I am trying to determine what statistical test I should use. I am comparing the results of a control group after treatment to the rest of the population. in this case it is teat scores from one class of students who receive an intervention to the rest of the students at that grade level in the district. the control group is about 20 students and the rest of the population is about 600 students.
Unpaired t- test is suggested, if you take only means of both treatments you can pick another test. Regards
You don't need a test if you have the data from the entire population.
Hi,
I would suggest a t-test since you are obviously dealing with a numerical variable (average). However care should be taken concerning the data sets size's and their meaning. If the second group (600 students) represents a 'statistical population' then it is meaningless to perform a test. If not, you can sample about 20 from 600 students to form a control group, while keeping the first group of 20 students as treatment group.
Yves Yannick
- Badges
- Science topic