Explore BrainMass

Which statistical test to use?

You want to compare differences in sales dollars (DV) for 43 employees before and after they have attended a motivational seminar.

Since comparing differences is the goal, I presume that you are going to compare the means of a pre-post test of the 43 employees. Okay, I was going to say that this is the answer but lets be honest, I feel that I am guessing, so I am thinking it is a McNemar test because you are doing a before-after measurement of the same subjects.

At the same time, because you are doing a before and after test of the same subjects you are one could also use the t-test for two-related-samples tests because the events are closely matched.

Both almost the same but my intuition is that the answer should be dependent on how the data is going to be measured. If the data is measured at the ordinal or nominal level then McNemar is the way to go. If it is measured a the Interval and Ratio level, then the two-related samples appears to be the answer. I am confused. Can you please explain if in fact I am right? If not which other test is out there that I could use and why.

Solution Preview

Here is the situation.
You have one group of sales people. You record their sales dollars. You then send them to a motivational seminar, and then want to see if it helped improve their sales dollars.

You are completely correct that you do want to compare the sales dollars values of the two group pre and post seminar. In my opinion, since you are using the same sample, I would use a paired t-test.

You use the paired t-test when there is one measurement variable and two nominal variables. In this case, the one nominal variable represents different individuals, ...

Solution Summary

This posting looks at a real life situation and asks us to find which statistical test would be best suited to solve the problem.