digest
Short Summary
Main recommendation: Stop passively highlighting and reading papers straight through; instead, read actively by setting a question before reading, writing notes in your own words, structuring each paper into Claim, Evidence, Limits, and Link, and reviewing them later through recall. The speaker argues this process encodes information far better than reading or highlighting, and that the notes serve as rehearsal for real academic work rather than as a record.
Key Points
- Before reading, ask "why" and set an intention, such as "What is this paper's one claim?", "What evidence would convince me?", and "How does it connect to what I already know?"
- Paraphrase in your own words and never copy verbatim; the speaker cites a study finding that verbatim note-takers did worse on conceptual questions, while paraphrasing forces deeper processing.
- Scan figures and captions first, since captions often carry the take-home message, and then encode those findings into your own language.
- Build a Claim, Evidence, Limits, Link skeleton for every paper, with Link connecting the paper to ideas you already know or to other papers.
- Retrieve and retell your notes after a delay (a weekly review) so the material encodes properly, and look for connections between papers.
Details Worth Noting
- Concrete instructions:
- Get the full text of the paper, printed one-sided or on a tablet or computer, with room to write notes.
- Before reading, spend about 30 seconds writing down what you want to find out, and write the paper's topic and key question at the top of the page. Do not highlight; write notes.
- Read the figures and captions first; if the data is absent, the speaker says that without data, there is often no point reading further in his field (a personal view, not a universal rule). Write each key finding in your own words in the margins or on the back of the page.
- After reading, fill a skeleton on the back of the first page: Claim (the paper's main assertion), Evidence (the data supporting it), Limits (constraints, untested conditions, what was not tested), and Link (connection to prior knowledge or another paper, marked with an arrow).
- Keep each paper's notes with the paper, as they are your own rehearsal rather than a record for others. Messiness is acceptable.
- Hold a 20-minute Friday review: lay out the week's pages, review blind if possible by recalling the claim and evidence aloud from memory, then check against the page, and look for connections across papers (same mechanism, different population, etc.).
- You can run the Friday review with others in your research group by proposing a shared review and discussing the papers read.
- The speaker illustrates the approach with a graphene organic photovoltaic paper: writing the question "what is graphene?", noting the best device as 5% graphene loading with 20 minutes of annealing at 160°C, and linking it to an earlier carbon nanotube paper that used similar doping and annealing approaches (the example link is attributed to a hypothetical 'Stapleton et al. 2002' paper, used as an illustration).
- The speaker notes that the approach works at any stage of a PhD; early on, a broad intention such as "this is interesting" is fine.
- The speaker says that retelling a paper in your own words, as they learned from a mentor, is more valuable than trying to write perfect prose.
Bottom Line
Active, self-generated notes reviewed through recall are a rehearsal of academic thinking, so the value lies in the process of paraphrasing, structuring, and connecting ideas rather than in producing a polished record.
