Comparing a sub-population to its overall population?

[mrl]

Hi there,

I am collecting data from a population of about 150 or so, and also from a subpopulation within that of 10-15 which are receiving a specific treatment. I am aiming to compare their scores on the same measure, to see if this population receiving treatment is truly 'special' in some way, and that there is not a larger population that also requires this treatment. I am just trying to figure out what statistics to use, as t-test for some reason is not quite sitting right with myself or my supervisor.

Thanks for any help.

 

[hlsmith]

Well 50% of the problem is solved, you want to compare two groups. Now how are the other variables formatted (i.e., nominal, continuous). And can you provide an example of the types of data you will be working with (e.g., age, sex, etc.).

 

[mrl]

Hi, thanks for the reply. It's score data, so it's continuous from 0-15 on a standardised test. I'm working with data from a standardised test of distress, a clinical outcomes measure. The answers given by participants add to a score on a range from 0 (very low distress) to 15 (high levels of distress).

 

[hlsmith]

What is your sample size?

If you go to the literature how do others typically run these data?

Do you want to control for any other variables?

 

[mrl]

My baseline population (so larger population) is 150, my subpopulation is around 10.

Looking at the literature, this hasn't been done before for this area. Otherwise it has been new scores for the treatment group compared with published normative data using parametric statistics dependent on the hypotheses. However, these have generally been from 2 different samples (the treatment group and the published normative data). The reason for not using the normative data is that for several reasons it doesn't really apply to this population, and is outdated.
As there is such a small subpopulation size controlling for other variables is hard, but age/sex will probably be controlled for.