A note on the measurement of reversals of perspective

The contradictory findings in the classical literature on the rate of reversal of perspective illusions stem, at least partly, from a failure to isolate three major sources of variance: set, subjects, and sequence. This paper provides a model for the statistical isolation of these variables. It consists of the square-root transformation of reversible data which appear to constitute a Poisson distribution. Such a transformation permits the full utilization of the analysis of variance.