Still another step was included in the scale constructions. The fact that an item was
endorsed differently by the group of 724 Minnesota normal than by the patients from various clinical populations did not necessarily indicate that it could be used success- fully for clinical screening purposes. Thus, an at tempt was made to cross-validate the scales by selecting a new group of normal and comparing their responses with a different group of clinical patients.
The items that still provided significant differences between these groups were selected for the final version of the scales.
It was reasoned, then, that these items and the scales comprising these items would be valid for differential diagnosis in actual clinical settings.