Social-communicative abilities are important intervention goals for children with ASD. Research shows that interventions targeting these abilities should be intensive and should promote generalization of learnt skills. Parents and teachers are ideal intervention agents to stimulate social-communicative abilities in a natural context. Moreover, training natural interaction partners provides a cost-effective alternative to intensive one-on-one training by therapists. Several studies have shown that parents can effectively use intervention techniques to improve the social-communicative abilities of their children. There is also some evidence for teacher training. However more research is needed that compares the effect of training parents and teachers. The present study aims to look at the effect of project ImPACT, an intervention program to stimulate social-communicative abilities in children with ASD, which was proven to be effective for parent-training. We aim to address four research questions. First, we want to assess whether this program can be effectively used by teachers. The second goal is to compare the effect of training parents and teachers on the progress children make in their social-communicative abilities. The third goal is to study if a combined approach with training of parents and teachers is beneficial compared to training in a single context. The fourth aim is to see how skills learnt in one context, generalize to the other context. These questions are addressed through a combination of a multiple baseline and alternating treatments design. Participants are four children with ASD, their parents and teachers. The children attend special education and are between 2.5 and 6 years old. Treatment effect is assessed every two weeks through video-observations of parent-child and parent-teacher interaction at home and in school. Communication, social engagement, imitation and play skills of the children and parent and teacher synchronous and asynchronous behavior are coded from the video. Results are analyzed through visual analysis of the graphs and by computing effect sizes for the difference between several phases of the intervention.