How do I interpret hierarchical regression results?

Question

How do I interpret hierarchical regression results?

Answer 1

Welcome to the site

You haven't told us what you are trying to do and I can't read your tables, but let's consider the situation more generally. When we compare two regression models, and model B has all of the variables in model A, plus some additional ones, what might we be interested in?

1. How parameters change. Here, we probably ought not be too interested in which particular variables are significant in which models. The change from significant to not significant can be irrelevant - for example, if a p value for a variable is .049 in one model and .051 in another, who cares? We ought be more interested in how the effect sizes of the variables change. Do those changes matter?

2. What the parameters are. Again, effect size may well be more important than significance. Sometimes, finding a small effect size can be important - e.g. if other studies in the field have found large ones; or if theory suggests large ones.

3. What $R^2$ is. This is usually somewhat interesting. A very low $R^2$ may indicate a poor model, even if it is significant (if N is large, a low $R^2$ will be significant) But the definition of "low" varies from field to field. Given that one of your variables is marital status, I'm guessing this is a study about some aspect of human psychology or behavior; in these fields "low" is quite low!

Now, why might the addition of a variable that is not significant affect the parameter estimate for another variable? One possible reason is related to mediation and moderation (but without more details, it's hard to say). Another is collinearity.

If you edit your post to describe your problem better, we may be able to give more detailed help.