This study aimed at identifying the determining factor in making teacher-given rubric feedback effective on student performance (planning experiments in science class), motivation, and calibration accuracy. In a pre-post experimental design, secondary school students (N = 120) were randomly assigned to three different variations of rubric feedback incrementally graded with respect to the quantity of included feedback information: in condition 1 the rubric was handed out to the students to make learning goals transparent (transparency-information). In condition 2 the rubric as well as information on the students' actual task performance was provided (individual performance-information). Condition 3 included not only the rubric and actual task performance information but also individual cues on how to proceed (individual performance-improvement-information). Students who received improvement information showed a significantly better performance in planning scientific experiments, perceived themselves as being more competent, and were also more accurate in their self-evaluative performance judgements. This calibration accuracy could be identified to be a partial mediator of the positive effect of rubric feedback on performance. Thus, results indicate that the crucial variable in making teachers rubric feedback effective appears to be the successful use of performance-improvement-information. Implications for educational research and practice are discussed.