ABSTRACT
Active Support is a staff practice that increases people with intellectual disabilities’ engagement in purposeful activities and social interactions, but variability in its implementation has been found across accommodation services. To explain this variability, researchers have examined predictors of either the quality of Active Support or levels of engagement but have not analysed them together. The aim of this study was to examine predictors of the quality of Active Support and in turn levels of engagement. Multilevel path analysis was used to test models comprising predictors of two dimensions of Active Support—Technical and Interpersonal—as well as engagement in social and non-social activities. Data were available for 573 people with intellectual disabilities across 198 accommodation services. The multilevel path analysis showed that strong practice leadership, culture of Supporting Well-Being, smaller sized services and higher adaptive behaviour of people with intellectual disabilities predicted Technical and Interpersonal Active Support. In turn, Technical Active Support and adaptive behaviour predicted non-social engagement. Interpersonal Active Support, adaptive behaviour, age, and two dimensions of challenging behaviour predicted social engagement. The findings confirm and extend previous evidence about predictors of the quality of Active Support and levels of engagement. The path models provide evidence for previously assumed connections between key variables important for the quality of Active Support and higher levels of engagement. The path models could provide guidance in what to focus on to implement and improve Active Support and engagement in services.