repeated measure ancova

#1
Hi,
I do have a rather simple question but could not find a direct answer on other forums.

I ran a repeated measures ancova with time (2) as within-subjects factor on accuracy rates (so that one was measurd twice) and entered age as the covariate.

in the between-subjects table the effect of age is singificant, how do I name this effect, is it the main effect of the covariate?
and further, does it mean that across levels, age has a different relationship to accuracy? (e.g. running a linear regression on mean accuracy with age as independent variable to interpret that relationship-this actually gives exactly the same p-value)?

I am highly desperate to receive an answer r at least half an answer and would be very grateful for any help, Thank you!
 
#2
Hi,
thanks for your reply. sorry for the confusion, the dv is accuracy rates, measured twice, so time is the repeated measures factor and then age as a covariate. sounds pretty simple.

however, if I then find a significant interaction between time and the covariate, is it appropriate to regress age on the difference between timepoints to visualize what the interaction means?
 
#3
Well that was simple, but the design is still cloudy. The study itself, I mean. I don't know what the accuracy rates is, accuracy of what? what are its levels, is it ordinal or real continuous, etc. If ordinal, how many levels does it have? Why have you measured it twice? Was there any treatment in between? How many patients / subjects did you have? etc... Although these seem less relevant to your concerns, they are the juice of solving the puzzle for the responders.

More important than that, what is the effect of time? Was the accuracy improved in the second measurement? What is the effect of age? Are younger patients more accurate?

however, if I then find a significant interaction between time and the covariate, is it appropriate to regress age on the difference between timepoints to visualize what the interaction means?
Yes it is ok, but the interaction already tells you what's going on. Suppose my guess is correct and younger people are more accurate, and treatment has increased the accuracy (recorded in the second timepoint). Then suppose the interaction of time and age is significant positive. This means that the effect of time (which was positive) will be boosted in older ages. So the treatment is more effective for older people, and less effective (although still significantly effective) for younger people. If the interaction is negative, it means that the effect of time on the accuracy is more vivid in younger people and less obvious (although still significantly effective) in older people.