I have a question regarding quant that has always bugged me. When calculating the T-stat sometimes you calculate it as (coefficient/std error). And sometimes it is (Number-1/std error). Why do you subtract the one at times and other time you don’t. And example of this is question 9 and 10 on the BSAS exam. The first question had the null: Ho: B <= 0 Ha: B does not > 0 The second qustion was Ho: B = 1 Ha: B does not = 1 Is it that when you see the null as equal to 1 you subtract it out to get the T-stat. Sorry as I know this is Level 1 101.
because it’s (Coef-Coef0)/std error in the first problem your Coef0 = 0, in the second one Coef0 = 1.
you’re testing vs. 1 or 0 is X equal to 1?.. then coeff-1/std error is X equal to zero… the coeff-zero/std error. same for greater than 1 or zero… only thing that changes is that now it’s two-tailed. i find this stuff somewhat confusing too.