Parental language input is essential for children’s early language development, yet substantial disparities persist in both its quantity and quality. Most early language interventions adopt an explicit theory of change, educating parents about why and how to talk with their children, but such approaches tend to be resource-intensive, which limit their potential for scalability. The present study examines Chat2Learn, a low-cost, text-messaging tool that delivers open-ended prompts designed to elicit decontextualized conversations among parent-child dyads. Grounded in Cognitive Load Theory, Chat2Learn aims to reduce parents’ extraneous load of generating conversational topics, thereby increasing opportunities for rich parent–child interactions. Participants were 63 parent–child dyads (children ages 3–6; 65.1% girls) recruited from preschool programs serving families in under-resourced neighborhoods in Chicago. They were randomly assigned to an experimental condition, where parents had access to Chat2Learn, or a business-as-usual control condition, during a 10-minute unstructured waiting period. Results showed that parents in the Chat2Learn condition spent less time inattentive or withdrawn during interactions (p = .05). Chat2Learn parents also produced higher-quality language input, including longer utterances, more unique words, and more open-ended questions, compared to controls (all p’s < .05). No immediate differences emerged in children’s language output. These findings provide a proof-of-concept that an intervention that takes a learning by doing approach in providing concrete support for parents with ‘what to talk about’ rather than abstract training in ‘why” and “how’ to talk can positively shift parent–child interaction patterns.

More on this topic

BFI Working Paper·Nov 20, 2025

Social Dynamics of AI Adoption

Leonardo Bursztyn, Alex Imas, Rafael Jiménez-Durán, Aaron Leonard, and Christopher Roth
Topics: K-12 Education, Technology & Innovation
BFI Working Paper·Oct 13, 2025

Introducing The SPEAK: A Scalable Computer- Adaptive Tool to Measure Knowledge of Early Human Development

Caroline Gaudreau, Dani Levine, John List, and Dana Suskind
Topics: Early Childhood Education
BFI Working Paper·Sep 30, 2025

A Study of the Microdynamics of Early Childhood Learning

James Heckman and Jin Zhou
Topics: Early Childhood Education