How do I interpret r values? Ppl usually say <3 is weak, 4-6 is medium, and 7+ is strong, but that this is dependent on field - but i can't find anybody saying more about how this changes per field! I'd love to see a chart of what's a medium strength across different fields
Conversation
Replying to
In most scientific fields it's p values that have significance. Pun intended. You cannot make a conclusion on correlative strength alone. Probably why you're having as much difficulty as you are for finding metrics based on a particular field
3
2
Replying to
Report a correlation and let the reader decide if s/he thinks it's weak, moderate, or strong. And realize that it's just an opinion.
1
Show replies
Replying to
I think of an R value as describing how much of the variation can be predicted if you know the relevant variable (interestingly, this is literally what it means). Therefore the more complex the prediction, the lower your standard for a strong correlation.
3
6
Show replies
Replying to
I would take it a step further and say it really depends on the specific observed correlation, even intra field
This Tweet is from a suspended account. Learn more
Replying to
Def depends on field. I'm guessing you're thinking of the behavioral or social sciences. See:
journals.sagepub.com/doi/10.1177/25
Correlation between parent and child income is r ≈ 0.3. Up to you if you think familial transmission of income is strong or not. But that's approx the value.
1
7





