How to handle the interaction effect has been a topic of much discussion among performance analysts and especially with clients.  Some have defended the usefulness of the interaction effect and have suggested ways to explain it as an independent and significant attribution factor.  Here we present another view of interaction:  it is a spurious effect resulting from incorrect attribution analysis.  The author suggests that interaction is always a component of the selection effort and provides both logical and numerical defenses of this argument.

Debunking the Interaction Myth