Attendance at a parenting program is stable, but organizers do not know why some families return and others stop coming. Which evidence would add the most useful context?