Abstraction
How to Write a Scientific Abstract
Welcome to a dry and boring Shelf Life about a topic no one cares about, but wait—I told my very cool friend, “I think I’m going to do a Shelf Life on how to write a great scientific abstract?” and she was like, “Yes definitely do it” and she seemed excited, so now I’m questioning everything. Perhaps all this time I was holding onto the boring abstract topic for a rainy day when I didn’t have much to write about, but really this topic was a banger all along? Is it possible I’m not as good a judge of interesting topics after all? No, surely it’s everyone else who are wrong.
Do I write a lot of scientific abstracts? No I do not. My qualifications for writing this article are I publish a lot of scientific abstracts; that is, in the course of my work, I edit, format, and produce a lot of scientific abstracts—thousands each year—abstracts that are attached to scientific journal articles and abstracts that stand alone as their own publication, usually as part of meeting proceedings (more on these in a bit).
I cannot tell you anything about doing science but I can tell you a lot about preparing science for publication, and surely that’s almost as good.
An abstract is a brief summary of a longer work on a specific subject; that longer work may be:
A research article;
A dissertation or thesis;
A white paper;
A live presentation given, or poster displayed, at a conference; or
Any other in-depth analysis of a specific topic—a scientific topic, specifically, for our purposes today.
The purpose of the abstract is to give readers (or, otherwise, users) a quick overview of the full contents of the material being abstracted in case that reader (user) either:
Does not have time to read (or watch or listen to) the full work but wishes to get an overview of the findings;
Wants to quickly assess whether the full work is germane to their (the reader’s/user’s) interest and, therefore, worthy of reading in full; or
Does not have access to the full work; for instance, if the work was presented at a conference the user missed and they wish to get a summary of what was presented or displayed.
A positive thing to have come out of the last few years of restricted travel (and so on) is that many conferences now offer hybrid attendance options—attendees may participate in person or virtually—and sometimes also asynchronous options (users can watch recorded sessions after the conference has concluded).
I’ll go ahead and say I’m a person who works for a company who hosts large conferences and I’m also a person who sometimes attends conferences in my own field (editing, publishing), so I’ve experienced the reaction to conferences going virtual/hybrid/asynchronous from several angles. I realize that some folks regret the move to fewer co-located/synchronous attendees because they miss the sidebars, the hallway conversations, the hugs, the coffee breaks, the dinners—and I understand that. But I also have found that I, personally, can attend many more conferences and professional development opportunities now that I don’t have to physically travel to all of them—my budget stretches more if I don’t have to pay for flights, hotels, meals away. Further, more of my colleagues (and more of my junior) colleagues can attend these events as the budget stretches.
Putting budget aside, think about who is most likely to not be able to travel to an onsite conference for other reasons:
People with child- or eldercare responsibilities (often women);
People with chronic illnesses and disabilities;
People in low- and middle-income countries where conferences are less likely to be held, who must travel further and for longer.
All this to say I think it’s a good thing there are more hybrid options. And now that there are more of these hybrid options, you’re more likely to be able to watch a lecture or presentation post-facto, or see a poster online without having to go to the physical poster session. This is great. But abstracts are still really important.
So: Let’s get down to business. To abstract. Our stuff. (I had to explain this Mulan [1998] joke to my partner so if you don’t get the reference consider yourself explained.)
What Makes a Good Abstract
A good abstract, like all good writing says everything it needs to say and nothing it doesn’t. Don’t worry I’m going to explain. Obviously the above was hopelessly vague. A good abstract is:
Concise—most professors and publishers will give writers a strict word limit;
Complete—all the important aspects of the research must be included; and
Self-Contained—the abstract must stand alone and not reference other materials.
A bit more on each.
Concise
Whomever the abstract is submitted to will provide a word count limit for the abstract. The body you’re submitting to might be, for instance, a journal, a conference, or a professor who has assigned a research paper—that body should (but won’t always) provide clear instructions on how long your abstract may be. If they specify a length (word count), do not exceed that. Many manuscript processing systems and conference submitters are automated and will kick your submission back at you if the abstract word count is too long. More on abstract length later—
The main thing is, the abstract should omit anything that isn’t essential for the reader to know about the research, methodology, and findings. Limit the content of the abstract to only the important information; for instance, consider two studies of 500 patients each:
In a study that compares rates of a specific cancer in patients of different races (eg, what is the rate of this cancer in Black versus White versus Asian patients?), then it is important for the reader to know that 170 patients were Black, 140 patients were White, and 190 patients were Asian. The ages of the patients may not be important.
In a study that looks at the rate of a specific cancer in patients of different age groups (eg, what is the rate of this cancer in patients under 40, in patients 40 to 49, and in patients 50 and over?), then it is important for the reader to know that 170 patients were younger than 40, 140 were 40 to 49, and 190 were 50 or older. The racial identities of the patients may not be important.
In the full paper (or presentation), there will likely be a table of patient characteristics that includes all this information—ages, racial identities, gender identity and sex characteristics, cancer stage at diagnosis, and so on. But the abstract needn’t contain all that—only what is critical to understanding the research.
Complete
Conversely, the abstract should contain every important highlight of the research, methodology, and findings so that someone who reads the entire abstract (and who has a sufficient level of scientific literacy) can grasp the whole study without reading the full paper or watching the presentation. This doesn’t mean the abstract must include every detail, but it must include all the salient aspects. For original research, this typically means:
What was the purpose of the study?
What methodology was used to conduct the research?
What were the findings of the research?
What do those findings mean?
That’s an awful lot to fit into 100 to 300 words; I know. That’s why it’s so important to be concise and choose your inclusions and omissions wisely. You can do it. I believe in you.
Self-Contained
Here are a couple of things an abstract should not include.
See Table 1 in the Methods section.
. . . as found by Smith and colleagues in their 1991 study.1
In these examples, the text of the abstract asks the reader to reference the full paper (example 1) or another paper entirely (example 2). An abstract should never direct a reader to another source for the information they need to understand the research being abstracted; and especially should not direct the reader to another part of the full paper—remember, a key function of the abstract is to inform those who do not have access to the full paper!
An abstract will not typically be allowed to include its own reference list. However, it is usually okay to reference materials that are widely known and available and do not need to be looked up to be understood by the target audience. By this I mean, for instance, landmark studies in the field. (A landmark study is one that delivers highly important and visible breakthrough science and are known by name by many or most people in the specific field.) Other items of this nature might include census data (a writer could mention “the 2020 U.S. census” without having to include a reference); or the SEER Registry (eg, “we completed a retrospective analysis of 1,234 cases found in the SEER Registry”).
Specs, Tips, and Tricks
IMRaD
Authors of research papers may be asked to provide a “structured abstract” or “unstructured abstract.” An unstructured abstract is just a paragraph summarizing the research. “Structured abstract” usually means the author should follow the IMRaD model, which stands for:
Introduction
Methods (or Materials and Methods, or Patients and Methods)
Results
and Discussion (or Conclusion, or Findings)
If you’re asked to provide a structured abstract and the headings for the structure are not provided, the ones above are the ones you should use.
The introduction should be one or two sentences that mainly explain the purpose of the research (why did we undertake this research in the first place?) and may also introduce the current state of the field as it applies to the topic.
The methods section should explain the where, when, how, and what or whom of the research. Where was the research conducted, when was it conducted, what methodology was used (how was it conducted), and the materials (what) or patients (whom) it was conducted on.
The results section should summarize the findings of the research; what did the methodology turn up? This section should include as much data as you have available and can fit.
The discussion section should contextualize the results. What do the findings mean, and for whom? Do your findings support, extend, or contradict prior research? And, in light of the findings, what research is now indicated to learn more?
Abstract Length
If you’re responding to any call for submissions or journal that is open to submissions, this information should be easy to find. For instance, I was rapidly able to find that Nature wants an abstract not longer than 200 words, PLOS will take an abstract up to 300 words, Science requests abstracts of 125 words or fewer, and American Economic Review will take an abstract of 100 words or fewer. However, I also found that Cell wants an abstract but doesn’t say how long it should be, and The Lancet will take an abstract “not exceeding 300 words” but I had to click through several menus and open a PDF (shrieking with rage while Acrobat Reader fired up) to find this information. It’s not always as easy to find as it should be; but usually the information is there.
Because words in the abstract are counted automatically by the manuscript processing system, don’t count on technicalities (eg, “I was exactly at the word count limit before putting in the IMRaD headings, do they count?” Yes of course they count.)
If you cannot find guidance from a specific journal (looking at you, Cell), consult the style manual most likely to be used in the field. For instance, American Medical Association (AMA) style specifies abstracts should not exceed 400 words while American Psychological Association (APA) style (for the social sciences) suggests 150 to 250 words.
Title
This goes for all publications and not just abstracts, but if you are publishing an abstract not attached to a journal article (for instance, as proceedings of a meeting) and you must give it a title, a best practice is to not begin your title with an article (“a,” “an,” or “the”) and instead skip to the first meaningful word; and never to start your title with a symbol that is not a letter if you can possibly help it—for instance @, #, “, ‘, or $.
Some examples:
No: $50,000 per patient per year spent on lip balm in the hospital setting
Yes: Lip balm spending in the hospital settingNo: “A Room With a View”: Effect of hospital room décor on patient outcomes
Yes: Effect of hospital room décor on patient outcomesNo: An examination of ibuprofen ± acetominophen for stubbed-toe pain
Yes: Examination of ibuprofen ± acetominophen for stubbed-toe painNo: #Hashtag use in social media
Yes: Hashtag use in social media
It’s cute when you come up with it but what’s not cute is when the third-party aggregators choke on your abstract and delay dissemination of your research or your abstract publishes with garbage characters in the title (#Hashtag use in social media) and it takes six to eight weeks for someone to fix.
Language Use
As with pretty much all science writing, the passive voice should be used (see what I did there?). Use the same style and voice you will use for your full paper, and resist the urge to save words by using less-formal language (eg, avoid “I,” “we,” “our”).
Don’t use trade names for drugs, chemicals, or devices if a generic is available. For instance:
No: “Methods: We caffeinated our team using 3 cups of coffee per scientist per day brewed in a Keurig”
Yes: “Methods: We caffeinated our team using 3 cups of coffee per scientist per day brewed in a capsule coffee machine”
Haha just kidding scientists need much more caffeine than that.
If the subjects of your research are people, use patient-sensitive language. For instance, patients are people and cannot be “randomized among two arms”; they are, rather, “randomly assigned to one of two arms.” The term cases should also not be used interchangeably with patients (this doesn’t mean cases can never be used; sometimes one patient presents two or more cases). Use patient-first constructions, like “patients with heart disease” rather than “heart disease patients.” (There’s a neat article you can read in PubMed Central by Crocker and Smith on person-sensitive language for more on this.)
If any scientists are out there reading and found this useful please blind me with your science, I love it.
If you have questions that you'd like to see answered in Shelf Life, ideas for topics that you'd like to explore, or feedback on the newsletter, please feel free to contact me. I would love to hear from you.
For more information about who I am, what I do, and, most important, what my dog looks like, please visit my website.
After you have read a few posts, if you find that you're enjoying Shelf Life, please recommend it to your word-oriented friends.