Subscribe to E-Bulletin Donate to CEGA

Courting Safe Behaviors: Testing Courtyard-based Safe Water and Hygiene Interventions in Urban Bangladesh

Development Challenge

Globally there are approximately 1.7 billion cases of diarrheal disease per year, and 760,000 of those cases lead to death in children under five.[1]  While many interventions have been piloted to combat diarrheal disease, few have had widespread success.  Treating drinking water with chlorine or filters and washing hands with soap substantially reduce the incidence of these diseases; however, such practices remain uncommon in many developing communities. 


In Dhaka, Bangladesh, a growing network of urban slums has proliferated unhygienic water usage. Families typically live in courtyards with six to 20 other families, sharing a common water source and latrine. Although fecal matter is generally seen negatively across cultures, many Bangladeshis do not believe their drinking water is contaminated, or that the water can cause illness. Thus, in addition to the cost and inconvenience of water treatment, social norms may impede efforts to improve water quality.   

CEGA researchers conducted a randomized evaluation in the courtyards of Dhaka. This project sought to understand the effect of messaging appealing to emotions of disgust at consuming contaminated water and shame of being seen consuming contaminated water on handwashing with soap and use of a chlorine dispenser at the courtyard’s tap.

Evaluation Strategy

Researchers conducted this study in Dhaka communities with piped water sources to the courtyard that had poor water quality and contributed to high incidence of water-borne disease. All compounds were provided a four-month free trial of a chlorine dispenser, with the possibility of extending their subscription based on their bid (described below.)   They were subsequently randomly assigned to one of two messages.  Within each message arm, a subset of the compounds was given soap dispensers so that researchers could measure the effect of messaging on handwashing. There were 435 compounds across the four treatment groups.

Health Messages

Half the courtyards received disgust-and-shame messages to encourage water treatment. The messages used harsh language so that households would feel disgusted that their water was contaminated with shit (we used the corresponding vulgar term) and to create potential peer judgment of others who do not use chlorine treatment. The other half received standard health messages informing participants of the health consequences of consuming untreated water. The messages were modeled on those used by the International Center for Diarrheal Disease Research (ICDDR) Bangladesh.  Use of chlorine dispensers was measured by unannounced checks and observations in the courtyards.

Two-thirds of each message group compounds also received a handwashing treatment.  They were given supplies of soap plus a “soapy bottle” to hold soap and water near the latrine.  Compounds’ availability of soap and handwashing behaviors was observed on unannounced visits. 

Willingness to Pay

Finally, compounds were randomly assigned to either a household or compound bidding group to measure their willingness to pay (WTP) for the chlorine dispensers after the four month trial period, as well as for a year supply of soap if applicable.  Researchers asked household or compound to place a bid of the maximum amount they would pay for an item (in this case, either the dispenser or a year-supply of soap). Groups were given the product if their willing to pay was greater than a randomly generated list value. 

Results and Policy Implications

Health Messages

Even during the free trial, use of the chlorine dispensers remained low (about 12 percent), regardless of behavior change message. 

There are several reasons the disgust-and-shame message may not have been more effective than the traditional health and germ message.  First, most participants reported giving little importance to how neighbors thought of them; thus, social pressure may not have been not a large in these compounds.  Second, we delivered the message only once, while repetition may be crucial to change norms. Finally, we did not reach many household members.  For example, the messages were delivered during the day, when few men were present.   Thus, while the disgust-and-shame approach did not do strikingly well in this study, it is worth studying if a more intensive intervention in a more socially cohesive setting might work better.

The soapy bottle and handwashing intervention resulted in somewhat higher observed handwashing after using the latrine: 15-17 percent compared to 10-11 percent in the groups without the intervention.  This effect was driven by the free provision of the soapy bottle, with similar effects for both messages.

Willingness to Pay

Average willingness to pay for the dispenser was low: 5 taka ($0.14) per month per household.   Follow-up surveys with the treatment group revealed that many had zero or negative willingness to pay because they did not like the taste of chlorine or did not want to share the hardware with other families.   Furthermore, the availability of soap in the compound decreased precipitously at compounds that did not subscribe to the chlorine dispenser (as the chlorine subscription included refills on the soapy bottle). This suggests that households’ demand for handwashing products was very price sensitive, even at extremely low prices.

When proven interventions are not taken up by the poor, a standard conclusion is that the main barrier to adoption is the price. However, this study indicates that take-up can remain low even with free provision and targeted messaging. The social pressure messages used in this study were about as effective as standard health messages, and neither changed household behavior by a lot. Behavioral interventions designed to increase the demand for products such as the ones in this study may require more frequent or longer-lived campaigns to have a large and lasting effect.



[1] World Health Organization. (2013). WHO factsheet, Diarrhoeal diease.  Accessed June 30, 2014.

Photo: WorldFish via Creative Commons. A mother washes her children's hands before the midday meal in Rangpur, Bangladesh.