Hi Eugene For Independent/Unpaired t-test you must first check if the variances are equal or unequal, then you can know which formula to use, right?? How did you know, without checking the F-stat if the variances were equal or unequal, and you went straight ahead to the formula for unequal variances??
+Lynette 3132 Negative t statistic falls on the left side of the t distribution. Whetehr it is a one or two tail test, treat it in the same way as if it were positive and falling in the right side. If you t stat is less that the critical value in the left tail (eg - t stat = -3.45, t crit - 2.45) then reject your Null hypothesis. Be careful with negative values. In the t tables use absolute (no negatives) only - it does not show critical values as negative for the left tail. Dr E.
![[TF2] The Russian Bear Trap](http://i.ytimg.com/vi/YFNSwz2_rOA/mqdefault.jpg)
Hi Eugene
For Independent/Unpaired t-test you must first check if the variances are equal or unequal, then you can know which formula to use, right??
How did you know, without checking the F-stat if the variances were equal or unequal, and you went straight ahead to the formula for unequal variances??
What happens if the t-statistic is negative? and the degree of freedom have a decimal (90.5)
+Lynette 3132
Negative t statistic falls on the left side of the t distribution. Whetehr it is a one or two tail test, treat it in the same way as if it were positive and falling in the right side. If you t stat is less that the critical value in the left tail (eg - t stat = -3.45, t crit - 2.45) then reject your Null hypothesis.
Be careful with negative values. In the t tables use absolute (no negatives) only - it does not show critical values as negative for the left tail.
Dr E.