R-squared in regression

Hi,

I was several times told that in regression the r-squared should be
written with lowercase "R", and NOT the uppercase. But I've never
heard any argument for this. I checked and there is no consistency in
this, and for example in "Applied Regression Analysis. A Research
Tool" by JO Rawlings, SG Pastra and DA Dickey capital R-2 is used. I
would actually vote for the uppercase R since the lowercase one may be
misunderstood as a squared correlation coefficient (in simple
regression, of course), which it is not (even though the two
quantities are equal indeed). What is your opinion?

Thanks,
Wolf
.

Quantcast