How to find a research gap you can defend
A research gap is only useful if a reviewer agrees it exists. "Nobody has studied X" is easy to write and easy to refute. A defensible gap is one you can point to in the literature: papers that explicitly stop short of something, or a combination nobody has tested. Here is how to find one and show that it is real.
Start from limitations and future work
The most reliable gaps are the ones authors point out themselves. Read the limitations, discussion and future-work sections of the ten most relevant recent papers, and copy out every sentence that names something they did not do. A gap that several papers mention independently is usually real.
Map what has been tested
Make a simple table: one row per paper, columns for the problem setting, the method, the data and the evaluation metric. Gaps show up as empty combinations: a method tested only on small models, a metric nobody reports, a setting that every paper assumes away.
Look for untested combinations and assumptions
- A method from one field that has not been applied to a closely related problem in another.
- A common simplifying assumption (static conditions, a single user, clean data) that real use breaks.
- Results reported on one kind of data that nobody has checked on another.
- Conflicting results between papers that nobody has reconciled.
Back the gap with evidence
Write the gap as one sentence, then attach two or three quotes from published papers that show the boundary of current work. Quote the text exactly and cite the page or section. A reviewer can argue with your interpretation; they cannot argue with what a paper says about itself.
Check that the gap matters
Not every gap is worth filling. Ask what would change if it were filled: a decision someone could make, a result that would no longer hold, a method that would become usable. If the answer is "nothing much", the gap will not carry a paper.
Questions
How many papers do I need to read to find a gap?
Enough to be confident you have the closest recent work: usually the 20 to 40 most relevant papers, with the limitations sections of the closest ten read carefully.
Is "no one has done X" a research gap?
Only if you can show it, and only if X matters. Support it with quotes from the closest papers that show where their work stops.
What if my gap was filled while I was working on it?
Re-run the search before writing up. If someone published it, compare carefully: a difference in setting or evaluation can still be a contribution.