I compared 2 groups and found that their test scores are significantly different. However, I also found that they are significantly different in the mean age. I want to see if the scores between the groups are still significantly different, after removing the effect of age. The correlation between scores and age is very small, $r = 0.03$.
One option would be to use ANCOVA, but this is a bit outdated technique IMO.
My other option is to use a regression model, but I can't figure out the model, specifically what are my DVs and what is my IV.
Best Answer
I agree with @Sointu's comment.
There is nothing "outdated" about ANCOVA. In what I see it's most often just approached through regression. The techniques aren't alternatives, at least not for your set-up.
Your outcome or response is test score and your predictors are age and your group variable. There is one outcome (DV, in your terminology) and two predictors (IV, in your terminology). If you're thinking that regression is predicting one independent variable from one or more dependent variables, you have the terms the wrong way round. I've seen that confusion often, and it's one of several reasons why -- if anything is outdated here -- the terminology of dependent and independent variables is top of the list.
Whether any effect of age is linear is a further question.