body { -ms-overflow-style: scrollbar; overflow-y: scroll; overscroll-behavior-y: none; } .errorContainer { background-color: #FFF; color: #0F1419; max-width: 600px; margin: 0 auto; padding: 10%; font-family: Helvetica, sans-serif; font-size: 16px; } .errorButton { margin: 3em 0; } .errorButton a { background: #1DA1F2; border-radius: 2.5em; color: white; padding: 1em 2em; text-decoration: none; } .errorButton a:hover, .errorButton a:focus { background: rgb(26, 145, 218); } .errorFooter { color: #657786; font-size: 80%; line-height: 1.5; padding: 1em 0; } .errorFooter a, .errorFooter a:visited { color: #657786; text-decoration: none; padding-right: 1em; } .errorFooter a:hover, .errorFooter a:active { text-decoration: underline; } #placeholder, #react-root { display: none !important; } body { background-color: #FFF !important; }

JavaScript is not available.

We’ve detected that JavaScript is disabled in this browser. Please enable JavaScript or switch to a supported browser to continue using twitter.com. You can see a list of supported browsers in our Help Center.

Terms of Service Privacy Policy Cookie Policy Imprint Ads info © 2023 Twitter, Inc.

Conversation

Over the years I've done like eighty billion twitter polls and also a metric poopton of surveys with huge Ns where i stare at correlations a ton. I've developed my own personal intuition for correlation strengths; idk if it's "scientific" but the way I tend to look at things:

9

1

188

My rough threshhold for "this is kinda interesting" is around r=0.12. "ok i'm paying attention" starts at around r=0.18, and "i'm getting excited" starts at 0.22. Anything 0.3 and up gets first-time-watching-lord-of-the-rings level feelings of magic and r=0.5s mean i am deceased

5

7

147

r=0.7s and up are mythological creatures no one has ever seen one, and if u think u saw one it probably was just some dude in a suit made out of a pitiful sample size

9:42 PM · Jun 22, 2022

This Tweet was deleted by the Tweet author. Learn more

Replying to

yeah hugely dependent on the subject i think! For example, height and weight is correlated at around r=0.5.

1

6

Show replies

AI Actor-Critic

Replying to

If you are interested a bit more in modelling, you should definitely try proper graphical models (eg Bayesian nets) where you can ask questions directly, what's the possibility of A given B, and not look at correlations anymore, but posteriors.

Replying to

As a physicist, anything below R^2=0.98 is worthless. Interesting how the acceptable tolerance changes so much depending on the area

2

sarahsecrets - now in dark mode

Replying to

in physics anything under 0.9 is not even considered a result

2

Coltan Scrivner

Replying to

I got an r of .7 btw a scale I made and a behavioral paradigm and I almost cried

2

Vincent_the_narrator

@NarratorVincent

Replying to

My abs hurt from laughing

Replying to

What are some of the ones that have surprised you the most?

Show more replies