Both advocates of competition as a means to better school performance and economics-based research on this issue assume a direct relationship between a more competitive market structure (in terms of the number and concentration of schools in a local market) and better school performance. This is an application to schools of the structure-conduct-performance model. It is assumed that head teachers and other professionals are motivated solely by self-interest, so that lack of competition results in x-inefficiency. However, if educational professionals are motivated by other considerations, in particular their values and beliefs, there is no automatic link between competitive structure and forms of competitive conduct that lead to better school performance. Since it is competitive conduct that affects school performance, the hypothesis of a postitive relationship between competition and performance is investigated in this study by collecting and analysing data on perceptions of competitive conduct from a survey of headteachers. An analysis of these data combined with administrative data finds that: (a) the two measures of perceived competition are only weakly related to measures of structural competition; (b) the number of perceived competitors is positively and significantly related to school performance in terms of the percentage of students obtaining 5 or more grades A* to C at GCSE but not the percentage obtaining 5 + A*--G grades.