, the same names and the same handful of works start reappearing. That repetition is not a sign that you are searching badly. It is the sign that you have found the boundaries of the conversation.
At that point further chaining returns diminishing results and the useful move is to stop searching and start writing, which is a decision students postpone because searching feels safer than committing. Recognising saturation is what gives you permission.
The reverse signal matters too: if every round produces entirely new names with no overlap, your topic is probably broader than you think it is, and the answer is to narrow the question rather than to read more. That is the same diagnosis as a capstone topic that is too wide.
05A worked shape.
Find one relevant paper by any means, including an ordinary web search. Read it fast enough to confirm relevance. Take its subject headings and run those in a proper database, which usually produces better results than your original keywords.
Then take its reference list backwards for the foundations, and its citing list forwards for the current state. Note anything appearing in both directions, because that is the field's centre of gravity.
Repeat once, maybe twice. Somewhere in the second round the names begin repeating, and that is the whole search done, usually in under an hour and usually better than three hours of keyword variations. The mechanics of the database half sit at how a database really searches.
Questions at this desk.
Where do I find who cited a paper?
Most publisher pages show it, most academic databases have a cited-by link on the record, and general scholarly search tools display it prominently. If your usual database does not, that alone is a reason to look the paper up somewhere else, because the forwards direction is worth having.
Is this acceptable, or is it a shortcut?
It is standard practice and it is how researchers work. The thing that would be unacceptable is citing a source you found in a reference list without reading it, which is a different failure entirely and a visible one. Chaining tells you what to read; it does not replace reading it.
What if my one good paper is old?
Then the forwards direction matters more than usual and will be productive, because everything published since is downstream of it. An older paper with many citing works is an excellent starting point precisely because it sits at the root of a conversation you can then follow forward.
How do I find the first paper?
An ordinary keyword search is fine for this, and so is a web search, since the only requirement is relevance rather than quality. You can also ask an instructor for one starting reference, which is a small request most are glad to answer and which saves an afternoon.
Notes in the margin, from the people at the desks.
READ THE LETTER