Scientists are taught how to to plan their experiments, but not their reading. And if you don’t read the literature, how can you ask the right questions or do the proper experiments in your chosen field?
Here’s some tips for getting started.
With the volume of published scientific data increasing at an exponential rate, keeping on top of the literature has ceased to be something that can be done casually. It now requires serious attention and a structured approach. Unlike 2-3 scientific generations back, there is now too much to read even within a single research field (and that’s just the peer-reviewed, published part – preprints are increasing the volume even more), let alone broader categorisations. Here’s TIR’s short guide to getting on top of things.
Part 1: Getting the papers
First, you need to assemble your reading “to do” list. That means obtaining current papers relevant to your topic, and the classic/keystone papers from that area.
– (1) First, sign up to receive eTOCS (electronic tables of contents) from all reputable – yes, all, but strong emphasis on the “reputable”! – general and specialist journals relevant to your project. This is your unbiased screen for papers. It will undoubtedly increase your inbox traffic, but quickly scanning eTOCs is a good way of catching papers that are slightly outside your exact area of interest, and gives you a broad perspective on the activity of your field. Don’t forget to sign up for preprint alerts from bioRxiv as well. In a sense, eTOCs are like a window onto the mind of the scientific world. In time, you’ll get used to skimming these and letting title keywords jump out at you. If you’re not sure what to sign up for, ask your mentor(s) and colleagues what journals they rate. You don’t need to worry about missing anything using this approach because you will also…
– (2) Set up a battery of PubMed alerts. This is the candidate screen of papers, and will make sure you don’t miss anything directly relevant to your work. Set up alerts for all group leaders in your field (using the [au] search filter) using the “save search” function, and get results delivered weekly. Also set up alerts for specific searches that are relevant to your topic. Again, this will increase your inbox traffic but those two points will together ensure that almost nothing will escape your attention. If you’re unsure how to use search filters in PubMed, ask your colleagues or leave a comment here.
– (3) Download pdfs of all papers that you think you need to read and put them into a folder marked “reading”. They stay there until you’ve read them properly. Collecting all unread papers in this way also gives you a sense of how far behind you are in your reading, and lets you allocate time to it specifically if need be.
– (4) Name the pdfs in a way that they automatically arrange in date order, e.g. “2019_4.22_Golgi segregation” or “2019.4.22_Morriswood”. This ensures that you always have the oldest pdfs at the top of the pile, and also helps in later archiving.
– (5) Read the magazine sections of Nature and Science every week, or as often as possible. Nature and Science are laughably poor vehicles for the dissemination of research data due to figure and wordcount compression, but their magazine sections (which are editorially independent of the research sections) are essential reading. The editorials, news articles, features, world views, book reviews and so on are a great way of getting a perspective on the world of science as a whole.
Part 2: Reading the papers
Once you’ve assembled your “reading” folder full of pdfs, you now need to consume its contents. The key thing with reading is that you need to ensure you can rapidly retrieve the important information from it if you come back to it. Reading a paper and not remembering anything about it is a wasted investment of time (and your time is limited).
– (1) Use the “highlight” function to mark key sentences as you go along. This
will ensure that the important sentences jump out at you when you return to it. Something like 100-250 highlights per pdf is a decent target.
– (2) Use the “comment” function to note any impressions you have as you go along. This is especially important if you disagree with the authors’ interpretation, or have an opinion on the data quality.
– (3) When you’ve finished, put a text box at the top of the pdf above the title. Summarise the main findings and your overall impressions in 2-5 sentences. This is ESSENTIAL. This text box will let you rapidly retrieve your evaluation of the paper irrespective of time elapsed since you read it. Plus, the summary can be updated in light of later work (e.g. “This was later superseded by…”, “This was largely refuted by…”).
– (4) If your reading time is limited, or your rate of reading is slow (very possible if English is not your first language), then you need to be extra choosy how you commit. Just read the title, abstract, then figures and figure legends. If the data look relevant to your work and the abstract connects them to your work, then commit to reading the whole manuscript. If not, write your text box summary at that point and stop there, but note that you haven’t yet read it all the way through carefully. Reading intensity should be proportional to the relevance of the work – something from your own or a competitor’s group needs close scrutiny; something of tangential relevance or that uses techniques you can’t easily critique or follow only needs light acquaintance.
Part 3: Organising the papers
Once you’ve read a paper, it can be moved out of your “reading” folder and into a new home.
– (1) Set up a “literature” or “references” folder to hold all your pdfs.
– (2) Create subfolders by topic (e.g. “Golgi”, “cytoskeleton”, “evolution”, “general interest”), and move your pdf into the one that seems most relevant. If you’ve named the pdf using a date (see above), then the pdfs will automatically arrange into chronological order. You don’t need to stress too much about which folder is most appropriate, as your computer’s search function will help you find papers later. Having them archived in this way though helps you do faster and more targeted searches if you need to find appropriate citations or quick manual searches.
Part 4: Using the papers
– (1) In your presentations, always provide a citation with any figure taken from a paper. This is not just a courtesy (and an essential one) to your scientific peers. The figures that you use to explain the background to a topic (whether general or specific) in either your research or your teaching are a distillation of what’s most important, so you need to be able to rapidly go back to those papers if need be.
– (2) In your papers (or in your thesis), always cite primary literature rather than reviews. Reviews should only be cited for general introductions to a topic (i.e. a bit like citing a textbook) rather than anything specific, and over-citation of reviews is often a giveaway that a writer is not really on top of the primary literature. Once you’re on top of your reading, you’ll find that it gets easier and easier to refer back to the original papers in your citations.
Part 5: Reaching back in time
– (1) If you manage to empty your “reading” folder, then you need to start going back in time. Start assembling all the older papers from your field, or the important ones that describe key observations, development of key reagents, and so on. It is essential to read as many of these source papers as possible in order to get a proper perspective on the field.
– (2) When you’re preparing a paper or a thesis, take the time to track down the original primary literature for all the important past observations – you only have to find them once! Once you know what the correct original citation for something is, you can use that for ever after. Put that pdf into your topic folder and, when possible, move it into your active “reading” folder so that it gets read properly. One big advantage is that most older papers are a lot shorter than modern ones, so they can be consumed quite quickly.
Part 6: Finding the time
Time is obviously limited for all scientists these days, and it’s easy to fall behind or even stop regular reading. This is a mistake, and will seriously hamper your development of scientific independence.
– (1) Set aside regular time on a daily basis for reading – a coffee break is a good opportunity. Make this of defined duration, so that you can clock in a minimum amount every week. Even 30min a day will yield real benefits, and there are plenty of 30min incubations out there with the available downtime…
– (2) Set yourself a rough target for paper consumption. One paper per working week of the year is a reasonable starting point.
– (3) At the start of your project, consider spending the first 2 weeks doing nothing but reading – every major paper on the topic for the last 5-10 years is an ambitious but achievable target. This is your best and possibly only chance to establish a firm theoretical foundation before the data starts flowing, and will be invaluable later on.
Part 7: Don’t underestimate the importance
– (1) Collecting or citing papers without reading them is like going to an art gallery just to pose in front of the paintings. If you have serious ambitions, or ambitions to be taken seriously, then you need to be on top of your field’s work both past and present. The high volume of misquoted references in the literature (particularly in reviews) is a sign that people have only skimmed important papers. If you cite it, read it.
– (2) Far too many students only start reading properly once they have to write the introduction to their thesis. This leaves far too much ground to be covered in too short a time, and contributes to errors. Getting on top of the literature is an essential step towards gaining independence, will stimulate new ideas, and maybe stop you overcommitting to unprofitable ones.
– (3) There are nowadays plenty of software tools (and online communities) out there for recommending pdfs, organsing pdfs, and managing references, but remember that these tools cannot substitute for your own impressions. Ultimately, you have to do the reading yourself, because it’s your opinion of things that counts. You (yes, YOU) need first-hand critical impressions of the keystone and current papers in your field.
This posting suggested by Graham Warren (LMCB, UK) – many thanks!