Reading a paper in an hourএক ঘণ্টায় একটা কাগজ পড়া

A doctoral student reads hundreds of papers. Most of them deserve five minutes, a few deserve an hour, and a handful deserve a week. The mistake is to read all of them the same way, first word to last, and to reach the end of an afternoon with two papers finished and nothing written down. The method here is three passes, and each one ends with a decision about whether to make the next.

Pass one: the shape (five to ten minutes)

Read the title, the abstract, the section headings, every figure and table with its caption, and the conclusion. Do not read the introduction and do not read the method. You are asking four questions: what is this paper about, what does it claim to have found, what kind of paper is it (empirical, theoretical, a review), and is it about my question or only near it.

Write two or three lines. What the paper claims, what kind of evidence it seems to use, and a decision: stop here, or make the second pass.

Pass two: the argument (twenty minutes)

Now read the introduction and the whole of the results and discussion, but skip the proofs, the robustness appendix and the estimation detail. Read to follow the argument: what question, what gap, what data in broad terms, what finding, and what the authors say it means. Mark every claim you would want to check and every number you may want to use. Mark the references you do not know.

Write a short paragraph in your own words: the question, the design, the main result with its size, and what you doubt. Then decide again.

Pass three: the method (thirty minutes)

This pass is only for a paper that will be cited in your own argument or copied in your own design. Read the method and the identification line by line, and try to rebuild it: what is the unit of observation, what is compared with what, where does the variation come from, what would have to be true for the estimate to be causal, and what did they do about the things that could be untrue. Read the tables again, knowing now what is in them. Read the robustness section asking which check you would have asked for and whether it is there.

Write the notes a literature note needs: sample, method, effect size, limitation, and how it bears on your question. That is the next lesson.

The one-line takeaway

After whichever pass you stop at, write one line. Not a summary of the paper: a sentence that says what the paper found and what it means for what you are doing. "Flood losses to farm income were mostly recovered within two seasons, so a one-year window will miss the recovery" is a takeaway. "This paper studies floods and income" is a filing label. The reading room asks for this line because in six months the line is the only thing you will read.

Where it goes wrong

• Reading the introduction first. An introduction is written to make you continue. The abstract and the figures are written to tell you what happened.
• Not writing between passes. A pass with no note leaves nothing behind, and the next paper writes over it.
• Making the third pass out of guilt. Most papers do not need it. The hour is for the ones that do.
• Copying the abstract as the takeaway. The abstract says what the authors found. The takeaway says what it means for you.
• Trusting a figure without reading its axis. A dip that looks like 30 per cent is 3 per cent on an axis that starts at 0.9.

• Pass one done: title, abstract, headings, every figure, conclusion. Two lines written.
• Decision made: stop, or go on.
• Pass two done: introduction, results, discussion. A paragraph in my own words.
• Pass three, if needed: method rebuilt, robustness read, sample and attrition checked.
• One line written: what it found and what it means for my question.
• The paper is filed with that line attached, not with the abstract.

The reading room is where this happens: a PDF drawn on this origin, a highlight anchored to its words, and a box for the one-line takeaway before a source is marked as read. The library keeps that line beside the source.