We selected four items per scale to obtain a short, economical version of the questionnaire. We chose the items according to whether they reflect the core of the construct (see Table 1) and kept them linguistically simple enough to be answered easily, quickly, and reliably. The items should show high item-total correlations and neither ceiling nor floor effects. Table 2
displays the means, standard deviations, and internal consistencies of each goal scale separately for the samples of pre-service and professional teachers. The item-total correlations for all items in both samples were almost above the critical value of .30, with just two exceptions. The internal consistencies of the scales with four items each ranged from .56 to .89 in the pre-service teacher sample and from .66 to .84 in the teacher sample (see Table 2).