Although early behavioral intervention (EBI) is considered as empirically-supported for children with autism, estimating treatment prognosis is a challenge for practitioners. The purpose of this paper is to present an integrated series of three experiments that examined the differential effectiveness of EBI delivered to 233 children with autism in a community setting. The first experiment used latent growth curves analysis to examine changes in autistic symptoms and adaptive functioning, and identify potential predictors of changes. The intensity of intervention, age at enrolment, IQ and autistic symptoms were associated either with progress during the intervention or maintenance during the follow-up. The second experiment used latent profile analysis to identify more homogenous subgroups in the sample and investigate their associations with sociodemographic characteristics and response to EBI. We found four profiles that all made progress during the intervention, with varying magnitudes of change. The last experiment compared five machine learning algorithms in estimating prognosis of children receiving EBI on adaptive functioning and autistic symptoms. Each algorithm produced better predictions than random sampling. Due to the limitations of observational methods, the results of these experiments must be interpreted with caution, but they support the need for further research on differential effectiveness.