Nereid
Staff Emeritus
Science Advisor
Gold Member
- 3,392
- 3
Health warning! health warning! statistics and damned lies are about to be discussed! Parental discretion is advised!Phobos said:Correlation coefficients of 0.44, 0.25, 0.15, & 0.479 are terrible and, to me, indicate no correlation at all. Maybe if the correlations were 0.7 or higher, then you would have something to talk about.
Plus, I have yet to hear about a reliable way to quantify IQ.
Here are two time series*:
57643
58206
58737
59266
59827
60353
60878
61435
61958
62479
7298.3
7624.1
8113.8
8586.7
9066.6
9629.4
10021.5
10338.2
10744.6
11472.6
Bill Gates, via XL, tells me that they are highly correlated (r2=0.9954), which is far, far better than 'cranial size' and _g_.
So, I have a very tight correlation - where to from here?
Just as with 'cranial size' and _g_, it seems to me one can only take the next step if one has a theory to test; otherwise, what is the intrinsic meaning of an r2 that is very close to 1?
In the case of 'cranial size' and _g_, we have a physical measurement of a body part of one mammal (this is Biology, not General Physics or Chemistry), and an abstract concept with roots in psychology (to turn up the contrast, NOT neuroscience).
How does the correlation which is the subject of this thread relate to any theories in biology? AFAIKS, it doesn't; it's simply stamp collecting.
{Note to Monique and iansmith: can you please move thread to a more appropriate section of PF? 'Other Sciences', for example.}
*http://sunearth.gsfc.nasa.gov/eclipse/TYPE/pluto.html .
Last edited by a moderator:
