How to compare a small sample to a large control group sample?


The prupose of my research was to measure the impact of a dropout prevention program. I used a survey consisting mostly of Likert scales and I compared the answers to a control group.
The control group was relatively large, as it was based on a government study of roughly 1500 participants, aged 12 to 17. My group, however, was relatively small, consisting of 195 respondents, also aged 12 to 17, out of a total "population" (the total size of the program) of 250 students. I need to divide the answers by age.

I have already collected my data and created charts, and the results seemed promising at first, as our students "outperformed" the control group on almost every variables. However, someone asked me if these differences were significant or not...

Since it appears like I need to use non-parametric methods for my ordinal data, I don't think I can use a t-test, but what worries me the most at the moment is whether my sample is big enough to reach any conclusion at all On average, I have between 20 and 35 respondants per age level (6 levels from 12y.o. to 17 y.o.).

I read somewhere that a Median Mann-Whitney U test might be the way to go, what do you think?

I hope I'm making sense with this question...

Thank you so much for your help!


Less is more. Stay pure. Stay poor.
Seems like you are on the right track. So you would then conduct 6 different Median Mann-Whitney U tests. You can always run the tests to see if their is a difference, or do a power test.