PDA

View Full Version : negative t-Test statistic


lavster
Oct21-11, 09:01 AM
I have calculated the one sided t-test statistic and am getting a negative value as the sample mean is less than the mean under H_0. there are no negative t values in the table - do i take the absolute value of this?

thanks

Cant or Wont
Oct28-11, 11:01 AM
Take the negative of the critical value in the table and reject if your t-value is less than this.

I'm presuming you're doing a one-sided test on (<). In this instance you could take the absolute value and reject when t is greater than, but you can probably see how always taking the absolute value can go wrong for a one-sided test?