How to Read Papers Fast
Chris Lee, M298
Winter Quarter, 2006
Scientists are busy, busy people. They don't have time to spend hours on every paper they need to know
about. It's essential to learn how to read a paper quickly but insightfully, because otherwise you'll become
lazy and uncritical. Most papers, you'll probably read quickly as outlined below. On papers of great
interest to your work, you can then drill into an in-depth review of the whole paper.
Amazing but true fact: this kind of speed-reading is how most of your critical papers and grant proposals
get reviewed. Often the reviewer only gets a chance to read piles of manuscripts during a plane flight or an
hour before a meeting. Though you may have spent months writing, the reviewer may only spend fifteen
minutes to decide your fate! Thus, learning this "speed-reading" process is important for learning how to
write successful grants and papers.

General Hints
When you read, highlight key nouns and verbs so that the highlighted words form "sentence
fragments" that state the key points. This forces your brain to zero in on the key nouns / verbs, and
assemble them into concise statements of the main points. Imagine returning later to the paper, and
only having to read these few highlighted words to get the essential points!
DO NOT highlight whole sentences or paragraphs; doing so means your brain is on auto-pilot.
Don't get stuck in one place. Flip quickly back and forth between the Abstract, Figures, Results
text (and Methods if necessary). The Abstract states a claim, the Figure shows evidence for the claim
plus controls, the Methods answer any nagging questions about the data. By staying moving, you
keep yourself awake, and you quickly draw together the different pieces needed to understand the
point.

If you can't understand a Figure, go back to the Abstract and make sure you understood the claim.
Conversely, if you can't understand what they're really talking about in the Abstract, try going to the
Figure and see if you can understand the claims there.
Scribble copiously on the manuscript, especially the Figures. You have to understand the data in
the figures, so draw on the figures: label key items (e.g. bands); draw arrows between key things to
compare; make your own diagrams of anything you've figured out. If anything bothers you, try to
resolve it quickly by jumping back and forth as described above; if that doesn't clarify it, write an
annoyed comment!

Don't give yourself very much time. This may seem counter-intuitive, but one of the best ways to
break down barriers to reading a heavy paper is to sit down twenty minutes before some other
appointment absolutely determined to "figure out what this paper's about" within the twenty minutes.
Just do it; you will figure it out. And after that, coming back to the paper later is easy. Just imagine
you’re a reviewer whose plane is landing in twenty minutes, and you have to decide whether this paper
is crappy or good!

Pass One: What's the Point? (5 minutes)
Read Abstract: get the key Nouns & Verbs
Synopsis: using these nouns & verbs, what's the point of the paper in three sentences or less? Should say
what alternative hypotheses were tested, and what was ruled out by the data.
Figures: which figures show which points? Match nouns & verbs to figure legends.
You should now have a pretty good idea of what the paper claims to show.

Pass Two: What's the Evidence? (15 minutes)
For each key point, consider the relevant figures:
• What kind of experiment(s) was done, and what is it capable of showing?
• what do they claim the figure shows?
• how does the figure "show" the claim(s)? Mark or draw on the figure to indicate the key evidence.
• what could be wrong with the claim(s), and what controls address that?

By restricting my attempt to understand the paper, to the figures, I make sure I understand the data itself,
rather than relying on what the authors claim in the text. That means figuring out what the individual
bands are. It's also a lot faster than wading through the whole paper!

Data are frequently quite "dirty" and you can perceive many problems quickly. I sometimes jump to the
relevant Methods section to quickly resolve (or raise) troubling questions about the data.

More In Depth, if desired
At this point, reading the Introduction, Results or Discussion should be easy and interesting, because you
know all the players and understand the paper. Anything the authors introduce as background, raise as
questions, or speculate about, should shed interesting new light on the players and points you already
understand.
Of course, you can dissect a paper in endless detail if it's of great interest to you. For the purpose of this
class I suggest you consider the questions on the attached sheet ("Questions for reviewing a paper").