Reading a scientific paper in three passes

Students arriving at a literature review often do what school has trained them to do: open the first paper, start at the title, and read every word until the references. Three hours later they have read one paper, remember little of it, and have fifty more in the folder.

Researchers do not read this way. They read in passes, and most papers never get the third one.

The first pass: five minutes

Read the title, the abstract and the headings. Look at every figure and read its caption. Read the first paragraph of the introduction and the whole of the conclusion. Skim the reference list for names you recognize.

At the end, write one sentence: what did this paper try to do, and did it succeed? Then decide. Most papers stop here, because they are not about your question. File them with the sentence.

The second pass: an hour

For papers that survive, read the whole thing, but still not every word. Read the introduction carefully: it tells you what the field thinks the problem is and who else has worked on it. Read the method closely enough to judge whether you would trust the results. Study the figures until you can explain each one without the caption. Read the discussion for the authors' own account of the limits.

Mark anything you do not understand, but do not stop to look it up. At the end, write a paragraph: the main claim, the evidence for it, the strongest limitation, and the one thing you would ask the authors.

The third pass: as long as it takes

Only a handful of papers deserve this. These are the ones your project is built on, or will argue against. Reread the method as if you were going to repeat it. Check the statistics. Follow the most important references. Try to reconstruct the argument from the data without the authors' prose, then compare.

By the end you should be able to say where the paper is strong, where it is weak, and what it leaves open. That last part is where your research question lives.

Keep a reading log

Every pass ends in writing. A reading log, one entry per paper with the date, the pass, and your sentence or paragraph, turns reading into a resource you can search. When you write your literature review, it will be a synthesis of your log, not a scramble through PDFs.