Previous research has shown how home and parental characteristics support or hinder the development of children’s self-regulation in the family context. There have only been limited attempts to understand these mechanisms in early childhood education settings. This study used the NICHD Study of Early Child Care (when participating children were 36-months old) to examine the relations among various aspects of the early childhood education setting, the interactions in the setting, and children’s self-regulation in center-based and home-based settings. Structural equation modeling was used to test a model proposing the deconstruction of early childhood education quality into structural (i.e., environmental and caregiver characteristics) and process quality components (i.e., positive and negative interactions) and to examine these as predictors of three-years old children’s self-regulation abilities. A meditational model was tested in which positive and negative interactions in the classroom mediated the relations between the structural characteristics and self-regulation. There were three important findings. First, although there were no consistent patterns of associations between structural features and self-regulation across the two types of care, there were more significant relationships in home-based care compared to center-based care. These findings showed that the home-based caregiver characteristics were more closely tied to the processes in the classroom than those characteristics of caregivers in center care. Second, both positive and negative caregiving were associated with children’s compliance, which suggested that compliance may have been influenced differently by process quality compared to other self-regulation measures, such as self-control and emotion-, behavior-, and attention-regulation. It may be that high rates of compliance may be markers of highly restrictive caregiving rather than the result of good quality caregiving. Third, there were very few significant relationships between process quality measures and children’s self-regulation measures, which suggested that commonly used process quality measures may not be capturing the processes that are most important for the development of self-regulation. / text