"Un Buen Comienzo" [A Good Start] was a professional development program implemented with prekindergarten and kindergarten teachers in Chilean public schools serving low-income families. In a randomized trial, the program showed moderate to large impacts on classroom quality but no impacts on targeted child outcomes. To unpack these findings, we examined intervention fidelity (IF) in both treatment and control groups. Specifically, the study examined (a) whether teachers in the treatment group showed greater fidelity to teaching practices prescribed by the intervention, as measured by dosage and adherence, at the end of prekindergarten and the end of kindergarten, than their control-group counterparts; and (b) whether language and literacy instructional dosage predicted gains in children's language and literacy outcomes at the end of prekindergarten and the end of kindergarten. IF data were coded minute by minute from videotapes of study classrooms collected at the beginning of prekindergarten, the end of prekindergarten, and the end of kindergarten. There were significant and large impacts of the intervention on dosage and adherence. Additionally, small but statistically significant associations were found between 2-year accumulated overall program dosage and children's reading and writing skills at the end of kindergarten. Results reveal that teacher practices changed in response to the intervention, but that increases in time spent on language and literacy instruction were not very substantial. This pattern may explain the absence of an overall impact on children's skills in the experimental study. Findings have implications for the design of both future preschool interventions and IF studies.