Digital note-taking for research

Note-taking with the output in mind

Ever since starting the psychology programme in 2011, I have been taking my notes in a digital format. To me the benefits have been clear from the beginning: no lost papers and the ability to easily include images, links and other media. But the notes from different courses have been living in isolated folders unable to “speak” to one another, which means I have often ended up with multiple notes on the same topic. When I have written things it has been hard to find relevant notes to use without navigating through old folders. This is until I started using linked note-taking in Obsidian.

Obsidian is a software that just “sits” on top of your folders and reads the markdown-files inside. Compared to other note-taking software, I really appreciate the ability for me to open up the files in any app that reads markdown and the files are not locked into an app that I might abandon some time in the future.

The key feature that enables linked note-taking is using Wiki-style links (in this case double-brackets [[like so]]) to create bidirectional links between two separate notes. So now, when I read or hear something which reminds me of another idea/article/conversation, I can create a link between the notes and there is now a connection. The ideas become connected and start to interact with one another.

Here’s an example. In September last year, I watched the documentary about AlphaGo, an artificial intelligence algorithm trained to play the Chinese game of Go. One thing I reflected on while watching the movie was that that the AI came up with new and original strategies which a human would not have used. Why? Because the AI only cared about winning, even if it was by the smallest margin, and chose moves that gave the highest probability of winning in the end. A human, on the other hand, typically wants to win with a secure margin. A few days later I was listening to a podcast about the electoral system in the US and how the structure gives rise to insidious tactics like Gerrymandering. Winning by the smallest possible margin is ideal, because any votes beyond the win are “wasted”. AlphaGo and Gerrymanderin are two separate contexts, but to me it was an interesting connection nonetheless. In a traditional folders-based note-taking archive, this connection would have been lost as soon as my mind wandered elsewhere, but thanks to linked note-taking I was able to establish the connection.

If I learn about something else that creates a similar situation, it will be much easier to connect it to these previous observations because I now only have to remember either the AlphaGo documentary or the concept of Gerrymandering. Finding connections between seemingly distant observations is even more easy with the use of Maps of Content, described below.

The blue text is a link to another note using double brackets.

Linked notes in research

I have seen the biggest gains of linked note-taking while reading and creating draft outlines for research projects. My process is very much a work in progress but it goes something like this:

  1. Each paper gets its own note
  2. If the paper is related to a topic I work on, it gets linked. For example [[OCD]]
  3. Topics I currently write about or are interested in get their own “map of content (MOC)” note. This concept is from Nick Milo and can take many shapes but my MOCs look like outlines or tables of content.
  4. As new writing projects emerge, I scan one or multiple MOCs for papers to use and add them to a new note to start writing with a few key references already in place. I can also create a new MOC and start to build a tentative outline from scratch.

1. Each paper gets its own note

I use a template to quickly add some basic information about the paper. The note is named after the Zotero citekey (lastnameYYYY) and then the paper title. The screenshot below shows a typical paper note for an article I read about obsessive-compulsive disorder (OCD).

There are more advanced and automated strategies than mine (see “More reading” section below) but this strategy is easy to work with and suits me at the moment.

Example of a note on a scientific paper.

So pretty simple. The example citation can go straight into a draft if I don’t have anything else to add, but I can link to the note directly if I want to dive deeper or add more context from the paper.

The example paper is directly related to OCD, so I have added a link to the Map of Content for OCD. At this point nothing else is needed, if I open up the note for OCD I will see the link and can quickly browse papers related to the topic. However, in this case I chose to add it somewhere in the MOC because I often need to access papers on various aspects of OCD.

3. Maps of Content for topics I am interested in or write about

Since I am reading quite a lot of papers about OCD I have created a loose outline with papers from various topics. The blue links indicate specific papers that I have added to the note, and the example paper shown above is added to the “Prevalence, comorbidity, and natural course” subheading. I have been a bit sloppy and not properly contextualized every note, but it’s a work in progress after all!

Map of Content example.

The panel on the right-hand side is super useful. I see a list of both linked (with double brackets) and unlinked mentions (i.e., no formal link with brackets but mention of OCD somewhere in the file). I can also use the “local graph” at the top right hand side to navigate to specific linked notes. Zooming in on the graph reveals the note titles. This is helpful if I want to explore other potential connections.

4. New writing projects use MOCs as a starting point

In a very idea-driven note-taking system like a Zettelkasten, new writing projects can emerge organically just from the sheer clustering of notes. Let’s imagine I took more notes on the peculiarity of contexts with binary outcomes like Go and the US electoral system. Perhaps after a while I would see an interesting cluster and have something to write about, or have as a starting point for a writing project.

However, the type of research I am involved in is a bit more planned than that and there’s usually a pretty clear idea about the output when we start a randomized clinical trial, for example. But the linked notes are still be useful for developing my ideas about potential new projects. They also help me to reflect more actively about what I am reading.

In the last year I’ve started reading research that is slightly detached from my ongoing research but where some studies are about the same psychiatric disorder, OCD. After a while, I created a MOC (blue dot) for that topic and linked some papers that I had read previously. As I continued to read more papers (yellow dots) I could see some key studies that were linked both to OCD (big blue dot on the left) and the new topic (blue dot on the right). These were a great starting point to figure out some potential studies of my own. I also expect that the blue dot on the right will continue to sprawl out in other directions.

Two MOCs in the graph view.

Conclusion: Linked notes increases the value of my reading and enhances my creativity

This method is still fairly new to me but I can already see how the added functionality of linking notes together helps me contextualize individual papers I read. They are no longer just sitting in an isolated document somewhere, gradually forgotten, but can be used and re-used for various purposes.

Using maps of content is a very playful approach and feels like a digital sketch pad. I can link disparate ideas together and think about the possible connections, or create fairly structured outlines related to a specific research project.

I have used Obsidian to facilitate this linking so far and really enjoy the application. It’s free to use and can be heavily customized to fit individual needs. But if I choose to use another app some time in the future, the notes are still there and will be easy to move since it’s just a bunch of lightweight markdown files.

More reading