Hypothesis test, how do I calculate the mean from my observations?

So I'm a bit embarrassed to be asking what seems to me like it should be very straightforward, but the answer is not clear to me. To keep the scenario simple, I have 7 subject study, 4 receive the treatment and 3 do not. The measure of interest is a score per minute. I have five months worth of data for each subject before and after the treatment. However, for each subject I have a varying amount of observations. For example, for a single day I have

Subject - Minutes - Total Score - Score per minute
1 - 2 - 16 - 8
2 - 1 - 5 - 5
3 - 1 - 4 - 4

Total score for the day would be 6.25 ( (16+5+4)/(2 + 1 + 1) ). Or do I treat each as a sample and create a sample mean of (8 + 5 + 4) / 3 = 5.67. Which is my measure of interesting for performing a hypothesis test?


Active Member
Hi lonestartiger

First, no need to be embarrassed, asking questions is the goal of the forum ..,

Is it daily treatment? every day before and after the treatment?
Or is it one-time treatment?

Can you please be more specific, what are the minutes and what is the score?

What test do you plan to use?

Generally, I think that both calculations are correct, really depend on what you want to check ...

General group measurement of the group would be calculated as (16+5+4)/(2 + 1 + 1) ).
in this case, you don't treat each person as one equal observation, but the group

But if you want to treat each person as one observation (equal) then you should do average per person's results (8+5+4)/3