I have a continuous predictor and a time-to-event binary outcome in two population defined by their medical history.
My hypothesis is that the predictor is [a much stronger predictor/explains more variance/has more explanatory power] in one of the populations than in the other.
How should I go about quantifying this and testing a null hypothesis of “equal predictive power”?
The interpretation is critical. I need a strong enough message for a reader to be able to say “this predictor really says everything in population A, but is only part of the story in population B”.
Would be grateful for any advice.