Not naive. It means something like "how much variance can we explain on the ethical side." But the problem is circular. We can explain the majority of the variance for the probes we've written. But are our probes comprehensively testing the reasons medics choose one patient over another?