Your topic isn't wrong. It's still three topics in a trench coat.

Say your topic out loud, in one sentence. If you need a second "and" to finish it, that's the tell — not that the topic's wrong, but that it's still three topics standing in a trench coat.

That's usually where it started — not as one topic, but as an interest wide enough to hold three or four of them at once. Somewhere between the interest and the page, it's supposed to narrow twice: once into a topic, and again into a question. Most "topic trouble" is really a stall at the first narrowing, mistaken for the finish line.

How do you narrow down a dissertation topic in two steps?

Here's what that looks like written out, not just described.

Draft one — the interest. One sentence, no jargon, the kind of thing you'd say to someone outside your field. "I'm interested in how small nonprofits handle a founder leaving."

Draft two — the topic. Add a population or a context — enough to narrow the interest without making a claim yet. "Leadership transitions in small nonprofits after a founding executive director departs." This is usually where people stop, hand it to their advisor, and wait for it to be accepted as finished. It isn't finished. It's a label for a corner of the literature — accurate, but not yet answerable.

Draft three — the question. Add a relationship, a comparison, or a gap — something with an actual answer on the other side of it. "Does having a documented succession plan predict organizational stability in small nonprofits during the two years after a founder's departure?"

The second narrowing is the one that gets skipped. A topic can survive for months as a label before anyone notices it was never turned into a claim.

Why do two different topic problems feel identical?

Too broad and too narrow don't feel different from the inside. Both produce the same thing: a flat refusal to explain your topic in under thirty seconds without losing the thread. You can't tell them apart by the stuck feeling alone. You need an outside test.

Too broad still has the trench coat on. Say it out loud — if it needs that second "and" to hold together, it's still doing an interest's job, not a topic's.

Too narrow is the opposite failure, and it's easy to mistake for boldness. A real literature search turns up almost nothing nearby — not "my exact angle is under-covered" (normal, often a good sign), but no surrounding conversation at all. A question with nothing to argue with isn't an original contribution. It's usually been narrowed past the point where a field can recognize it.

Most of the time it's neither — it's a draft that stopped one narrowing too early, or went one narrowing too far.

What search actually tells you whether you're too broad or too narrow?

"Do a search and see what comes up" isn't specific enough to trust when you're the one who's stuck — so here's the actual version I walk candidates through.

Take your draft-two topic (not draft three yet — the labeled version, before you've forced a claim) and run it through your database of choice using just the core terms, no added qualifiers. Pull the first twenty results that are actually relevant, not just returned. You can use Google Scholar, JSTOR, ProQuest Dissertations & Theses, or your university library's database.

Now sort what you find into three piles: studies that ask close to your question already, studies that sit one step away — same population or same variable, different angle — and studies that only share a keyword with you but nothing else.

If most of what you found lands in the first pile, the question's likely been asked. Not disqualifying on its own, but it means your draft three needs to say specifically what you're doing differently — a different population, a different method, a gap the existing answers left open.

If most of it lands in the second pile, that's the healthy result. A real conversation exists nearby, with room in it for your specific angle. This is what "under-covered, not uncovered" looks like in practice.

If almost everything lands in the third pile — a handful of keyword matches and nothing that engages your actual question — that's the narrowed-too-far signal, not evidence you've found untouched ground. A field that can't recognize your question yet is usually a sign the question needs to widen back toward a conversation that already exists, not proof you're the first person clever enough to ask it.

A gap in the literature and a gap in your search terms look identical from where you're sitting. The three-pile sort is what tells them apart. (If the search itself feels shaky rather than just the topic, building a lit review with its own thesis makes this whole process easier.)

What does this look like with a real pile of ideas?

One candidate I worked with had been talking his topic through with ChatGPT since the start of his degree — not to write anything, just to think out loud with, on and off, for months. By the time he sat down with me, he had somewhere north of a dozen "directions." He couldn't tell me which one was his topic, because none of them had ever really stopped being an interest.

I didn't pick a favorite out of the dozen. I took the one he kept returning to and ran it through the three drafts. As an interest, it was broad enough to fit half the pile inside it. As a topic, it narrowed to something nameable — but he still couldn't say it in one sentence without an "and" pulling in a second variable that belonged to a different one of his twelve ideas. That was the tell: not that the direction was wrong, just that two of his twelve ideas had never actually been separate — they were one topic that hadn't finished narrowing.

Splitting them, then running the smaller one through the search check, took about forty minutes. Draft three came out of that same session. The other eleven ideas didn't get discarded so much as quietly absorbed — most of them turned out to be the same handful of variables, recombined.

What do you do when your committee disagrees with your scope?

This is the part almost nobody writes about directly, and it's one of the more common reasons a candidate concludes the topic itself is doomed: one reader says narrow further, another says it's fine as-is, and your own instinct is doing a third, different thing entirely.

That's not a sign the topic is unsalvageable. It's a scope disagreement, and scope disagreements get resolved by asking a specific question — not by quietly picking a side and hoping nobody notices, and not by trying to satisfy both notes at once and ending up with a topic that's blurry in a new way. (This is the same underlying skill as reading vague advisor feedback — turning an ambiguous note into a specific, answerable question.)

That does two things at once. It names the disagreement as a normal committee-coordination problem instead of a verdict on your judgment. And it puts the resolution back where it actually belongs — with the people who can settle it — instead of leaving you to guess and redraft in the dark for another two weeks.

There's a quieter version of this same problem, and it's arguably more common than open disagreement: nobody has said anything at all. No conflicting notes, no pushback — just silence on scope, while your own uncertainty about whether it's too broad or too narrow sits there unresolved. It's tempting to read that silence as approval. Sometimes it is. Just as often, it means scope hasn't actually been reviewed yet — your advisor is waiting to weigh in until the question is further along, or assumed you'd flag it if you were unsure.

Don't wait for that to resolve itself. The same move works here as with open disagreement, just addressed to one person instead of two: name the specific uncertainty, and ask for it directly. (If the silence extends beyond just scope feedback, here's what to do when your advisor stops responding altogether.)

One sentence, one direct question, answerable in one reply. It converts a vague worry you're carrying alone into a concrete question someone else can actually resolve — which is the same move underneath both versions of this problem: uncertainty about scope isn't yours to sit with silently, whether it showed up as two conflicting notes or as no notes at all.

What's a five-minute check before you touch the topic again?

Before you conclude anything about the topic, run this:

Say it in one sentence. Count the "and"s. Two or more — still a trench coat.

Spend fifteen minutes searching your closest sources. Found some, but not your exact angle? Healthy. Found almost nothing? Likely narrowed too far.

Try writing draft three — the version with an actual claim in it. Stuck there, not at draft two? That's not the topic failing. That's exactly where the work is.

FAQ

Final thoughts

None of this requires a new topic. Learning how to narrow down a dissertation topic isn't about finding a better idea somewhere out there — it's about finishing the draft you're already standing in the middle of.

Say it out loud, count the "and"s, run the search, and see which of the two narrowings you actually stalled on. Most of the time, the topic was never the problem. It just hadn't been asked yet.