Childhood socioeconomic status (SES) is associated with persistent academic achievement gaps, which necessitates evidence-based, scalable interventions to improve children's outcomes. The present study reports results from a replication and extension of a family-based training program previously found to improve cognitive development in lower-SES preschoolers. One hundred and one primarily low-SES families with 107 children aged 4-7 years were randomly assigned to the intervention or passive control group. Intent-to-treat regression models revealed that children whose families were assigned to the intervention group did not exhibit significant benefit on composite measures of nonverbal IQ, executive functioning, or language skills, though post-hoc analyses suggested marginal improvement on the fluid reasoning subcomponent of nonverbal IQ. Treatment-on-treated models revealed a significant positive effect of intervention attendance on fluid reasoning and a negative effect on vocabulary. We discuss potential causes for the non-replication, including differences in the sample composition, size, and assessment choices. Results suggest the need to more broadly assess scalable interventions with varying populations and ensure appropriate cultural and geographical adaptations to achieve maximum benefits for children from diverse backgrounds.