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.