Published on July 14, 2025 8:20 PM GMT
Tl;dr how can I improve my literature-review based posts?
I write a fair number of blog posts that present the data from scientific papers. There’s a balancing act to this- too much detail and people bounce off, too little and I’m misleading people. I don’t even think I’m on the pareto frontier of this- probably I could get better at which details I share and how I share them, to improve readability and rigor at the same time. This post is a little bit my thoughts on the matter and a lot of requests for input from readers- what do you actually want to see? What are examples of doing this well? Any requests for me personally?
I ask for audience feedback explicitly at a few points, but please don’t limit yourself to those. I’m interested in all suggestions and examples .
Context
If you’re just tuning in, here’s a few examples of posts I mean:
- Salt water gargling as an anti-viralLong Covid Is Not Necessarily Your Biggest ProblemIron deficiencies are very bad and you should treat them
These are all posts where the bulk of the text is describing individual papers, but I have some conclusion I would like the reader to consider.
My motivating example is my project on the risks of long term ketamine use. Right now I’m working on a technical post on how to translate doses consumed by humans into concentrations in the cerebro spinal fluid (draft ), which is reference material for a post people might actually read.
Principles
Epistemic Legibility
My goal is always to present information to people they can interpret for themselves, rather than rely on my summaries. My proudest moment as a researcher was when I was hired by a couple to investigate a particular risk during pregnancy, and due to different risk tolerances they came to opposite conclusions from the same model. To accomplish this, I need to give people the relevant details, in as digestible a format as possible.
What helps you connect with scientific posts? Some ideas:
- My search processMy selection criteriaMy conclusionsMotivationYour ideas here
Then there’s the papers themselves. For the ketamine dosing post, there’s <20, maybe <10 papers in the world that meet my criteria for inclusion, so it’s feasible to include details on each of them. But which details help people understand, and which aren’t worth the attention they cost?
Some paper details I could include:
- Sample size. Experimental set upKey graphsDescription of results
- Averages, or with confidence intervals?
Readability
All else equal, it’s better for a post to take less energy to read than more. Actually that’s not quite true- for posts that would be especially costly if I’m wrong or I expect to be misinterpreted, I will often bury the conclusion, like I did in this post on binge drinking. But we’ll ignore that for now and focus on the much more common case of wanting posts to be as accessible as possible.
Detail and readability often trade off against each other, but what I’m looking for here is ways to improve readability while holding detail constant. Some ideas I have:
- Formatting, probably? Seems like it should help but I don’t know what specifically.Humor
- Unfortunately the easiest way to do this is to make fun of bad studies, which gets repetitive.
Audience
Everyone says to have an audience in mind. There are two major audiences and two minor.
People who are Interested in the Opinions of Uncredentialed Internet Weirdos
This is a tautology, but refers to something much more specific than it looks at first. People who are interested in hearing uncredentialed randos describe and interpret academic papers have a lot more in common than just their willingness to do that.
Some other traits they share:
- Statistical literacyDesire for interpretations to be quantifiedHigher risk toleranceYour ideas hereInterested in the specific topic- such as ketamine use, or long covid risk.
- It’s rare I want to convince people that they should be in a topic when they weren’t before.
Fishing for Corrections
Some posts aren’t meant to be read widely. They’re meant to be a reference in other, more readable post, and to invite corrections from the three people who will read them. This is my intention for the ketamine dosage translation post– it’ll be lucky if it’s read by 10 people when it’s first published, but one of those might be quite useful.
The primary benefit to me is catching mistakes before I write an entire 10,000 word post with information that could hurt people I’m wrong that depends on the false conclusion. It also feels virtuous to explain my reasoning in detail, even if nothing specifically good comes from it.
Myself
Writing lets me think through things. I always budget at least as much time for the “writing” phase of a project as research, because there are gaps I don’t notice until I start writing them down.
I’m interested in how this works for other people- have you found ways to improve your writing for yourself?
Potential clients
I make my living as a freelance researcher, with my blog being the major evidence I am good at this. I’d like clients who read my posts to be able to assess my skill level, even if they’re not interested in the topic and have no context.
Conclusion, such as it is
I would like to get better at writing the kind of posts I write. In particular, I’d like to get better at conveying relevant information, in ways that take as little work from the reader as possible, but no less than that. I will be very grateful for feedback that helps me improve or that helps me create a framework by which I can improve. I expect that to mostly be critical, but compliments are helpful too- I’d hate to throw out the baby with the bathwater.
Discuss
