Abstract:
|
School mobility, the proportion of non-promotional school changes among students, can reach over 60% in US schools, but little research has been done into how this mobility affects interventional research in schools. This contribution examines how contamination caused by student transfer between schools under intervention and control can affect stepped-wedge cluster randomized trials. The stepped-wedge design employs randomized, sequential roll-out of the studied intervention and all participating clusters ultimately receive the intervention, making stepped-wedge a common trial design choice for educational settings. Simulation studies were used to assess the impacts of changes in mobility rate on power and Type I error. Additional changes to school and district size, ICC, and study length were employed to assess how altering study design may protect against power loss due to contamination.
|