How I Take Notes When Doing Research
How I Take Notes When Doing Research
Following a suggestion by Clive Thompson.
Clive Thompson recently wrote How I Take Notes When I’m Doing Research, in which he lays out his thinking about research, and — to a limited extent — the tools and techniques he employs. He ends the piece asking ‘Let me know if you do something like this, what it looks like, and how it works for you!’, so here I go.
Although Thompson talks with other reporters about work,
one thing we almost never talk about, I realize, is how to take notes when you’re doing research.
Really? Why don’t they talk about that? Is it a secret? Are people’s approaches so different to be incommensurable? Well, a lot of people in the Tools for Thinking community talk about it, although not necessarily with the same intense focus on research for writing.
He then lays the context: he’s a researcher poring over a lot of text.
When I say “doing research” I mean specifically when I’m doing textual research — like, reading books or scholarly articles or news-site posts.
[…]
As it happens, my journalism often requires I read a mountain of material. For any given Wired column, for example, I might read dozens of white papers, reports, and news articles. I’ll also do ten or twelve interviews and transcribe them. When I’m researching a longer feature for a magazine? This number quickly grows to scores of documents, and several dozen interviews. And with a book — like my last one, Coders — we’re talking about literally hundreds and hundreds of documents (books, papers, etc) and several hundred transcribed interviews.
This is similar to my work. A recent piece I did for WebexAhead on the history and future of the work week involved 30 or 40 sources, heavily annotated.
He then makes the case for writing notes about what you are researching, so that it gets internalized, and is also easier to grasp later on:
Early in my writing career I found myself drowning in all this research material. I’d read stuff, but then months or weeks later — when time came to write — I’d forget what I’d read, or what some interviewee had said to me.
But all that reading and clipping never entirely stuck in my mind. It rarely became useful knowledge.
Merely reading something isn’t enough to make it stick in your mind and your memory. Nor is taking a clipping of something — nor highlighting a passage, either digitally or physically.
No, if you want to actually internalize what you’re encountering, and turn it into knowledge — something that’s truly part of your mind, and that you can think with — you need to do a bit more work.
Specifically, when you’re reading, you ought to be writing your own little summaries of the most important things you find.
This might be nothing more than a quick sentence reiterating the main point; or it might be a quick little reflection; or maybe even your own “aha” moment building on top of what you’ve read.
Said differently: It’s the reframing of material into summaries or marginalia that adds value and meaning for the researcher.
Then he talks tools. His is Scrivener.
So what I wound up doing is pretty simple. Every time I start a new journalistic project, I create a new place to store all my notes — which for me is usually a new “project” in Scrivener, in essence a little database for chunks of text.
This is where Thompson and I part company. He starts a new project in Scrivener. I have thousands of notes in Obsidian already at the point when I start a writing project, like this post. But in his case, he starts the project and after that point begins collecting reference materials:
Then, as I’m reading, whenever I stumble upon a useful passage — something I think I might want to refer to in my final piece of journalism — I clip the interesting part, and write a quick sentence or two reflecting on it.
Yes, I do that too, but as I said, I already have materials in Obsidian, tagged and annotated. And I also add new ones. For example, I added his post, and annotated it with tags, highlights, and annotations. Like this:
In the fourth paragraph, which starts ‘As it happens’, you can see a footnote I created for an annotation:
^[A recent piece I did for WebexAhead involved 30 or 40 sources, heavily annotated.]
Note that I can use these footnotes in many ways. I can search in a single file for all footnotes (which I seldom do), search across many documents for footnotes including certain terms or tags, or I can view the document in reading mode and all the footnotes appear at the end of the document for rapid review.
What I really dream of is sidenotes as championed by Edward Tufte, as shown here, where the text of footnotes appear in the right margin.
Someday, sidenotes will be supported in Obsidian.
So, when Thompson finds materials that are important for his research, he does something like my footnotes:
Then, as I’m reading, whenever I stumble upon a useful passage — something I think I might want to refer to in my final piece of journalism — I clip the interesting part, and write a quick sentence or two reflecting on it. Sometimes it’s just a single sentence reiterating, in my own words, what the passage says.
[..] writing those little paraphrases is just enough mental work that it helps internalize this research.
When I’m reading scientific papers, I often try to summarize the most useful findings in my own voice. This is, again, because it makes it far speedier to re-encounter this info months later. But it’s also that the act of explaining to myself what the heck it is I’ve just read helps me really internalize the info.
He doesn’t talk much about searching in Scrivener, but I use Obsidian’s search and Dataview capabilities to find materials relevant to a research project. For example, I searched for the tag ‘#people/clive-thompson’ and found several references in my Obsidian vault.
Thompson’s article ‘hundreds of ways to get shit down’ was something I wrote about in early August, and has some slight overlap with this topic, but I am mostly demonstrating how search makes things easier for re-finding materials.
I only recently adopted the people-tagging approach, but I have notes going back to 2013 in my vault. Here’s a brute force text search on “clive thompson” using the Query Control plugin, that yielded a dozen or so hits:
As Thompson is quick to admit, this is a lot of work:
But it is a lot of work. When I’m reading and clipping out key passages, it slows me down a lot. But this is the whole point of what those cognitive psychologists explained to me, years ago. To turn notes into knowledge, you need to do some active work.
And have a system to retain it and re-find (re-search?) it later.
One final note: When I am starting to build a body of research on a topic, I will use some additional tools in Obsidian to help. For example, I can extract highlights from a document — like the one based on Thompson’s article — and convert those to a Kanban with cards based on the highlights, like this:
Note that the footnotes are pulled to the footer of each card, and not the entire document, which makes the annotation more obvious and contextualized. These cards can be reorganized, edited, deleted, and so on. When I have many documents being used as source materials, each document can be pulled into its own list in the Kanban, and the outline of the final essay can be composed by copying cards and moving into the outline. This is the approximation of moving around cards in Scrivener in Obsidian.
In Conclusion
Clive Thompson’s approach to research is quite similar to what I do, just a transposition from the affordances of Scrivener to Obsidian. And the real takeaway isn’t the ways the tools work to support our efforts, but how our minds work to make sense of the process and particles that make up our research activities.
The way I see it, the writing part of deep research projects is made dramatically easier because of all the upfront work during the finding and annotating of research sources. But, as Thompson says, it’s a lot of work. It doesn’t diminish the time spent thinking, but it shortens the time spent refining, rationalizing, and writing.