Purpose <p>Applied behavioral analysis is the most common intervention for autistic children. It has long been proposed that more hours per week in ABA produce better outcomes. ABA practice guidelines typically recommend 30–40&#xa0;h per week for young autistic children. While some studies support a dosage effect, other studies find no predictive relationship between ABA treatment hours and a range of child outcomes. Thus, the dosage question remains open, leaving parents and providers unsure of the ideal dosage of ABA for an autistic child to make meaningful progress. Real world data (RWD) can fill the gap between community treatment and controlled research often conducted by universities.</p> Methods <p>In this study, we examined the RWD data from 725 autistic children receiving ABA as part of community-based care. Using longitudinal multilevel modeling, we examined the relationship between average hours of ABA per month and change on proximal and distal measures: adaptive behavior, goal attainment, and dangerous behavior.</p> Results <p>Dosage was a significant predictor of goal attainment, indicating that clients who spent more time in treatment met more goals. Conversely, the relationship was negative for adaptive behavior and communication, such that higher hours predicted slower progress in these domains. Finally, dangerous behavior decreased over time and was independent of hours. Secondary analyses demonstrated that communication abilities at baseline predicted treatment outcomes above and beyond dosage, except for dangerous behavior.</p> Conclusion <p>While more research is needed, this study suggests that practice guidelines should temper language around dosage and ABA.</p>