This question threw me off in their calculation of the standard error. I had not seen it done this way before. Can anyone point me to a section in the reading where they cover this? Basically, we were given a regression table and asked to calculate a t-statistic. The answer given is: t-statistic calc = (coefficient - null hypothesis)/standard error standard error = coefficient / t-statistic (this is from the regression table given above) had not seen std. error calc’ed like that before…

because they assume that null = 0 in the table, but later on, they assume alternative hyp = 1; thus, null does not equal to 1. So you need to use 1 instead of 0, thats why you have to find the standard error because it does not change if you’re calculating for t test with different hyp.

Thanks. Makes perfect sense.