Using AI-Assisted Parenting Guidance to Promote Early Childhood Development: A Randomized Experiment in Rural China

How to cite

Shi, Yaojiang; Yue, Ai. 2026. "Using AI-Assisted Parenting Guidance to Promote Early Childhood Development: A Randomized Experiment in Rural China." AEA RCT Registry. AEARCTR-0018272. https://doi.org/10.1257/rct.18272-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Yaojiang Shi* , Ai Yue

* Primary investigator

Registry ID
AEARCTR-0018272
DOI
https://doi.org/10.1257/rct.18272-1.0
First registered
Last updated
Status
In development
Trial dates
2026-04-01 – 2026-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18272

Abstract

While experimental evidence supports the effectiveness of parenting interventions, the impact of Artificial Intelligence (AI)-based parenting interventions on caregivers and children remains understudied. In this study, we aim to evaluate whether AI-assisted parenting guidance can improve caregivers’ parenting mental health, stress, and self-efficacy, as well as early childhood developmental outcomes in rural China. In a pre-existing center-based parenting program in rural China, we randomly assign 20 centers (over 1000 registered caregiver-child dyads) to treatment and control arms. In the treatment centers, caregivers are encouraged to use an AI application integrated with a comprehensive curriculum covering six domains on early childhood development: psychological development, nutrition and feeding, childcare, safety care, first aid for accidental injuries, and caregiver-child interaction and early reading. This curriculum can well complement the psychosocial stimulation services provided by the parenting centers. The control group receives standard parenting center services.

Experimental design

Randomized controlled trial

Design details

Randomization method
Randomization was conducted by computer using STATA 18.0 and the “minmax method” (Bruhn and McKenzie, 2009)
Randomization unit
Pre-existing parenting center
Clusters
At least 20 parenting centers
Observations
At least 1023 caregiver-child dyads
Arms
10 parenting centers in the treatment group; 10 in control group
Primary outcomes
Caregiver's parenting stress measured using PSI-SF (Parenting Stress Index-Short Form), mental health measured using DASS-21 (Depression Anxiety Stress Scale), and parenting self-efficacy measured using PSOC (Parenting Sense of Competence)
Secondary outcomes
Child social-emotional outcome measured by ASQ: SE, and language outcome measured by ASQ language sub-scale

Cited by

No citation data available for this registration yet.

Keywords