statistics

khadeja
Posts: 85
Joined: Wed Feb 07, 2018 10:42 am

statistics

Postby khadeja » Mon Apr 16, 2018 11:47 am

What are the R values for a weak, moderate or strong correlation?
NS_Tutor_Andrew
Site Admin
Posts: 520
Joined: Mon May 23, 2016 1:47 pm

Re: statistics

Postby NS_Tutor_Andrew » Tue Apr 17, 2018 4:34 pm

Hi khadeja,

There's not really a fixed answer to that, because it depends on the context. You can find sources like this one that will give you some general guidelines, but to be blunt, it's not accidental that this source is Statistics for Dummies. In reality it depends on the system that you're investigating. If you're doing chemistry/physics research in a lab, analyzing a system that is very well understood and has minimal confounders, then you'd really want to see a high r value. A result of r=0.7 might not be that strong in such a context -- it might still suggest you're missing something. In contrast, if you're doing social sciences research, r=0.7 would be super strong, because societies are messy and people are weird. It's also worth keeping in mind that r values are different from statistical significance, which also depends on factors like the number of observations.

Hope this helps provide some clarity, even though it's not a definitive answer! My snarky comment about the Statistics for Dummies source aside, those numbers serve as a rough starting point, w/ the understanding that you have to interpret r values in light of context & statistical significance.
Andrew D.
Content Manager, Next Step Test Prep.
khadeja
Posts: 85
Joined: Wed Feb 07, 2018 10:42 am

Re: statistics

Postby khadeja » Tue Apr 17, 2018 4:52 pm

hahah nope i knew it was a dumb question to begin with; but what you said is just what i wanted to confirm :D
NS_Tutor_Andrew
Site Admin
Posts: 520
Joined: Mon May 23, 2016 1:47 pm

Re: statistics

Postby NS_Tutor_Andrew » Tue Apr 17, 2018 4:56 pm

Hi khadeja,

It's absolutely not a dumb question. My joke was more that any source that just gave you a list of ranges without providing context is giving a bad answer to a good question.
Andrew D.
Content Manager, Next Step Test Prep.
khadeja
Posts: 85
Joined: Wed Feb 07, 2018 10:42 am

Re: statistics

Postby khadeja » Tue Apr 17, 2018 5:49 pm

ohhh haha right :idea:
NS_Tutor_Andrew
Site Admin
Posts: 520
Joined: Mon May 23, 2016 1:47 pm

Re: statistics

Postby NS_Tutor_Andrew » Tue Apr 17, 2018 6:29 pm

Yes, apologies for my unclear humor (or attempt at humor), but the target was 100% the explanation for not giving context!
Andrew D.
Content Manager, Next Step Test Prep.
khadeja
Posts: 85
Joined: Wed Feb 07, 2018 10:42 am

Re: statistics

Postby khadeja » Tue Apr 17, 2018 6:44 pm

omg please don't apologise!

Return to “General Discussion & FAQ”