00:01
The only difference between a one -tailed test and a two -tailed test is the critical region that we use for our test.
00:09
If we look at the first example in blue, we see that we are testing to show that mu is not equal to 100.
00:18
So the critical region, if we were to do a 90 % confidence interval or 90 % alpha level, would be having 5 % on either end of the spell curve.
00:39
So if we had a 5 % shade on the left and 5 % shaded on the right, if any of, so this middle area would be 90 % of the area, if our test statistic ends up being in the shaded region on either side, then we will reject the null hypothesis.
01:04
So this one would be a two -tailed test.
01:13
The reason for this is that it does not matter whether it's lower than 100 or higher than 100.
01:21
If it's far enough away from 100, then we will conclude that mu is not equal to 100.
01:29
Now, if we look at the second example, these are both going to be one -tailed test, the red and the green.
01:37
But the red example specifically wants muted to be greater than 100.
01:43
So again, we will make sure that the center area is 90 % of the area, and this time we would accept anything in the top 10 % of the bell curve, and so this would be 10 % right here, whereas with the two -tailed test, we would only have 5 % on each side...