Background: Angelman syndrome (AS) is a neurogenetic disorder characterized by severe intellectual disability, lack of speech, and low threshold for laughter; it is considered a "syndromic" form of autism spectrum disorder (ASD). Previous studies have indicated overlap of ASD and AS, primarily in individuals with larger (approximately 6 Mb) Class I deletions of chromosome 15q11-13. Questions remain regarding whether intellectual disability solely contributes to ASD features in AS and how ASD features in AS change over time. In this study, we used a dimensional approach to examine ASD symptom severity in individuals with AS Class I versus Class II deletions within the context of cognitive development over time. Methods: A total of 17 participants with a larger, Class I deletion and 25 participants with a smaller Class II deletion (approximately 5 Mb) were enrolled (age range = 2-25 years; 5 years 5 months). Standardized measures of cognition, language, motor skills, adaptive skills, maladaptive behavior, autism, and sensory-seeking behaviors/aversions were given at baseline and after 12 months. Results: Despite equivalent cognition and adaptive behavior, the results of repeated measures analyses of variance indicate that participants with Class I deletions have greater impairment in social affect (F = 8.65; p = 0.006) and more repetitive behaviors (F = 7.92; p = 0.008) compared to participants with Class II deletions. Although both groups improve in cognition over time, differences in ASD behaviors persist. Conclusions: Despite a lack of differences in cognition or adaptive behavior, individuals with Class I deletions have greater severity in ASD features and sensory aversions that remain over time. There are four genes ("NIPA 1," "NIPA 2," "CYFIP1," and "GCP5") missing in Class I and present in Class II deletions, one or more of which may have a role in modifying the severity of social affect impairment, and level of restricted/repetitive behaviors in AS. Our results also suggest the utility of a dimensional, longitudinal approach to the assessment of ASD features in populations of individuals who are low functioning.