Hacker News new | ask | show | jobs
by bsmith89 1030 days ago
While I strongly empathize with the author's feelings on this (I've had similar feelings in my own field), I also want to voice an opposing viewpoint:

> Good/useful/valuable/important/positive-ROI science doesn't necessarily require everyone to know what the major discoveries are nor even care when 60% of studies fail to replicate.

Handful of reasons I feel this way:

- If I care deeply about exactly 10 of the 100 publications in my sub-field, and you care about a different 10, it may not matter much to either of us when 60 of them are later refuted. I may have not cared about those specific conclusions, already been skeptical about them, or have several other studies and my own unpublished results to maintain my confidence in the broader idea.

- While we all have great examples of dramatic upheavals in _other_ people's fields — pick your favorite of cosmology, genetic engineering, mathematics, etc. — when you're immersed in it, science is much more incremental, subtle, and complex. Congratulations! You've ventured on beyond the Dunning-Kruger effect. Scientific progress is not a series of miracles.

- Relatedly, I'm guessing most scientists are much more aware of the shortcomings in their own and their peers' research. Do keep this in mind while reading the perspective of an insider. Be skeptical, by all means, but not only of psychology.

2 comments

If Einstein's paper on Special Relativity were proven wrong there would be quite a stir. Is there such a paper in psychology carrying such weight? Or even a central idea?
All of the central ideas in psychology are actually cribbed from philosophy (very selectively and almost never with attribution because you know, we're doing "science"; biased toward material that can be wrenched into guidance for "positive outcomes" that get the grants and TED talk invites), while all the data is extraordinarily messy, mantled under mouldering traditions of associating ones product with the group protection of a figurehead "school" (i.e. so-and-so's Biopsychoturbochemo Model) highly resistant or essentially unsuited to normalization, and of course unrepeatable in any rigorous sense.

The biggest difference between psychology, (alongsid its armed infantry counterpart psychiatry,) is that unlike math or physics or chemistry, which are all intrinsically descriptive, psychology is concerned with how to change behavior that has not yet happened. Not to describe phenomena that already exist but have not been encountered or formally described yet.

> Congratulations! You've ventured on beyond the Dunning-Kruger effect.

https://www.psychologytoday.com/us/blog/how-do-you-know/2020...

I find the irony in this context particularly rich.