Abstract: This study investigates the effects of a collaborative multimodal writing intervention in different settings on the development of writing skills of beginner learners of Dutch as a second language. The study extends previous research by using multiple measures to assess writing development and by comparing the impact of different settings, i.e., a task-based in-school intervention and an out-of-school leisure intervention. All participants in both interventions (N = 52) completed two writing tasks before and after the intervention (in a pretestposttest design). Participants in a non-intervention group (N = 32)did the same before and after a two-week holiday in which they did not receive writing instruction. Participants' output was assessed by analytic CAF-measures as well as for communicative effectiveness. Multilevel analyses revealed significantly greater growth for both intervention groups compared to the non-intervention group. The in-school group showed growth for six measures (complexity and text length on one task and communicative effectiveness, content, lexical diversity and text length on the other). Also the out-of-school group revealed greater gains for six measures (complexity on one task and content, two measures of syntactic complexity, lexical diversity and text length on the other). In all, this study shows that multimodal composition can promote writing of beginner L2 learners, regardless of the exact context in which the writing takes place. Furthermore, the study shows the importance of using multiple measures and multiple tasks to evaluate writing development.