Conceptual
Login

Overlapping Distributions and the Limits of Predicting One Person

Makes the variance split visual and consequential. Two distributions whose means differ by a fraction of an SD overlap almost entirely, so knowing a person's group membership barely narrows the guess about that person. This is the node that turns 'the averages differ' into 'and yet I still cannot tell you about this individual', which is the single most-needed move in F.

Questions this Concept answers

  • Why does a difference in group means that is a small fraction of a standard deviation help so little in describing one person?

This Concept is waiting for its first lesson!

Makes the variance split visual and consequential. Two distributions whose means differ by a fraction of an SD overlap almost entirely, so knowing a person's group membership barely narrows the guess about that person. This is the node that turns 'the averages differ' into 'and yet I still cannot tell you about this individual', which is the single most-needed move in F.

Are you a teacher? Sign in to start contributing.

Sign In