In partnership with
Encouraging teachers and principals to turn up to school in Peru
Teacher absenteeism can be fun for students. If we ask you to remember a day at school when a teacher didn’t show up, it probably brings back memories of happy faces, a movie, or a game outside. Unfortunately, research shows that even a few days of teacher absenteeism can negatively affect the reading and maths abilities of students.
That’s the reason behind our recent work with the Peruvian Government. Using data from random spot checks, we found that on average 7% of teachers and 17% of principals are absent on any given day. Each year, this corresponds to teachers missing over two school weeks (12.5 days – about 3 times as much as in the UK) and principals missing a full month (30.5 days) of schooling.
How do we get teachers and principals to turn up to school? Working with the Peruvian Ministry of Education (Minedu) and the World Bank’s Mind, Behavior, and Development Unit (eMBeD), we set to find out – designing a randomised trial to test whether or not attendance could be influenced by emails sent by the Ministry of Education.
Close to 100,000 teachers and principals in 27,000 schools around Peru were divided into three groups and received either:
- no email (control);
- two emails highlighting that the large majority of teachers are usually present at school (social norm, see figure 1), or;
- two emails emphasising the positive effects of teacher attendance on student performance (prosocial).
Figure 1: Example of email highlighting the social norm (Spanish)
In order to measure impacts on attendance, we designed the trial to be implemented within an existing nationwide survey being conducted by the Ministry of Education – randomly allocating the schools to be visited each month to each of our three trial arms.
The survey includes spot checks where survey staff check directly if teachers are present in classrooms, providing us with a more objective measure than absences reported by school principals (who our qualitative research revealed are likely to under-report their teachers’, and their own, absences). This “embedding” method within an existing survey is a great example of how BIT and eMBeD aim to design trials that are both rigorous, and cost-effective for governments.
What did we find? Our results show that the emails had no impact on the attendance of teachers, but that the `social norms’ emails significantly increased average attendance by principals, from 83% to 87%.
This corresponds to approximately 7 fewer missed days of school a year per principal (out of 179 days) – a substantive effect, especially for an intervention that had near zero cost to the Peruvian government.
Figure 2: Impact of behavioural emails on principal attendance
There are a few reasons that may explain why we see an impact on principals but not on teachers.
First, principals were twice as likely to open emails received at their institutional addresses than teachers. The difference in effect could therefore be simply explained by the fact that more principals actually read the emails we sent.
Second, principals’ attendance was lower at baseline, which means that it had greater room for improvement than teachers’ attendance.
What comes next? In future work, we hope to test more cost-effective ways to reach out to teachers, revisiting options we had initially considered (but excluded because of implementation issues) such as adding behavioural messages on payslips. These will be more complicated to implement than emails, but would present an opportunity to make direct links between absences and their negative consequences on pay – which should be a strong motivator.
- Australia (15)
- BI Ventures (2)
- BX Conference (12)
- BX2018 (4)
- BX2019 (1)
- Charitable giving (9)
- Consumer affairs (6)
- Data science (1)
- Education and skills (20)
- Energy and sustainability (6)
- Equality and discrimination (6)
- Health (22)
- Home affairs and security (9)
- International (9)
- Labour market & economic growth (9)
- New Zealand (1)
- News (72)
- North America (15)
- Reports (28)
- Singapore (1)
- Tax (4)
- Trial design (6)
- Trial results (28)
- Uncategorized (99)
- Vacancies (11)