Abstract
Application of a design of multiple time series to a case of intervention in two classrooms of primary education. In the designs of interrupted time series carried out with a single subject, two main difficulties appear. The first one lies in the fact that it isn't quite clear whether the changes, which appear in the execution, are due to the circumstances introduced by the investigator (treatment) or to other circumstances that may occur at the same time. The second one consists in the difficult generalization of the results obtained from a single subject. In this work, we present a case to which we have applied the extension of the inter rupted time series design analysis technique (ARIMA models) (Vallejo & Escudero, 1998) with slight modifications, and extending it to the use of several answer variables registered during the treatment.