As per Eqn (10) on Page 311 Book 1 (CFAI) , T stat = (b1^ - b1 ) / sb^1

and Solution to EOR Question 5 (Page 341 book 1) says , T stat = [r*((n-2)^.5)]/[(1-r^2)^.5]

How do we know when to use which formula for calculating T stat ?

Thanks

Hi

I havent had the time to pick up the book and check the exact question, but the first formula is for testing significance of regression or slope coefficient b1 while the second one is for testing significance of correlation coefficient .

These are 2 different things and there should be ideally no confusion regarding which one to use since for the first you will needed information from the regression including slope coefficient (b1 value) and slope std dev (sb1) while for the second one you need correlation coefficient r (which is also ( Square root of the coefficient of determination R^2)). Also question will be very clear as to what it wishes you to check - the slope coefficient or correlation coefficient.

Hope that helps

Cheers

That explained everything !! Thanks a lot Manâ€¦

Gd lk