Reading papers and keeping track
The problem is rarely reading a paper. It is reading two hundred of them, deciding which deserve real attention, and being able to find what one said long after you have forgotten reading it.
Updated
Every research degree contains a moment where the folder stops being a resource and starts being a source of dread. There are more papers in it than hours available, several have been read twice by accident, and the note file for the important one says "useful, check later".
The pages below deal with that directly. They cover triage before reading, what a note has to contain to survive six months, why highlighting stops working past a certain library size, reading in a second language, and how to get back into a project after months away.
One principle underlies most of the answers. The expensive part of research reading is not the reading, it is the retrieval: getting back to the passage that made a paper matter. Systems that optimize for capture, more notes, more highlights, more tags, usually make retrieval worse. Systems that optimize for finding your way back to a specific passage tend to be smaller than people expect.
How do I organize and extract information from more than 100 papers?
Why folder systems collapse past a hundred papers, what to record about each one so the note is still useful later, and how to keep the system small.
Read the answer →How do I triage 200 PDFs before deciding which ones to read in full?
How to triage a large pile of PDFs before deciding what to read in full, what the three-pass reading method is, and how to record a decision not to read something.
Read the answer →How many papers can a PhD student realistically read in a day?
A realistic figure for how many papers a PhD student can read in a day, how thoroughly you are actually supposed to read, and whether not finishing papers is normal.
Read the answer →How do I remember what was in papers I read months ago?
Why you forget papers you definitely read, what a literature note has to contain to still be useful later, and why highlighting fails as a memory system.
Read the answer →How do I find a paper when I only remember the argument it made?
How to find a paper when you only remember the argument it made, which search moves work when keywords fail, and how to stop it happening again.
Read the answer →What do I do when my note-taking system takes longer than reading the papers?
What to do when a note system costs more than it returns, what the minimum useful literature note is, and how to simplify a system that has grown out of control.
Read the answer →Why can I never find my PDF highlights again?
Why you can never find your PDF highlights again, what to do while reading instead, and how to make annotations searchable across a whole library of papers.
Read the answer →How do I keep track of connections between papers?
How to keep track of how papers relate to each other, whether an atomic note system is necessary, and how to see which papers speak to the same claim.
Read the answer →How do I read research papers efficiently when English is my second language?
How to read academic papers efficiently in a second language, whether to translate as you read, and the fastest way to build the vocabulary of your own field.
Read the answer →How do I include non-English studies in a literature review?
How to include non-English studies in a literature review, whether machine translation is good enough for screening, and how to report a language restriction.
Read the answer →How do I keep up with new papers without reading everything?
How to keep up with new publications without reading everything, how to set alerts that are signal rather than noise, and how to track work citing your key studies.
Read the answer →How do I get back into a research project after months away?
How to get back into a research project after a long break, whether you have to reread everything, and what to write down before a planned interruption.
Read the answer →