A very high stakes confession

I have a long learn-a-language app streak, but my abilities plateaued [redacted] time ago. The app’s inherent limitations combine with my insistence on doing lessons in the worst possible way.

Questions are not created equal

I open the app at 11:52 PM, click past the conceptual lessons, and frantically finish a vocab lesson to preserve my streak.

We've got research indicating I'm sabotaging myself. A team added self-quizzing to The Rust Programming Language textbook. Learners quizzes themselves more than a million times before the team analyzed the data.

Conceptual questions like “why would this program not compile?” could determine which learners grasped a concept vs not. And they worked far better than more mechanical, vocab-equivalent questions like “will this program compile?”

Per my last email (complimentary)

In my last newsletter I advocated for quizzing ourselves early in the learning process. Let's build on that and quiz ourselves with “why” questions. That's not the only way to do conceptual self-quizzing, and it's a decent heuristic.

Learning update 00

I've found this approach super useful while teaching myself neovim. Instead of asking myself “how do I make the cursor go 3 words to the right?” I ask “Why is using ‘f’ instead of ‘l’ to advance in a line often more useful?“ I feel less like I'm answering trivia questions and more like I'm building a mental map.

A data thing I liked

Chapter 9 on design in the Experimentology open access textbook by a murderers row of stats + methods folks

A not-data thing I liked

How to catch an arsonist by Howtown