Manova, covariates and corrected model - significant?

Hi, I'm writing my dissertation that is in on thursday and I am very very desperate for some help for a statistics problem I am having. Anyhelp anyone can give me I would be soooo grateful for.
Basically am testing whether people in different stages of change with regards to exercise remember a different amount or positive or negative statements about exercsie on a memory test. I haven't found any significance using a manova test and so tried to take account of covariates. When taking account of age, I got the results below, and i thought the fact the negative statements corrected model had a significance value of 0.040 meant that when age was accounted for there was a significant difference in the no of negative statements remembered by the different groups. However having got my first draft feedback my supervisor says this isn't right and I have misinterpreted my results. I'm so confused an I've tried looking through spss textbooks, but I still don't understand what the corrected model refers to then :confused: please can someone help me??
Here are the relevant results:
Corrected Model Positive .599
Negative 0.040

Age Positive .344
Negative .024

Stage change Positive .540
Negative .159

Thank you so much for any help anyone can give, I don't understand statistics at all :(