This methodological review examines sixteen empirical studies on dynamic assessment of reading comprehension published between 2000 and 2021. Systematic content analysis was used to highlight methodological trends in the publications and analyze construct, consequential, and social validity. Our findings describe the following tendencies: a preponderance of quasi-experimental designs with non-probabilistic sample methods and a predominance of interventionist dynamic testing methods. The research outcomes were grouped in three categories: (b) half of the studies showed that dynamic assessment promoted students’ reading comprehension abilities, (a) more than a quarter demonstrated that students performed significantly better on dynamic testing than on static testing, and (c) a fifth of the studies demonstrated that dynamic assessment could contribute to diagnosis or student classification. Our analysis emphasized opportunities for improvement in construct validity in research focused on dynamic assessment of reading comprehension, including: (1) defining key constructs, (2) providing detailed information on how the constructs were operationalized, (3) explaining how reliability was assessed, and (4) explicitly linking the results to the method used to theory. Most studies focused on the proximal consequential basis of the test used, notably to demonstrate the effectiveness or the feasibility of dynamic testing. Therefore, it may be necessary for future research to examine the distal consequences, notably on students' social, emotional, and cognitive development. Finally, our review underlines the need to better document the social relevance of dynamic assessment in the classroom by giving a voice to students, teachers, and other stakeholders who may be most concerned.