Experimental Design

This study takes place in the Udaipur district of Rajasthan, India. For this study we used a cluster-randomized design with stratification by geographic area (the Udaipur district is comprised of 5 blocks). Because daycare programs are made available to the entire community it was not desirable to randomize at the level of individual families because of potential contamination. Instead, we randomized at the level of the village hamlet (hamlets are clusters of houses that surround a village center but constitute an independent unit). Based on focus group data and prior knowledge of heterogeneity in the Udaipur district, we chose to stratify by block group to prevent variations in the distributions of blocks across treatment groups (e.g., if women in treated hamlets were more likely to reside in blocks with more economic opportunities). Because four of the five blocks contained an odd number of hamlets, the randomization was done so that two of the four blocks would have an additional treated hamlet and the remaining two would have an additional control hamlet.
Due to the nature of the intervention it was not possible to blind study personnel or participants after the implementation of the intervention. However, we concealed the allocation of hamlets to treatment or control status until after the baseline survey in order to minimize opportunities for bias in recruitment of participants and the baseline survey.

Intervention

The intervention established, in each of 80 treated clusters, a program of full time, affordable, community-run daycare. The daycare centers are called balwadis, and provide full day child-care, nutritious food and supplements, basic medicines, and preschool education to children one to six years old. Additionally, the balwadis support increasing child immunization coverage by maintaining immunization records and following-up with parents and government nurses. The centers are operated by trained local women, called sanchalikas, who are hired and trained by a local non-governmental organization called Seva Mandir. The sanchalikas meet quarterly with parents to provide feedback on child progress. To encourage community enrollment in the program, the balwadi program was also accompanied by a household marketing campaign.

Monitoring

Fieldworkers are conducting monthly site visits to each balwadi to verify that the balwadi is operating and the structure is adequate, and that the sanchalika is providing sufficient food and education to the children. Data from the monthly site visits on the number of days each balwadi is operating and any reasons for temporary closings will be derived from a camera-monitoring system. Sanchalikas are required to take three self-timed photos each day, which are used to evaluate the number of days when the balwadi was run for at least 6 hours, which is considered a full day of operation. We will use the monitoring data to assess the impact of treatment assignment on actual receipt and use of daycare among women and children.