such as discussing the forces involved as the same object sits in the student's hand, hangs from a spring, and as the object is pushed off the edge of the table. Throughout the unit, the teacher listens carefully to his students' responses and explanations. Without an understanding of both student learning and the science involved, upon hearing the proper terms from his students, he may have proceeded with his unit with the impression that the students shared a scientific understanding of force (for a class transcript and analysis by the teacher, see Minstrell, 1992).
The data produced from the variety of assessments illustrated in the vignettes are not only useful for the teachers but also as essential tools in helping students to realize where they stand in relation to their goals. Thus for the students, the journals with the teacher 's comments added, serve as a repository for one form of feedback so they can maintain a continuing record of their work and progress. It is important to emphasize that assigning grades on a student' s work does not help them to grasp what it takes on their part to understand something more accurately or deeply. Comments on a student 's work that indicate specific actions to close the gap between the student's current understanding and the desired goal provide crucial help if the student takes them seriously. There is well-researched evidence that grades on student work do not help learning in the way that specific comments do. The same research shows that students generally look only at the grades and take little notice of the comments if provided (Butler, 1987). The opportunity that Ms. R's students had to design, build, and then rebuild instruments based on their trials gives them a chance to make good use of feedback to improve their piece of work.
Providing information to students is not solely a cognitive exchange. It is intertwined with issues of affect, motivation, self-esteem, self-attribution, self-concept, self-efficacy, and one's beliefs about the nature of learning. From many studies in this area (Butler, 1988; Butler & Neuman, 1995; Cameron & Pierce, 1994; Kluger & deNisi, 1996), a further generalization emerges. This is the distinction between feedback that emphasizes learning goals and the associated targets and feedback that focuses on self-esteem, often linked to the giving of grades and other reward and punishment schemes. Upon comparison of feedback in experimental studies, it is the feedback about learning goals that shows better learning gains. Feedback of the self-esteem type (trying to make the student feel better, irrespective of the quality of the work) leads less successful students to attribute their shortcomings to lack of ability. The