3 ms·
Research papers demand more read-throughs compared to other genres. My fellow grad students had their own reading styles, but I think my approach was fairly typ
by psychomugs 5y ago
Research papers demand more read-throughs compared to other genres. My fellow grad students had their own reading styles, but I think my approach was fairly typical:
- Review: amass several papers related to the topic, either through search (e.g. Google Scholar) or through a survey paper on the topic.
- Filter: read the abstracts and decide which papers are worth skimming.
- Skim: read the introduction, skim the figures and their captions, and conclusion. This step is another filter; ask yourself whether diving into the technical minutia is worth it.
- Read: read the paper in-depth, taking notes in your manner of choice (which seems to be slides for OP).
- Reread: read and skim the paper a few more times until you could summarize it to someone else. Divorce yourself from the paper by reading other papers, and revisit it to contextualize its contribution.
- jll29 5y agoI upvoted the parent, and I would add another important step, namely: - Compare: read similar papers (such as papers trying to solve the same problem both before and after the current paper) to get a sense of the whole space and range of methods tried to date. Also, look what later papers say about the paper under scrutiny. Finally, reflect on how the various papers found in the 'Gather' phase related to each other: do they refine each other, refute, supersede etc.?
- a_e_k 5y agoThat's pretty close to my approach. I'd suggest not just limiting yourself to just the initial set of papers amassed in your "Review" step, though. Feel free to add new papers to your queue at any time if you come across an interesting term or citation. It's okay for this to be a breadth-first search.
- psychomugs 5y agoI agree on this and definitely left this out, though I do find this step difficult - it’s a recursive loop with a fuzzy stopping condition that can get pretty deep.