NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Back to History
Improving Full Immunization Rates in Haryana: Evaluating Incentives and Communication Methods
Initial registration date
October 06, 2016
July 09, 2018 2:58 PM EDT
Other Primary Investigator(s)
J-PAL Global, MIT
J-PAL South Asia at IFMR
J-PAL Global, MIT
Additional Trial Information
The primary purpose of this program is to increase the coverage of full immunization rates among children under 12 months of age in seven poor-performing districts of the state of Haryana, India. This will be done by providing non-cash incentives to parents to motivate them to consistently immunize their children. Previous research conducted in Rajasthan, India shows that lack of demand for immunization may be a key reason for low rates of full immunization, even when supply-side bottlenecks have been removed. Thus, this study will rigorously evaluate the effectiveness of providing incentives to beneficiaries in increasing full first-year routine immunization rates in Haryana. It will also look at the scalability of this approach in improving immunization coverage under the current government framework. Along with the immunization incentives experiment, the study conducts two additional experiments to evaluate various communication strategies the government could use to boost immunization rates. The first of these analyzes the potential role of village social networks: community members are asked to nominate individuals from their village who they consider good at spreading information, trustworthy in the village, or a combination of the two. These nominated 'seeds' will be expected to spread the information that they receive about immunization to other members in their community. This experiment evaluates different approaches that policymakers could use to mobilize well-suited community members to increase immunization rates. The second one, targeted reminders experiment, will provide reminders via text messages to primary caregivers, customized to the vaccination schedule of their children. This experiment is designed to evaluate a low-cost and scalable targeted immunization information system. Both of these experiments will be cross-randomized with the larger incentive experiment, which will shed more light on the efficacy of the interactions of these experiments with the incentive program.
Banerjee, Abhijit et al. 2018. "Improving Full Immunization Rates in Haryana: Evaluating Incentives and Communication Methods." AEA RCT Registry. July 09.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
a. Full first year immunization rate among children 1-3 years of age. Full first year immunization refers to whether a child has received all five vaccines of the first year immunization schedule viz. BCG, DPT/Penta - I, II, III, and Measles.
b. Flow at immunization session sites. This is measured by the volume of children in the session sites. This data will be collected by the Auxiliary Nurses (ANMs) conducting the vaccination sessions using a data entry application designed specifically for this project.
Primary Outcomes (explanation)
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
The program comprises three experiments:
1. Incentives Experiment: This experiment tries to evaluate the impact of incentives to generate demand for child immunization. 140 Primary Health Centers (PHCs), each of which provides health services to about 40 villages, selected from our 7 sample districts were randomly divided in two groups- incentive treatment (group that receives incentives) and incentive control (group that receives no incentives). Incentives will be in the form of mobile credits. Within the incentive group, we further randomize sub-centers (SC), of which there are approximately 5 per PHC, into various levels and slopes of the incentive amount tied with each vaccine: a. low incentive, flat payment for each vaccine b. high incentive, flat payment for each vaccine
c. low incentive, increasing payment (incentive amount increases after the third vaccine)
d. high incentive, increasing payment (incentive amount increases after the third vaccine)
2. Communication Experiment: This experiment, which is cross-randomized with the incentive experiment, will study the effectiveness of using various 'seeds' to diffuse information about vaccination as well as convince individuals to immunize their children. For this experiment, 943 villages will be divided into five groups: control group, random seeds group, gossip seeds group, trust seeds group, and trusted gossip seeds group. Each village will have 6 seeds, each of which are identified by seeking nominations from villagers.
3. Targeted Reminders Experiment: The targeted reminders experiment will be conducted across the program districts. Half of the villages will be randomly selected to receive text messages that are tailored to the immunization schedule of the eligible child based on their date of birth collected during the immunization session. The text message will remind primary caregivers about the specific vaccine that the child is due for and when and where it can be administered. This experiment will also be cross-randomized with the incentives experiment.
Experimental Design Details
Randomization done in office by a computer
1. Incentives experiment: level of randomization is Primary Health Centre (PHC). The variation of incentives is randomized at the Sub-centre level.
2. Communication Experiment: Level of randomization is the village
3. Targeted Reminders Experiment: Level of randomization is the village
Was the treatment clustered?
Sample size: planned number of clusters
1. Incentives Experiment: 140 PHCs, 755 SCs
2. Communication Experiment: 943 villages
3. Targeted Reminders Experiment: 2319 villages
Sample size: planned number of observations
Incentives Experiment: 14500 households from the survey
Also plan to use administrative data from the application used by the ANMs during immunization session. The application is designed specifically for this project and will collect data on all the shots provided during the program period.
Sample size (or number of clusters) by treatment arms
Incentives Experiment: 70 PHCs -control, 70 PHCs -treatment. Within these 70 treatment PHCs:120 SCs- high incentive, increasing payment, 120 SCs- low incentive, increasing payment, 73 SCs- high incentive, flat payment, 73 SCs- low incentive, flat payment
Communication Experiment: 392 villages- pure control, 130 villages- random seed treatment, 134 villages- gossip seed treatment, 148 villages- trusted seed treatment, 139 villages- trusted gossip treatment
Targeted Reminders Experiment: 1160 villages- control, 1159 villages- treatment
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
INSTITUTIONAL REVIEW BOARDS (IRBs)
Committee on the Use of Humans as Experimental Subjects, Massachusetts Institute of Technology
IRB Approval Date
IRB Approval Number
Analysis Plan Documents
Pre Analysis Plan
July 09, 2018