On the use of latin squares in psychology.

Abstract

The author contends that the use of the latin square is seldom appropriate in psychological research. "It is defensible only in those rare instances where one has sound a priori reasons for believing that the interactions are zero.".