Why writers hoard research (and when to stop)

Every so often a manuscript arrives with a research folder attached that’s several times the length of the piece itself: interviews transcribed in full, a dozen papers annotated in the margins, whole books’ worth of notes distilled down to a paragraph or two that actually made it onto the page.

The writer usually presents this as diligence, and some of it is. But past a certain point, more research stops being preparation for the writing and starts being a way of not doing it.

The folder that keeps growing is doing a job

Piers Steel’s review of several decades of procrastination research is the most thorough attempt I know of to work out what delay is actually for, and one of his conclusions is useful here: procrastination correlates most strongly not with laziness but with task aversiveness — how unpleasant or threatening the task feels in the moment.

Writing the argument down is where a claim can be shown to be weak, a structure can collapse, or a deadline can be missed in full view. Reading one more source rarely carries that risk. It has the shape of work, produces a stack of highlighted pages, and defers the moment where the writing itself is judged.

The folder keeps growing because growing it is comfortable, and the piece it’s meant to support stays unfinished because finishing it is not.

It solves today’s problem at tomorrow’s expense

Fuschia Sirois and Timothy Pychyl’s account of procrastination frames it as a matter of priorities rather than a character flaw: when a task produces an unpleasant feeling — doubt about whether the argument holds, anxiety about how it will be received — people tend to prioritise repairing that feeling immediately over making progress on the task that caused it, even when they know the deferral will cost them later. Opening one more source and finding it says something useful repairs the feeling of not being ready. It doesn’t touch the actual problem, which is a draft that still needs to be written, and their research is clear that the relief is short-lived: the discomfort returns, usually worse, closer to the deadline the extra reading was meant to prepare for.

None of this makes over-researching identical to more ordinary forms of procrastination, like checking email instead of writing. It’s a more respectable version, because it looks like exactly the kind of care a serious writer is supposed to take. That’s what makes it harder to catch in yourself than in someone else’s manuscript.

Not every research pass is the same pass

None of this is an argument against research itself, and I want to be careful not to overstate the case. Early in a project, before the argument has taken shape, wide reading is exactly the right instinct — you don’t yet know which sources matter, so casting a broad net is legitimate work, not avoidance of it. The pattern I’m describing shows up later, once a writer has a draft, or at least a clear argument, sitting unfinished on the desk. At that stage the research has already done its job of shaping the piece. Continuing to research from that point forward is a different activity wearing the same clothes, and the giveaway is usually where the writer’s attention has moved: away from the gaps in the draft and onto whatever is easiest to keep reading about.

There is no natural stopping point in the material itself

Herbert Simon’s work on decision-making under real constraints, rather than the idealised kind economists like to assume, gives the clearest account I know of why “just a bit more research” never resolves on its own.

Simon argued that organisms with limited time and limited capacity for weighing every option don’t optimise; they satisfice — they search until they reach a solution that clears a threshold they’ve set in advance, then stop, rather than continuing to search for something better. The material itself supplies no such threshold. There is always another study, another expert, another angle nobody has covered yet.

Without a stopping rule decided ahead of time, research doesn’t run out on its own; it simply continues until something external — a deadline, an editor, exhaustion — cuts it off first.

That reframes the question usefully. The problem was never how much research is enough in some absolute sense. It’s that most writers never set the threshold at all, and a search with no threshold has no reason to stop.

What a threshold looks like in practice

I spent the first part of my working life in front of a classroom rather than a manuscript, and one habit from that work has stayed with me: a lesson plan has to specify, in advance, what “prepared enough” looks like, because the bell doesn’t wait for you to feel ready.

Editing has taught me the same discipline applies to research. Before I’ll accept that a manuscript needs another source, I ask what specific gap that source is meant to close — a claim that’s currently unsupported, a counterargument that hasn’t been addressed, a fact that needs checking.

If the honest answer is that the source would be interesting rather than necessary, that’s the clue the research pass has become the more comfortable alternative to the writing pass. A writer who sets that threshold before starting — three sources on the central claim, one strong counterargument addressed, and no more — has a stopping rule. A writer without one is relying on a feeling of readiness that, as the research above suggests, has no reliable relationship to how much work is actually left to do.

The manuscript doesn’t know how much you read

A reader encountering the finished piece has no way of seeing the size of the folder behind it, and no reason to care. What they can see is whether the argument holds, whether the claims are supported, and whether the piece finishes what it started.

Past the point where research is closing a specific, identifiable gap, more of it adds weight to a folder nobody but the writer will ever open, while the actual manuscript — the thing that has to survive contact with a reader — sits exactly where it was before the extra reading began.