Preparation of the annual census surveys in Mayotte : study of the survey design and redefinition of the collection areas

The first annual census survey was hold in Mayotte in 2021. The methodology usually used for the census of the population in a « large » municipality (municipality of 10,000 inhabitants or more) in the French overseas departments is based on the questioning of approximately 8 % of the dwellings in the municipality each year, randomly selected by survey within areas called « islets ». Prior to this draw, a cartographic survey is carried out between five and ten months before the census survey.

The importance of makeshift housing in Mayotte – which represents 39 % of the dwellings in the exhaustive census of 2017 – is likely to significantly affect the performance of this methodology. Indeed, this type of dwellings evolves very quickly (construction, destruction, etc.) and makes if difficult to find the real ones selected. Based on simulations carried out from the census of 2017, this document provides an assessment of the impact of makeshift housing on census estimators as well as the performance of alternative drawing scenarios.

In addition, the islets mobilized for the census of 2017 are very large and heterogeneous in size, which implies methodological and field difficulties. In close collaboration with the Insee regional service in Mayotte, these islets have been redefined so that their borders are durably identifiable on the ground and that they are reasonably homogeneous in size.

This document also describes the process implemented to do this.