Hacker News new | past | comments | ask | show | jobs | submit login

Not exactly the best experimental design here. Instead, canadian cross should be used which would reduce and detect variation between individuals.

(Same devs on multiple occasions, sometimes well slept, sometimes not. Matching tasks for both groups. Learning effect is controlled by swapping group order of task execution. This is fully factorial fully crossed design.)

However, the magnitude is indeed big and the results probably valid but might not quite match up for experienced devs.




IMO, as far as research has been done in this field: the only point they need to make with experiments at the moment is that it has a huge impact.

Hopefully, they now are able to get amazing grants because of these results to study it more in-depth.

The politics of science >= doing actual good science [1]

[1] Provided the right intentions are there (advancing truthful knowledge of humanity)


(Shameless plug) we are developing a “study as a service” project to drive home the individuality and impact of sleep.

The plan is to take a week with someone - comparing various “in bed” times along with quality of sleep to a simple, dynamic cognitive task.

If anyone finds that interesting, please reach out to the email in my profile.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: