Do you compare the absolute value of the t-statistic to the critical value, or the absolute value of the critical value to the t-statistic or the absolute value of both? Example: t-stat=-0.5, t-critical=-0.4. Do you reject the null?

Reject the null in that case. Absolute value of both. But if you look at the tables, they don’t give you negative critical t values.

Isn’t t-critical usually just stated as a positive number? I think you would compare abs. value of t-value to t-critical. In the case above, reject.

rellison Wrote: ------------------------------------------------------- > Example: t-stat=-0.5, t-critical=-0.4. Do you > reject the null? abs(-0.5) > abs(-0.4) = Reject Null

You dont have to use absolute values. For a two tail test, the t-c that you obtain from the table will represent the +t-c(positive )and -t-c (negative) because you need two intervals You will reject Ho if t-s+t-c (t-s is outside of the two intervals) and fail to reject Ho when -t-c