In its initial version, the questionnaire comprised 221 items, which could be reduced
to 98 in the course of scale analysis (see Table 2). Item reduction was effected
on the basis of the usual criteria according to classical test theory requirements
(e.g., Lienert & Raatz, 1998). Items with too high or too low item diffi culty (P < .20
or < .80 respectively), too low selectivity (< .35) or too low a loading on the respective
factor of the factor analysis (< .40) were dropped. Retest reliability was collected
over a period of 14 to 18 weeks. All results were satisfactory. They are within the