Last week, we discussed how we designed and implemented behavioural nudges in Philippines schools. In this post, we’ll share what we found from evaluating their impact through a randomized controlled trial.
“Wash your hands with soap” is an imperative we often hear these days. During a pandemic like COVID-19, handwashing is key to slowing the spread of the virus. Even in non-COVID times, it is important for preventing the transmission of infectious diseases such as respiratory tract infection or diarrhoea.
In the Philippines, the Department of Education (DepEd) has sought to increase handwashing among schoolchildren with their WASH in School (WinS) policy. WinS promotes group tooth-brushing with toothpaste, biannual de-worming in public elementary schools, daily group handwashing with soap, and other WASH-related activities and infrastructural improvements.
In previous research in public schools funded by UNICEF, however, IDinsight found that pupils in the provinces of Camarines del Norte and Puerto Princesa washed their hands with water and soap in fewer than 1 in 10 instances after using the toilet, even when adequate facilities were available. The HiFive program, a teacher-led handwashing educational campaign we previously evaluated, modestly improved handwashing rates — but not enough to lead to nationwide scale-up.
As our previous study and others indicate, large public education campaigns, even when they are effective at increasing knowledge about hygiene, do not automatically lead to behavioural change1. Even when people know they should wash their hands and can wash their hands, in the moment they may forget, or be in a hurry, or be too tired2 to actually do so.
In 2019, with the support of UNICEF, USAID, and DepEd we began to test a different approach: low-cost, behavioural nudges to encourage handwashing in schools.
Behavioural nudges can help address the gap between knowledge and behaviour. By changing the way choices are presented, nudges can make people more likely to do something desirable3. For example, to encourage handwashing, previous studies have found painting a bright footpath leading from the toilet to the handwashing station to be effective4. This is a contextual cue that makes heading to the sink and washing your hands after toilet use more automatic.
We adapted successful designs from previous studies to the local context and added new nudges based on field visits to school sites. This nudges design process is detailed in our companion blog post.
In the end, we installed:
To understand how the nudges affected handwashing, we conducted a cluster randomized controlled trial among public elementary schools in Zamboanga del Norte. We randomly allocated 66 schools to the “treatment group”, where we installed nudges. We also randomly selected 66 “control group” schools, where nudges would not be installed5. In order to be eligible for the study, schools had to have adequate handwashing infrastructure such as toilet availability and regular water access. The schools in control and treatment groups were randomly assigned. This meant that each group of schools would on average be the same, except for the nudges. In other words, if the two groups had any differences in handwashing outcomes, we could attribute those differences to the impact of the nudges.
We installed the nudges in October 2019 and returned four months later in February 2020 to assess handwashing outcomes. To do so, we conducted direct classroom observation. Direct classroom observation leads to more objective estimates of handwashing than student self-report. To minimize the possibility of teachers and students behaving differently because they were being watched (observer effects), we did not disclose that we were specifically observing handwashing, and only stated that we were observing “regular classroom activities.”
You might have noticed that February 2020 was at the beginning of the COVID-19 pandemic, and reasonably asked if that had any effect on handwashing. The answer is that it very likely did. Though our data collection occurred just before widespread school cancellations and movement restrictions, teachers and students may have already been showing heightened vigilance over handwashing. In fact, shortly before we arrived, DepEd Zamboanga Del Norte division sent a memo about COVID-19 and the importance of handwashing to all schools. Luckily, our evaluation design randomized between treatment and control. Since we can reasonably assume any boost to base handwashing rates due to COVID would be similar across treatment and control, we were still able to obtain unbiased estimates of the impact of the nudges on handwashing rates.
Nudges improved student handwashing behaviour. Compared to students in control group schools, nudges increased handwashing rates among students in treatment group schools by 17 percentage points (equivalent to an increase of 148 per cent), even four months after implementation. This indicates the impact of the nudges persisted over time. Because we measured outcomes only once, we are unable to determine whether and how their impact changed over time. The percentage point (pp) increase of 17.3 was comparable to the results of a study testing arrows & stickers in the US6, though lower than the impact estimates of school-based handwashing nudges in South Asia7. It is also higher than the HiFive program we previously evaluated. Nudges also led to a much smaller, 8.2pp increase in water-only handwashing, which can reduce bacterial contamination, but is significantly less effective than handwashing with water and soap8. The combined effect was an increase of 25.6 pp in handwashing with at least water. Pupils responded similarly across age groups and genders.
The nudges also led to improved access to functional handwashing facilities. Compared to students in control schools, students in treatment schools were slightly more likely to have access to a functional handwashing facility near toilets, and 38 per cent more likely to have access to a facility with soap.
But these results led us to other questions.
First, how exactly did schools improve access to functional facilities?
To get more insight into this process, we compared treatment impact between two types of handwashing stations: those with faucets and running water; and those with buckets holding stored water for handwashing next to the sink basin. It turned out that improved access to functional facilities is driven primarily by the impact on facilities with stored water. Since teachers primarily maintain handwashing facilities day-to-day, the nudges appeared to have encouraged them to more consistently supply stored water and soap. The nudges not only positively affected student behaviour, but also teacher behaviour.
Second, since we saw a large increase in soap availability, did students wash their hands more simply because they had more opportunity (soap access) to do so, or because they actually changed their behaviour?
We conducted a mediation analysis to answer this question. Results indicate nudges increased handwashing rates, even after controlling for the nudges’ effect on soap availability. That is, this program simultaneously nudged students to wash hands with soap in classrooms that already had soap, and nudged teachers to provide soap where it was not already available.
Despite their low cost, nudges generally remained in good condition after four months. More than 75 per cent of every type of nudge we installed remained at the end of our observation, and the vast majority were still in good condition. The principals, teachers, and students we interviewed were also most enthusiastic about the nudges, though several school staff members emphasized that the school needed infrastructure improvements, particularly consistent water access, in order to take full advantage of the nudges. As one principal stated, “I am happy if there are new things in my school, especially if they are lasting, useable, and useful.”
These results show that low-cost changes to the environment can drive longer-lasting, “sticky” behaviour change. We found notable increases in handwashing rates four months after nudges were installed, which suggests that the contextual and visual cues we installed helped trigger habit formation. This finding is in line with other studies using incentives9 and behavioural nudges,10 which also found behavioural impacts several months post-program installation. However, since we only measured handwashing at one point in time, we’re unable to estimate treatment effects over time. Future studies might consider alternative (but more costly) data collection approaches such as using real-time sensor data or liquid soap volume in order to measure persistence effects.
At a cost of less than $60 USD per school, we are recommending that the Department of Education scale-up school-based nudges in other elementary schools in the Philippines, as a part of its WinS policy. Before implementing nudges, we recommend schools have minimum WASH infrastructure in place: functional toilets, functional handwashing stations, and water availability at least some hours of the day (running or stored), and dedicated school funds for soap.
We are in discussions with UNICEF and DepEd about how to implement these recommendations in the near future. Many schools are currently suspended due to COVID-19, both in the Philippines and in other countries. The nudges, which can improve handwashing rates after toilet use, and other measures promoting hand hygiene at other COVID-specific critical times, could help ensure a safer environment in schools for students when they return.
If you are interested in installing nudges in your school, stay tuned! In the next couple of weeks, we will be releasing a ‘how-to’ guide for others interested in installing nudges in schools.
This research is made possible by the generous support of the American people through the United States Agency for International Development (USAID). The contents are the responsibility of IDinsight and do not necessarily reflect the views of USAID or the United States Government. The project is also supported by the UNICEF WASH team in the Philippines.
Asterisks in graphs refer to: *** p<0.01, ** p<0.05, * p<0.1
We collaborate with government leaders to develop and roll out data-driven policy solutions aligned with their priorities and within their budgets.
We work with international agencies to ensure their agendas and operations are founded on the best available data and evidence.
15 May 2020
28 May 2020
28 April 2020
21 June 2019