Impact Evaluation

The impact evaluation aimed to understand whether the demand generation and habit formation interventions had any significant effect on community toilet usage, open defecation practices and some health-related outcomes. Researchers used the randomized control trial (RCT) methodology to conduct the evaluation.

Study sample

The primary sample for the study were all households within a 5-minute walking distance of the facility who did not have a private or shared toilet. These households were considered to be the primary potential users of the community toilet facility.

Data Collection

Our team conducted multiple rounds of data collection to understand the socio-economic status, and sanitation access, behaviors and preferences of households. This included a census and baseline survey before toilet construction to identify households and collect baseline information. We also collected real-time data on usage of the Sammaan facility for one year through a tablet-based Management Information System (MIS) developed for the project and operated by facility caretakers. After one year, we conducted an endline survey. (Learn more about the Sammaan MIS by clicking on the MIS tile below.)

Experimental Groups

In an RCT, the interventions are randomly assigned. The analysis is conducted by comparing the outcomes from different groups: i.e. toilet usage of households in demand generation communities versus toilet usage of households in communities where demand generation was not conducted.

The diagram below shows the different experimental groups in the Sammaan RCT.

  • Demand Generation


The demand generation intervention was assigned through a computer lottery. Some communities were randomly assigned to receive the intervention and the rest of the communities did not receive it during the study period.

14 neighbourhoods randomly assigned toGET Demand Generation

15 neighbourhoods randomly assigned toNOT GET Demand Generation

  • Hygeine faucets

  • at the gender section level

The facility was divided into gender sections and the hygiene faucets were randomly allocated via computer-based lottery.

  • Calendars and discount vouchers

  • at the household level

The calendar and the discount voucher interventions were assigned through a doorstep lottery.

Results fourtcoming

  • Resources

Check out the resources to see J-PAL’s introduction to evaluation, including RCTs, and the analysis plan for the Sammaan study.

J-PAL’s Introduction to Evaluations

Pre-analysis plan