Just wanted to check if i understood this correctly. the estimated variance of prediction error (s2=s2[1+1/n+(x-x)2… p 322 in the CFA reading) is used when finding the confidence level of the equation as a whole Y+_ ts
While the standard error of estimate (formulae page 306) is used to find the confidence level of a coeefficient b±ts
Correct? no?
Hi Zanalist,
It’s not correct.
In the formula calculating the sf2 to get the confidence interval for Y, you use SEE^2 (see example 18 page 323 in CFA book).
to calculate the confidence interval of a parameter, you use its own standard deviation, not the SEE (see practice problem 12 to calculate X11 and X12). SEE measures the standard error of the error term, so it has nothing to do with the parameters itselves.
Hope this helps.
Picur