Enquête Harmonisée sur le Conditions de Vie des Ménages, 2021-2022

The sampling frame for the 2018/2019 survey is based on the 2009 general population and housing census (RGPH). In 2021/2022, an enumeration was conducted in the same clusters. The Mali EHCVM 2018/19 utilized the sixth edition of the 2017/18 modular and permanent household survey (EMOP) as the sampling frame. This frame comes from the 2009 Census of Population and Housing (RGPH) and contains 1153 enumeration areas, is nationally representative, and covers all ten regions and the Bamako district. The survey collected data in rural and urban areas in all regions in each of the two waves, except Kidal, Bamako, Toaudénit, and Menaka. More specifically, during the first wave, the survey did not collect data in rural Kidal. Moreover, due to Bamako's purely urban nature, no rural area data are available for this district. Likewise, concerning Taoudenit and Menaka, the survey did not collect data for those regions during the first wave. In addition, there is no urban data for Taoudenit in wave 2.

In Mali, the survey design team implemented a 2-stage stratified sampling methodology - without replacement. At the first stage, 500 enumeration areas (EAs) were selected with Probability Proportional to Size (PPS) using the number of households as a measure of size. In the second stage, 12 households were selected in each enumeration area randomly. Furthermore, 51 additional EAs were selected from the sampling frame during the second wave to cover Menaka and Taoudenit as the wave 1 EAs did not contain those two regions.

In 2021/2022, the strategy is to revisit the same clusters. This involves either surveying the 12 households from 2018/2019 if they are found (after the enumeration phase), or surveying the found households and completing the sample to 12 in clusters where fewer households are found during the enumeration phase (either because there were fewer than 12 households in the final 2018/2019 database or because some households were not found).

The total estimated survey sample size was 6 708 households – 3528 from urban areas and 3180 from rural areas. After that, the survey design randomly divided each enumeration area into two equal groups. The survey team interrogated the first group in wave one and the other in wave 2. Finally, due to the household and EA replacement strategy put in place, the final sample size was very close to the expectations, with 2,732 and 3,411 households surveyed respectively in urban and rural areas. For the replacement of households, households within the EA replaced vacant/refusing households. These replacement households were assigned to agents by the field team leaders and came from the list of replacement households in the enumeration file. In practice, in each EA, 18 households were selected, including 12 households to be surveyed and six replacement households. Concerning the replacement of EAs, during the first wave, seven (07) EAs could not be surveyed due to insecurity (2 in Timbuktu and 5 in Mopti). These seven EAs were replaced by the EMOP EAs and surveyed in wave 2. The survey team made those replacements considering the area of residence of the EA. In terms of detailed actual surveys, the survey teams interviewed 1,352 households in urban areas and 1,479 households in rural areas during wave one. In wave two, the teams interviewed 1,380 households in urban areas and 1,932 households in rural areas.