The type of tail you're measuring your stat result against depends on what you're trying to find out.

So if you want to perform a statistical test to see if something is equivalent or not, then you refer to the Two Tailed number.

Alternatively if you want to test to see if something is greater than or less than, then you have to use the One Tailed number.

So for example, if I want to compare two arrays of numbers to see if their means are statistically equivalent (at a 95% level of confidence), then I need a two tailed test.

Hypothesis: Array 1 = Array 2

Alt/Null Hypothesis: Array 1 <> Array 2

If the confidence level I want in my results has to pass the 95% threshold, then my two tailed P Value has to be less than or equal to .05. (1-.95 = .05)

In a one tailed test my hypotheses are a little different, as is my p-value.

Hypothesis: Array 1 > Array 2

Alt/Null Hypothesis: Array 1 <= Array 2

You make the same interpretation based on your confidence threshold as above, but use the P-value for the one tailed test instead: e.g. Is my 1 tailed P-value less than .05.

Best,

Eric~