This thesis explores the diagnostic, predictive and informative values of Dynamic Assessment (DA) in children with Typical Development (TD) and Developmental Language Disorder (DLD). DA evaluates children’s learning potential by incorporating teaching within the task itself, either through a graduated prompts procedure or a pretest/teaching/posttest format. Unlike traditional static assessments, DA is particularly effective for bilingual children and provides valuable insight into children's learning processes.

The thesis is organized into three sections, each addressing one of DA’s functions: diagnostic, predictive, and informative.

Section 1: Diagnostic Value

Three studies examined DA’s accuracy in distinguishing children with DLD from their peers with TD.

• Study 1 involved a DA of morphosyntax with 79 children (37 with DLD), assessing six structures. Children with DLD required more prompts, with ROC curves confirming strong diagnostic power. Interestingly, a link was found between performance in morphosyntax and non-verbal Analogical Reasoning (AR).
• Study 2 used a narrative DA in a pretest/teaching/posttest format with 60 children (30 with DLD). Only the experimental group showed gains (and not the control one, not participating in the teaching phase), indicating specific learning effects and demonstrating DA’s ability to differentiate between DLD and TD without bias against bilinguals.
• Study 3 used a narrative DA in a pretest/teaching/posttest format with 192 children (88 with DLD). Children with TD showed both direct and generalization effects, while those with DLD only showed direct learning effects, highlighting known generalization difficulties. The task again effectively distinguished children with TD and DLD, without disadvantaging bilinguals.

Section 2: Predictive Value

• Study 4 followed 21 children with DLD who completed the narrative DA from Study 3. A delayed posttest 12 weeks later showed that DA posttest scores predicted future macrostructural narrative abilities, while microstructural elements were better predicted by pretest scores. This suggests DA offers added predictive power beyond static measures.

Section 3: Informative Value

Two studies explored DA’s ability to reveal learning trajectories and cognitive mechanisms.

• Study 5 included 157 children (40 with DLD) who learned invented morphological rules. Learning curves showed that TD children quickly stabilized performance, while DLD children required more support and showed less stable learning. Non-verbal AR influenced performance only in TD children.

• Study 6 assessed 195 children (98 with DLD) on static and dynamic tasks of AR and morphosyntax. No relation appeared in static tasks, but a significant link emerged in dynamic tasks for children with DLD. This suggests that they relied on perceptual rather than rule-based learning, aligning with the Procedural Deficit Hypothesis.

Conclusion

This thesis provides novel French-language DA tools and supports the idea that children with DLD face difficulties in language learning. Across six studies and 802 participants (331 with DLD), the findings reinforce the relevance of DA for both research and clinical practice.