Research Design

[Anne]

Hello,

 

This is a little off topic for this section as I see it is mostly around research and how it relates to med school (finding it, getting involved, what schools think about it). I have a question about an experiment I am designing and was hoping someone could help me to figure it out. It is about experimental controls.

 

Ok, so basically I have a test group, Group X, and I am putting Group X participants through a program. I am going to have a control for Group X (Group X2) that are people with the same attributes as Group X that do not go through the program.

 

Here is where I am confused as a colleague suggested that I should also have Group P- a group of the general population that goes through the program and Group P2 (general population that does not go through the program). I suppose the point of these would be so that I can see if the program affects the general population the same way as for those with the specific attributes of Group X. However, to me this isn't just a control it seems like a whole new experiment. The point of the experiment is to see the program's effects on Group X in which case you need Group X2 to do that. Adding in Group P and P2 seems like a new experiment to test the differences in response to the program among two different populations. It might make the experiment thorough, but seems like too much for one paper/test.

 

Am I wrong to view it this way? Should P and P2 be considered controls?

 

Thanks

[thatonekid]

P and P2 would be controls if you want to compare the effects of a program between your target group and the general population. To me, it seems like all you want to look at is how the program impacts people in your target group. That makes P and P2 unnecessary.

[OIC]

I agree with tok, P and P2 seem unnecessary.From the way you've described it, P and P2 would be like phase 1 of a drug study to evaluate the effects of the drug on healthy controls and determine the maximum allowable dose which you would then administer in phase 2 on the actual group of interest. However, I get the feeling that you are likely talking about something closer to psycho-social intervention as opposed to a drug regimen. There is probably little reason to put healthy controls who are not the target group through a type of psychosocial therapy when they don't need it (and are likely to just twiddle their thumbs).

 

If, say, you were creating a program that instituted a psychosocial therapy schedule for people who were depressed, it could be interesting to see if you manage to improve the mood of people who are not depressed at all (as would be determined from your P vs P2 analysis). But, I don't think this is experimentally necessary and it should be regarded, if at all, as a complementary addition.