Page 1 of 1

Confidence intervals

Posted: Sun Feb 17, 2013 9:36 am
by Jonas3636
Hi,

I wish to construct a 90% confidence interval for my regression, E[log(score)|log(sat)=250, log(merit)=14, female=1]

Do I restrict the sample to just the data where female =1 (dummy) and then define new variables [log(sat) - 250] and [log(merit) - 14] ?

And then the intercept is used for computing the interval?

Thanks

Re: Confidence intervals

Posted: Sat Mar 02, 2013 5:00 am
by Jonas3636
Can anyone confirm this for me? Thanks