Writing effective titles for science papers

Screen Shot 2018-03-01 at 20.39.43How do you find a paper? How do you decide which abstracts to read? The answer is the title. Getting it right will help your work get noticed,  get read and ultimately get cited. Get it wrong and you might get your paper rejected or lose marks unnecessarily. Worse of all, a poor title might mean that the work you did does not get seen. Research that no-one knows about might as well not exist.

The advice on this page refers to primary data-style papers or for project reports in undergrad/masters programs. i.e. situations where you have collected new data. Titles for review articles, essays, dissertations, opinion pieces are a bit different as you have much more flexibility.

Styles of Titles

These are the four main types of titles, in order of effectiveness

  1. Statement (declarative)  “LaNts influence laminin organisation”
  2. Compound (statement+question or statement+implication) “LaNts influence laminin organisation: implications for wound repair”
  3. Descriptive “Investigation into LaNt effects on laminin organisation”
  4. Question (interrogative) Do LaNts influence laminin organisation?”

Statements work best as they are, by far, the most useful to the reader, it tells them exactly what they need to know without having to go any further. Your paper will get remembered and cited more frequently if you have a simple clear statement. People will see it in their reference library, remember what you actually found out and therefore choose to reference you.

In contrast, teaser style questions are really annoying. Clickbait might work for getting people to open the abstract but you’ll miss out on citations.

  • Shorter and more general statements work best You need to be specific and not overreach but aim for as direct as to the point as possible.
  • If you need subclauses, put the primary finding clause first. “LaNts regulate laminin deposition by corneal epithelial cells” is better than “In corneal epithelial cells, LaNts regulate laminin deposition”
  • Active voice sounds better than passive “LaNts regulate laminins” rather than “Laminins are regulated by LaNts”
  • One clear message is better than two less clear ones; focus on the key finding
  • Think big picture; (within reason) write your title in a way that will appeal not only to the target journal but also to a wider audience


Start writing your manuscript by coming up with a draft title, then edit later

The rest of your writing depends on what story you want to tell. Therefore, before you do anything else, decide what you consider to be the key takeaway message from your paper to be and write a draft title that  states that message as simply as possible.

Note that at this point you don’t need to worry about the precise wording too much, you are will adjust that in the editing stage later  but it will help the rest of your writing to be on message from the start. Your figure order and therefore your results, your introduction and discussion all depend on what the key story is. If you are genuinely not in a position to write even a draft title you can get your methods written and make some individual data panels for your figures, you’ll struggle writing the rest until you have an idea where you are going!

Seriously, draft a title for your paper early!

Having a draft title in place for a manuscript where you have some of the main story elements but aren’t quite complete is a good way to focus your next set of experiments. Your draft title may need to change to reflect the next set of data but having one in place will help you going forward. Being clear about what you are working toward will help you identify what proofs are required.

With a draft title in place you can  go ahead and make your figures and write the results, intro and discussion sections (click the links to see our guides). Come back here again when those sections are drafted!


Writing a student paper… state the answer!

Screen Shot 2018-03-01 at 20.47.41I know this is repetition but I really want it to sink in! Definitive results statements will get you better marks. 

In a student paper (e.g. project report) your work is going to get marked. If you use a question or descriptive title then the first thing that your marker will think is that you don’t know what your data means!! Even if the rest of your work is super clear, the very first impression that they will have of your work will be negative.

What about journal articles? What works?

There is quite a bit of published literature on what titles work, in terms of downloads and citations.some of which is subject specific. Here are some trends:

  • Papers whose titles emphasize broader conceptual or comparative issues fare better both pre- and post-publication than papers with organism-specific titles (ref)
  • Articles with question titles tend to be downloaded more but cited less (ref)
  • Journals that use shorter titles tend to receive more citations (ref)
  • Articles with longer titles are downloaded slightly less than the articles with shorter titles (ref)
  • Titles with a colon tend to receive fewer downloads and citations (and be longer) than those without (ref)


But, use the right title for your story!

Be aware that the correlations for each of the findings above aren’t super tight and there are quite a lot of caveats to the interpretation. Moreover, like all stats, these data tell you about the population, they don’t say definitively what will work for your story, don’t try to shoehorn your title into a framework based solely on these findings. Full disclosure, a  paper of mine that is about to come out that has a descriptive title. I tried to reshape it into a statement a bunch of different ways but I couldn’t come up with something that effectively captured the message in a succinct and effective way.

Don’t undersell but also don’t overstate your findings!

If you legitimately have cured cancer then you should shout it from the rooftops. However, if you haven’t then you shouldn’t say you have in your title! Extreme example, but the concept stays true; your title should capture the biggest finding of your data but only go as far as your data can support.

Scientists need proof. A reviewer/editor/marker should spot the overstatement and should ask you to prove any aspect that you have stated but not supported.

Screen Shot 2018-03-01 at 21.36.35

Generalisability. A specific form of overstating to be aware of. If your work has been performed in a limited capacity and you don’t yet know if the findings translate to other systems then you do need to indicate the limitation.

For example; “LaNts regulate cell migration” sounds bold and dramatic but if I’ve only checked this in corneal epithelial cells then it would be inappropriate to imply the effects are general to all cell migration.

Presenting implications as concrete findings. Again a specific type of overstating, if your data shows that treatment with drug A decreases protein levels of B and also that drug A changes cell behaviour C but you haven’t shown that C depends on B then you shouldn’t claim that is does in your title! Sounds obvious, but it happens a lot.

Danger words to look out for:

via – have you really connected the two findings? “mir 485 regulates cell migration via downregulating pax6” – unless you have shown conditions where mir 485 treatment was unable to regulate pax6 but could do all its other stuff then via is risky.

to – are you implying a deliberate mechanism (watch for anthropomorphism; assigning character traits to things). Using just “to” as shorthand for “a pathway has evolved to” or equivalent, is quite common but there might be more elegant ways to make that point (simply switching to with which might do the trick)

Write lots of versions of your title

Final tip – try out lots of different versions of your title. Get people who are unfamiliar with the data to look at your different versions to select the one that is clearest in its message. Get people who are familiar to check that you are stating the findings appropriately, not over or understating them and not (accidentally) used a question!

Really, don’t use a question!

Return to the rest of our writing guides

%d bloggers like this: