Statement of the problem Due to the increasing autonomy, schools are held responsible for developing and guaranteeing educational quality. They are expected to gather information on their results and functioning, to relate that information to input and context characteristics, and to use it for school development. Information on pupils’ achievement provided by School (performance) feedback (SF) systems can be used as starting points for self-evaluation and quality assurance (Visscher & Coe, 2003). It can help schools to see where their strengths and weaknesses lie. Yet our knowledge base of how schools use SF and what the facilitating or hindering conditions are, is at present somewhat limited. This paper is part of a research project addressing the question which feedback modes (in terms of content and presentation) are most useful for schools and which conditions within schools promote the effective use of school feedback. The SF delivered to 200 primary schools by the SiBO project (School Trajectories in Primary Education) in Flanders is used as an instructive case. That feedback contains information on pupils’ learning gains and the schools’ value added from a longitudinal perspective involving data on mathematics and Dutch language achievement. The present paper focuses on the following questions: 1. What problems do head teachers experience in interpreting SF? What characteristics of the feedback system (with regard to content, presentation and support) contribute to a correct interpretation by schools? 2. How do schools use SF? What problems do head teachers experience in using SF? Theoretical framework Two aspects of working with SF are of key importance: the interpretation of the school reports and the use of it as a component of school development. Both are situated in a comprehensive framework, based on Visscher’s model (Visscher, 2002). Various models for effective SF (Fitz-Gibbon & Tymms, 2002; Van Petegem & Vanhoof, 2004; Schildkamp, 2007) and for the inventarization of SF use (Rossi, 1993; Liket, 1992) are integrated into this theoretical framework. In addition to answering the above research questions, a further explicit objective of this paper is to contribute to theory development.

Method In order to validate and further develop the existing theoretical framework in confrontation with empirical data, a qualitative approach was adopted (Maso & Smaling, 1998). Data were collected through semi-structured in-depth interviews with 16 school principals who had received a school feedback report a few months before. We used theory-driven sampling (Silverman, 2000) representing a wide diversity of school contexts. The coding process was carried out by two researchers using ATLAS.TI. An inter-rater reliability of 0.9 was established (Miles & Huberman, 1994).

Results and discussion Only few effects and a rather limited use of school reports were reported by the respondents. Explanations for this and details on the particular ways our respondents have used the SF reports will be presented in the paper. Most school principals state that reading the report requires a great deal of time and effort. A number of them say that the feedback is too difficult, due to different reasons. A comprehension test conducted during the interview revealed misconceptions regarding both the interpretation of concepts and graphical representations in the report. Most school principals indicated that they did lack sufficient knowledge and appropriate skills to arrive at a correct interpretation. They also expressed a need for extra support going beyond the interpretation of the SF report, addressing how SF can be integrated into school policy.