So what you’re basically saying is that EDT is vulnerable to Simpson’s Paradox?
But then, aren’t all conclusions drawn from incomplete sets of data potentially at risk from unobserved causations? And complete sets of data are ridiculously hard (if not impossible) to obtain anyway.
So what you’re basically saying is that EDT is vulnerable to Simpson’s Paradox?
But then, aren’t all conclusions drawn from incomplete sets of data potentially at risk from unobserved causations? And complete sets of data are ridiculously hard (if not impossible) to obtain anyway.