Intervention dosage is foundational to realizing intended impacts but is often variable, particularly when interventions are implemented under real-world conditions. In this study, we examined dosage of small-group emergent literacy intervention experienced by preschool children (n = 154) identified as at risk for later reading difficulties in authentic classroom settings. We documented considerable variability in dosage that was largely due to when instructors stopped offering lessons. Drawing from extant literature and an ecological orientation, we found that instructor factors (i.e., instructor self-efficacy for teaching language and literacy, instructor perception of lesson acceptability, average small-group size) and classroom factors (i.e., classroom teachers' self-efficacy for decision making), but not child factors, significantly predicted children's intervention dosage. Moreover, most variance could be attributed to differences between small groups/instructors rather than individual differences among children. We discuss implications for preschool teachers, administrators, researchers, and intervention developers seeking to better support successful small-group intervention implementation. [This is the online version of an article published in "Journal of Early Intervention."]