School inspections, bureaucrats, and school quality: Experimental evidence from Tanzania

• In Event: From where you are to where you want to be: How information shapes goals, behaviours and motivation of actors in education

Proposal

The government of Tanzania has recently reformed its school inspection program, reframing the formerly punitive function as a friendly source of support to schools, now called the School Quality Assurance (SQA) program. A key part of the reform is to replace the traditional school inspections with Whole School Visits (WSVs), conducted by 3-4 School Quality Assurance Officers (SQAOs). Each SQA visit produces a set of diagnostics and recommendations across a range of domains that are shared with all stakeholders. The Ward Education Officers (WEOs) are then responsible for conducting follow-up visits to the schools to monitor compliance to these recommendations. We conducted a randomized evaluation of the SQA program (Visit), as well as an additional treatment to increase monitoring and oversight by the WEOs (Visit&Text). Out of a nationally representative sample of 397 schools we randomly assigned 198 schools to receive the WSVs. Moreover, in half of these schools, also randomly assigned, WEOs were encouraged through text-based reminders to engage head teachers on how they are addressing the recommendations of the SQAOs. We collected baseline data in February-March 2019 and midline data in Februray-March 2020.

We document three main findings. First, head teachers in Visit&Text arm revised downwards their beliefs about the quality of school leadership at the start of 2019 (i.e. prior to when most schools received the WSV). There were no commensurate changes in beliefs about the quality of teaching or extent of community engagement. Second, teacher behavior changed in the Visit&Text arm: teacher presence increased by 7.9 percentage points and teaching practice, as measured using the Teach classroom observation toolkit, improved. We find no evidence of improvements in school management, community involvement, nor the overall quality of the school environment. Third, there was a modest improvement in student learning of 0.05SD in Kiswahili, but no improvements in mathematics.

Author

• Jacobus Cilliers, Georgetown University