CODIMENSIONALITY WITHOUT HIGH CORRELATION

The product moment correlation of discrete variables is depressed by differences in the variables' distributions which are consistent with their variables are actually codimensional and so to bias convergent validation studies toward rejectring a true hypothesis of (sufficient) convergence. 3 being codimensional. This makes it prone to indicate multidimensionality when variables are actually codimensional and so to bias convergent validation studies toward rejectring a true hypothesis of (sufficient) convergence.