Fiabilité, validité et pertinence-critères de la recherche sur l'enseignement de la mathématique

The terms reliability and validity here are used in a broader context than that of testing. Educational research often stresses reliability, that is internal consistency at the expense of validity and relevance. How research may start, is shown by a study of the IREM of Grenoble, which exhibits the clash between text and context, in particular the magic context. As medicine did in the past, educational research may start with phenomenological diagnosis but then progress to etiology in order finally to focus on prevention, which requires a more profound understanding of individual development. Meanwhile we are faced with a new aspect of research: legitimation.

---

---

Conférence faite le 29 novembre 1980 à l'UER de Didactique des Disciplines, Université Paris VII.