Abstract :
Early Behavioral Interventions (EBIs) are a cornerstone of treatment for autism spectrum disorder, yet their differential impact on the development of generalized social skills remains a critical area of inquiry. An important aspect of the treatment of autism spectrum disorder is Early Behavioral Interventions (EBIs), the effectiveness of which in the formation of generalized social skills is still a burning subject of research. This quasi-experimental study involved comparison of the efficacy of two models of EBI a structured Early Intensive Behavioral Intervention (EIBI) and a Naturalistic Developmental Behavioral Intervention (NDBI) on the development of social skills in 60 preschool-aged children with autism. The subjects were provided with six months of the designated intervention, 20 hours per week. Social skills were measured at baseline, post intervention and at three months follow-up by the Vineland-3 Socialization domain, the Social Responsiveness Scale-2 (SRS-2) and direct observation of social initiations. Mixed-model ANOVAs revealed that Group x Time interaction was significant (asterisked) at the size of *p below .05, but that that NDBI group showed considerably higher increases in socialization and social initiations in both post-test and follow-up. Importantly, NDBI group demonstrated a further increase in the follow-up period, and the gains of the EIBI group reached a level. The moderation results indicated that the results of the EIBI group were heavily influenced by the baseline non-verbal IQ and language level, but the NDBI model portrayed greater success in a wider variety of child traits. The results indicate that NDBIs prove to be more successful, compared to the traditional EIBI, in encouraging long-term, externalized social abilities and possibly offer a more inclusive intervention to a heterogeneous clinical population. It can be recommended to focus on NDBI frameworks, to provide personalized planning of interventions using pre-treatment assessment, and require explicit generalization programming.