Ensino por Múltiplos Exemplares: Revisão sistemática de estudos experimentais

Abstract

The multiple exemplar instruction procedure (MEI) is characterized by the rapid and randomized rotation of training of different verbal repertoires in consecutive trails and has been used in the investigation of induction of 'bidirectional naming' (BiN) and promotion of interdependence between verbal repertoires (IVR). This work performed a systematic review of experimental studies that have used MEI in the conditions mentioned above, with the aim of identify and analyze the profile of the participants, tests’ structure and tested repertoires, the aim and structure of the MEI training, and the obtained results. A search in the databases Wiley, PUBMED, SciELO, Web of Science and Scopus, using the term 'multiple exemplar', resulted in 24 articles selected for analysis (12 of them used MEI to induce BiN and 12 to promote IRV). In general, it was observed that: (1) the majority of participants were children with atypical development; (2) the most used test structures were: pre/post-test through training identity matching to sample with the tact of the sample stimulus by the experimenter (IDMTS+tact) up to performance criteria, followed by tests of auditory-visual matching to sample (AVMTS), pure and impure tacts (for BiN), and pre/post-test of the repertoires trained in the MEI (e.g., tact, mand) and pre/post-test through training a repertoire with one set and testing of another repertoire with the same set, and vice versa (for IRV); and the main tested repertoires were 'speaker unidirectional naming', mand and pure tact; (3) the most used MEI structure were: the rotation of trials with IDMTS+tact, AVMTS, pure and impure tacts (for BiN), and rotation of trials with mands and pure tact (for IVR); and (4) the results were generally positive for establishing IVR and ‘speaker unidirectional naming’. It is argued that adequate tests of ‘joint bidirectional naming’ or ‘incidental bidirectional naming’ have not been conducted, and research is suggested to overcome this limitation and to advance knowledge about IVR.