# Resolved: Alpha value: when alpha itself and alpha / 2 is used?

Hi,

I was wondering when the alpha value divided by 2 and alpha without division is used? In Z and T table we divide it because we are looking for CI, right?

And hypothesis testing does depend on the two-sided or one-sided testing?

I have some questions of my own. But from what I have seen, we use just alpha when there is just one region of rejection, like a case where the null hypothesis is over or under a certain value, and not a specific value. On the other hand, we use alpha divided by two when the null hypothesis is an exact value, thus creating two rejection regions, one which is over the value, and one that is under the value.

Hypothesis testing will be the same for both types of tests, one and two sided, what changes is the rejection region. We use the z-test and we compare the result of that formula to the rejection regions we have made up.