Yes, that is a notorious test deviser's fault. For a while I have used a Dutch translation of the Young Schema-Questionnaire version 2 in which the 205 items of the 16 subscales had been well-mixed, so that problem had been dealt with. However, I stopped using the YSQ-2 because clients with very serious personality problems scored high in so many schemas at the same time. On an item level the results could be informative (however, rarely surprising after a few sessions) but the subscale scores were rather non-discriminative. In a CFA by Rijkeboer & van den Berg (2006, Cognitive Therapy and Research, 30, 263-278) 62 of the 120 factor correlations were equal to or higher than .70. The scales are a heterogeneous mix of past experiences, upbringing styles, present attitudes and beliefs, present experiences, self-concept etc. I doubt whether the various versions of the YSQ have ever been adequately psychometrically analyzed.

However, that does not mean that I do not endorse Young's ideas.

Thanks Peter, that is very helpful indeed. I cannot understand the reason for which items were ordered in such way! Perhaps they found it to be more clinically useful despite being psycometric weakness! But as you said, it is not the case even clinically.

The lack of discriminative power as indicated by high levels of intercorrelations have been reported, and it is sometimes interpreted as a sign of internal validity of the scale!

Factor analytic findings, both EFA & CFA, tend to be inconclusive in general.

Thanks again for your insightful points.