When looking for a research topic, start with what recent studies suggest

You have a proposal due and no topic. Your supervisor says, "Read the recent papers." So you do, and near the end of most of them you find a paragraph that starts "Future research should…". It feels like a gift. Sometimes it is. Often it is the study's own weak spots, dressed up as advice. It pays to know which one you are holding before you give it a year of your life.

Many suggestions are limitations turned inside out

Three management researchers went through 1,276 journal articles and set each paper's limitations beside its suggestions for future research (Brutus et al., 2013). The two lists closely matched. The study looked at one moment in time? Then the next one should follow people over years. It covered one setting? Then try others. The researchers had a name for it: "limitations turned inside out". Only about one article in nine pointed future research towards new theory.

Health research has a different problem: suggestions too vague to use. Fewer than a third of 2,535 Cochrane reviews said who should be studied next (Clarke et al., 2007). When a team at Johns Hopkins dug the specific research gaps out of evidence reports, the reports' own suggestions turned out to be more general than the gaps they found (Robinson et al., 2011). And in one corner of computer science, future-work statements were "often unspecific or ambiguous", and later papers rarely took them up (Suray et al., 2024).

None of this makes the paragraph useless. The same researchers who coined "limitations turned inside out" wanted these sections to point to "immediate and incremental opportunities" (Brutus et al., 2013), and a small, careful next step can be exactly the right size for an Honours or master's project. It just needs more from you before it becomes your topic.

Three questions turn a suggestion into your topic

Who, exactly? Most suggestions do not say. Seven in ten Cochrane reviews did not (Clarke et al., 2007). Decide it yourself, and be ready to explain your choice. A topic that quietly means "everyone" will come straight back from your supervisor.

What would it teach us that the first study could not? This is the one that matters most. The easy move is "test it in South Africa". That can be a fine project, but only if you can finish this sentence: "Here, we might see something the original study could not, because…"

Here is an invented example. A reading programme was tested in schools that teach children in their home language, and the paper ends: "Future research should test the programme in other settings."

The quick version: "The reading programme in South African schools."

The version with a real question in it: "The programme assumes children keep learning in their home language. In many South African schools, lessons switch to English in Grade 4. Does the effect survive the switch?"

The second one says what South Africa could teach everyone else.

Has anyone done it since? Suggestions go stale. When a group of UK health research bodies proposed a standard way to write research recommendations, they made the date a required part, because it "served to define its shelf life" (Brown et al., 2006). Check the year the paper came out, then look for what has been published since. One Cochrane review in five already pointed to a relevant study under way (Clarke et al., 2007).

Two habits help. Read the discussion, not just the abstract: in six major medical journals, between 18 and 68 per cent of abstracts contained data that contradicted the article or was not in it at all (Pitkin et al., 1999). And copy down the authors' exact sentence. "The authors suggested X; I am asking Y because Z" is a far better way into a supervision meeting than "there is a gap in the literature".

Try it with one paper you have already read. Find its future-research paragraph and answer the three questions, one sentence each. If they come easily, you may have a topic. If one will not come, that is where your work starts.

Often it is the second question that opens things up, because it shows what the original authors took for granted, like the language children are taught in. That is the beginning of a question of your own. The next step is to ask who would decide differently if you were right.

Most of this evidence comes from management, health and computer science, and some of it is two decades old. Other fields may differ.

References

Brown, P., Brunnhuber, K., Chalkidou, K., Chalmers, I., Clarke, M., Fenton, M., Forbes, C., Glanville, J., Hicks, N. J., Moody, J., Twaddle, S., Timimi, H., & Young, P. (2006). How to formulate research recommendations. BMJ, 333(7572), 804–806. https://doi.org/10.1136/bmj.38987.492014.94

Brutus, S., Aguinis, H., & Wassmer, U. (2013). Self-reported limitations and future directions in scholarly reports: Analysis and recommendations. Journal of Management, 39(1), 48–75. https://doi.org/10.1177/0149206312455245

Clarke, L., Clarke, M., & Clarke, T. (2007). How useful are Cochrane reviews in identifying research needs? Journal of Health Services Research & Policy, 12(2), 101–103. https://doi.org/10.1258/135581907780279648

Pitkin, R. M., Branagan, M. A., & Burmeister, L. F. (1999). Accuracy of data in abstracts of published research articles. JAMA, 281(12), 1110–1111. https://doi.org/10.1001/jama.281.12.1110

Robinson, K. A., Saldanha, I. J., & McKoy, N. A. (2011). Development of a framework to identify research gaps from systematic reviews. Journal of Clinical Epidemiology, 64(12), 1325–1330. https://doi.org/10.1016/j.jclinepi.2011.06.009

Suray, J., Klemmer, J. H., Schmüser, J., & Fahl, S. (2024). How the future works at SOUPS: Analyzing future work statements and their impact on usable security and privacy research [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2405.20785