How should you read academic papers efficiently?

Not front to back — that is the single habit that makes academic reading slow and demoralising, and nobody who does it professionally reads that way.

The core principle: read to answer a question, not to absorb a document. Decide what you need from the paper before opening it.

The three-pass approach, which is the standard advice and works:

Pass one — five to ten minutes. Title, abstract, section headings, figures, and conclusion. The aim is a single decision: is this paper relevant enough to read further? Most are not, and abandoning them quickly is the skill.

Pass two — around an hour. Read for the argument. Introduction, then results and discussion, paying close attention to figures and tables, which frequently carry the substance more directly than the prose. Skip methods detail and proofs. You should be able to state what they did, what they found, and whether you believe it.

Pass three — only for papers central to your work. Read the methods properly, check whether the analysis supports the claims, and consider how you would have done it differently.

Read in a useful order within the paper: abstract, then conclusion, then figures, then introduction, then the rest. The conclusion early tells you where the argument is going, which makes everything before it easier to follow.

Practical habits:

Take notes in your own words. Highlighting produces almost nothing; re-expression forces comprehension.

Record the one-sentence contribution of every paper you read, and you will remember them.

Follow citations backwards to find the foundational papers, and forwards — via citation counts — to find what happened since.

Read the limitations section, which authors write carefully and readers routinely skip.

Accept partial understanding. Everyone reads papers they only partly follow, including experts reading outside their subfield.