160 secondary school teachers were administered questionnaires on attributions for success and failure in students familiar to them. Student profiles were obtained by a 2 (success or failure) x 2 (extreme or moderate grades) design. The list of attributions employs classic dimensions (internal/external), stable/unstable, controlable/non controlable) plus a new dimension (academic/non academic) shown to be relevant in two preliminary surveys. The results indicate an interaction between the two independent variables as a function of the internal/external dimension. Responses characterizing for success and failure reveala specific dimension termed subject/object, that parallels the a-priori academic/non academic dimension.Studies of perceived causality of success and failure have shown that the internal/external dimension cannot totally account for attributions. As Weiner (1972, 1979) points out, individual success and failure can be ascribed to four types of causes: ability, effort, task difficulty and luck. Causes can be classified along three dimensions: locus (internal or external cause), stablity (a stable or unstable cause) and control (a cause that can be controlled by the actor or is beyond control). For example, Weiner (1972) defines ability as an internal, stable and uncontrollable cause, luck as an external, unstable, and uncontrollable cause. The criteria used to classify causes into dimensions are not always made explicit and have given rise to a variety of interpretations. The internal/external dimension is often ambiguous as well: financial reward, which appears to be an external source of reinforcement, can be presented in such a way This work has received the financial support of INElOP and Laboratory of Social Psychology 'Social Representations and ideological process' of Paris X-Nanterre University.