I am reading now many studies about juvenile offending and delinquency, and there are always compared offenders´ results from different tests (for example personality, IQ) with control group. I am asking myself why, when there exist norms that are representative for normal population?
Say your juvenile offender are on average 16 year old boys from Britain, and the control group consists of (on average) 16 year old British boys who haven't done anything wrong. The only difference between these groups are the fact one has a criminal record or not. If you would look at the true norm, you could be looking at the world wide norm of both boys and girls from both Britain and other countries that may or may not have a criminal record.
The control groups (and actual participants) are thus chosen (and tested) specifically to factor out unforeseen variability. Depending on the question authors want to answer, different control groups may be necessary.