Résumé
Precise assessment of the different stages of the word problem solving process is an important clinical issue for speech and language therapists. However, existing tools still lack of precise and reliable evaluation of integration, which correspond to the mental representation of the problem. The aim of this study is to develop and experiment a protocol for integration assessment of additive word problem in elementary school children. Our protocol is drawn up with a dual modality-verbal and visual-in order to emancipate from the potential bias of the verbal constraint. Experiment with children without any difficulties (CTRL) showed the relevance of the verbal modality: demand of paraphrasing shows a good construct validity. Visual modality, which consists of demanding a frieze designation among distractors, presents some structural limitations, that should be reworked in order to reassess its relevance. Experiments with children presenting language disorders (TLO) or mathematics difficulties (TCM) revealed integration impairments in these two groups, but children with TLO present a greater deficit. Besides the quantitative analysis of TLO and TCM performances, additional markers such as the type of paraphrasing and the linkage between paraphrasing and execution help identify different deficit profiles between TLO and TCM facing integration evaluation. Ours results did not allow to verify the relevance of a visual modality intended for children with language disorders (TLO). All these results must be tempered by the small sample size of our different groups, and by the limitations we pointed concerning the construction of our visual modality protocol.