Improving Households' Attitudes and Behaviours to Increase Toilet Use (HABIT) in Bihar, India

"[T]he findings highlight the limitations of short-term behavioural interventions in changing entrenched social attitudes. These interventions can work only in conjunction with bottom-up social change and will not be sufficient, in isolation, to change deep-seated biases."

India's sanitation problem has long puzzled health researchers and economists. To try to explain the persistently high rates of open defaecation (OD), social scientists have established a link between deep-rooted caste bias and sanitation practices, which help explain household resistance to using affordable pit latrines and aversion to pit emptying, a task that has traditionally been looked upon as one to be done by particular groups. In response, Oxford Policy Management (OPM), ideas42, and World Vision India (WVI) designed and developed a behavioural intervention directed at shifting norms, correcting underlying cognitive biases, and promoting the development of intention to use latrines. This International Initiative for Impact Evaluation (3ie) study evaluates the impact of the intervention in increasing the intent and habit of toilet use in rural Bihar, India, where up to 70% of people practice OD.

The eligible population for the study were households that have functional twin-pit toilets in six blocks across six districts of Bihar (Jamui, Khagaria, Madhepura, Nalanda, Nawada, and Saharsa) within which WVI was already working. Based on formative research conducted between November 2016 and May 2017, the project partners entered an iterative design phase, where different design ideas were prototyped and piloted, leveraging WVI's network of community facilitators and volunteers. What emerged was the improving households' attitudes and behaviours to increase toilet use (HABIT) intervention, which used a set of inter-related activities and tools to create and activate intentions to use latrines by shifting norms related to the acceptability of OD. Specifically:

• Two community meetings that featured games targeting behavioural barriers such as overestimation of pit filling rates, unclear rewards to latrine use, and ambiguity around pit decomposition and emptying.
• Household visits that included additional components addressing barriers to intention and habit formation: (i) a card game helped impart correct information regarding the pit-filling rate for the household; (ii) a poster with the household's pledge to reaffirm their commitment and help form a habit; (iii) a chalkboard to track the toilet use of different family members and hold family members to their commitment; and (iv) a lockbox to save money for expenses associated with toilet maintenance and toilet repair, which also acted as a commitment device to help the family maintain their habit of use.

The qualitative components of the evaluation included:

• In-depth interviews with household members conducted with one adult member per household, who was present for at least one of the two interventions (community meeting and household visit). The respondents in the sample belonged to scheduled caste (SC) and other backward caste (OBC) households;
• Focus group discussions (FGDs) with community members who attended community meetings - primarily women from 24 to 50 years of age. FGD respondents belonged to either SC or OBC households, depending on the caste predominant in the sampled ward; and
• Key informant interviews conducted with the following people: members of village institutions, people who empty pits for a living, and WVI representatives.

There were 44 treatment wards and 44 control wards. A quantitative baseline household and community survey were conducted in February 2018. The Improving HABIT intervention started in May 2018. A process assessment of the intervention was conducted in August 2018. Following the completion of the intervention in September 2018, the researchers conducted the endline data collection, both quantitative and qualitative, in January 2019.

The study found significantly high improvements in toilet use. At baseline, around 48% of households in the survey sample were OD free (ODF). At endline, this figure increased to over 80%. While at baseline, over 60% of individuals in the sample were usually using a toilet, this figure increased to over 90% at the endline. However, this increase was largely similar across intervention and control areas, suggesting that the increase was not the result of the intervention.

Despite the significant increases in toilet use, the study found limited change in intermediate outcomes - i.e., knowledge and attitudes around pit-filling and pit emptying. A majority of households continued to mis-estimate rates of pit filling and perceived pit emptying as inconvenient. However, households in the intervention areas were six percentage points more likely to correctly estimate the pit-filling rate than their control counterparts. In addition, while the number of people underestimating the amount of time taken for pits to fill increased in control areas, this number fell substantially in the intervention areas from 43% to 23%.

The study findings suggest that the Swattch Bharat Mission (SBM) and a range of other activities related to behaviour change in Bihar have significantly increased the use of twin-pit toilets, constructed under the SBM. Despite this, sustainability of toilet use is a concern, given the persisting underlying misconceptions around pit filling and a lack of awareness about the process of decomposition outcomes, which the intervention was able to influence to a certain extent.

Some findings from the HABIT study that can provide input into the next phase of SBM programming include the need to focus on pit emptying and greater messaging on the design of the twin pit. Construction of toilet pits and inconsistency in type and size presents a challenge for implementers, making it difficult to tailor standardised messaging around pit filling. Providing adequate information to ensure conversion of faecal matter to non-pathogenic decomposed matter and framing the sanitation issues around social attitudes and behaviours would be necessary to promote self-emptying and to avoid further entrenchment of caste-based pit emptying.

For those designing behavioural interventions, the study's findings highlight the need to contextualise programme components. The use of a chalkboard to track and motivate habit formation was not successful, with interviews suggesting that households were uncomfortable with tracking their toilet use and having it visible to others, especially guests in their households. Piloting and detailed qualitative interviews during the piloting phase would help gauge adverse reactions to intervention design.

In conclusion: "This study provides support to the evidence around aversion to pit-emptying and deeply entrenched notions of caste, purity and pollution. This highlights the importance of treating sanitation as a social issue, and not merely as one related to access and toilet use behaviours....Given deep-seated social biases about pit emptying, the findings also point to the need for a concentrated effort from the government, civil society and programme implementers to design messaging to change attitudes around caste-based pit emptying and the need to look at sanitation behaviours from a social change standpoint."