On the Procrustean analogue of individual differences scaling (INDSCAL)