How to Read a Research Paper Efficiently

Know the usual structure

Many empirical papers follow an Introduction, Methods, Results and Discussion pattern, often called IMRaD, with an abstract at the start and references at the end. Knowing where to look makes skimming faster: the abstract summarizes the whole paper, the methods describe what was done, and the discussion explains what the authors think it means and what limits it.

A multi-pass routine

A well-known approach, described by S. Keshav in his 2007 article 'How to Read a Paper', uses three passes. This is a simplified version:

1. First pass (5 to 10 minutes): read the title, abstract, headings, figures and conclusion. Decide what the paper is about and whether you need it.
2. Second pass (about an hour): read the paper with care but skip fine details such as proofs. Study the figures and tables and note references you may want to follow up.
3. Third pass (only if needed): work through the details and try to reconstruct the argument as if you were the author, checking every assumption.

Questions to ask while reading

• What question is the paper trying to answer, and why does it matter?
• What did the authors actually do? Who or what was studied, and how large was the sample?
• Do the results support the conclusions, or do the conclusions go beyond the data?
• What limitations do the authors admit, and what others can you see?
• How does this connect to the other sources you are reading?

Taking notes you can reuse

For each paper, write down the full reference first, then a one or two sentence summary in your own words, the main finding, the method, a limitation, and one or two quotes with page numbers if you may quote them. Recording the page number at the time saves a lot of searching later. A reference tool can format the citation, but check it against the paper itself.

Reading for a literature review

When you must read many papers on one topic, a simple comparison table keeps them organized. Give each paper a row and use columns for the full reference, the research question, the method and sample, the main finding, and a limitation. After ten or fifteen rows, patterns appear: which methods are common, where results agree or conflict, and what has not been studied. Those patterns are the raw material of a literature review, far more useful than ten separate summaries.

Using tools carefully

A reading planner and reading-time estimate can tell you how many pages a day a reading list needs, but they cannot tell you how hard a paper is. Dense methods sections and statistics take far longer than the introduction, so add margin, and do not cite a statement you have only seen in a summary: confirm important details by reading the relevant passage yourself.