Happy Monday!
I hope you had a wonderful weekend. Summer looks like it has itself gone on holiday here, so it’s all rather grey and wet. Never mind, it was nice while it lasted!
This week, I want to step away from the day-to-day news to offer more of a philosophical reflection on an ongoing tension I have been grappling with: the conflict between the need to regularly produce material and a desire for greater depth. And I want to suggest micro datasets as one possible solution — and solicit some ideas on which ones could add value to your work.
With that in mind, here’s what you can expect this week:
- Depth, responsiveness and a lack of harmony
- Productive tension
- Micro datasets as stepping stones to bigger things
To read this newsletter in your browser, click here.
Depth, responsiveness and a lack of harmony
Truly deep work takes time. An in-depth profile of a person could require weeks or even months of research. An investigation into a covert group’s activities could take even longer. Any time you want to move away from the here-and-now — the latest incident, the superficial news report — to move beyond the obvious and look at the bigger picture, you need to be prepared to invest considerable time and effort. Adding value and saying something new is rarely a quick task.
AI, used well, might speed up some parts of the research process, but it doesn’t change this fundamental truth. AI generally produces work that is derivative. If you want a truly original idea, if you want to say something fundamentally new, you’re still going to have to do it yourself.
And yet we all have more immediate pressures. If you produce a newsletter like this and aim for a regular schedule, the tension between your production schedule and the deep work that requires weeks or months is fairly obvious. If you’re a journalist with articles to write, or work in any other sector where a regular output is required, there will be a similar tension at play. The needs of the moment are often antithetical to what might matter in the long-term. Entire careers are lost in the gaps between the two.
Productive tension
Many of us probably entertain a vague dream of shaking off these immediate pressures to concentrate on something more meaningful, more enduring. If I didn’t have to do tasks X, Y, Z, then I could really focus on A, B, C — and deliver the value I know I’m capable of. Blame capitalism: X, Y, Z are, more often than not, tied to the requirements of a job or some other need to earn money. Which in turn pays the bills. A, B, C, unless you’re lucky, might not pay much at all.
Yet it is also important to acknowledge that the tension between the immediate and the deep can be a productive one. The worst thing you can do is spend years squirrelled away, working on something and not telling anyone. Not only will you fail to put your knowledge to productive use in the interim, but you’ll probably produce something less valuable, less interesting as a result. Anyone who engages in any form of public writing knows that it sharpens the mind, forces you to work through ideas, and hones them as they are exposed to public scrutiny. You would, in short, be worse off if you did not produce something in the here and now. More academics should blog for this very reason.
Micro datasets as stepping stones to bigger things
One of both my strengths and weaknesses — and most of my strengths are weaknesses and vice versa — is that I want things to be rigorous and comprehensive. I don’t merely want to know something about a topic. I want to know everything and to be confident that the answers I produce come from the data, not its flaws. Very noble, but very obviously flawed too. First, it’s almost impossible to know everything about a subject (side note: the traditional academic literature review, positioned as a comprehensive survey of knowledge on a topic, should acknowledge its own futility and adopt different goals and forms). Second, the pursuit of comprehensiveness can produce a kind of analytical paralysis. You spend all your time trying to get to the bottom of a bottomless well — rather than focusing on making the best use out of what you already have.
With this — and the forthcoming launch of the Russian Security Research Lab (RSRL) — in mind, I’ve been considering devoting more time to producing what I call micro datasets. These are things that are robust in their own terms, but lack any aspirations to comprehensiveness in relation to the wider topic.
Let me offer an example. If I want to understand terrorism in Russia — its geographical spread, different forms, different connections — I could compile a multi-source dataset of incidents and prosecutions. To be comprehensive, it would need to ensure that events in the periphery are adequately captured; to look at things that didn’t happen or almost happened, or were claimed to happen but actually didn’t; and then to resolve the contradictions that inevitably emerge. It would be an immensely resource-intensive project, even more so than my current tracking of political violence in the North Caucasus. It could only be pursued at the expense of other (more valuable?) projects. A micro dataset, by contrast, would be more manageable. I could, for example, look at what terrorism looks like through the eyes of the National Antiterrorism Committee. This would require only only source, cataloguing its reporting and claims. It obviously wouldn’t answer all the questions, but it could answer some of them, such as how the Russian authorities portray the terrorist threat. The comprehensive dataset could, given other commitments, take six months to stand up; the micro dataset, maybe a week.
A micro dataset, then, is something that adds value but remains conscious of its own limitations. Over time, it could be combined with other micro datasets to form a more complete picture.
This naturally gives rise to the question of which micro datasets in the security realm would be genuinely useful? Hit reply and let me know!
Think someone else would find this useful? Why not forward this to them?