Enquête Ménage sur la Migration et les Transferts de Fonds 2009
Migration and Remittances Household Survey 2009
The main objective of the assignment is to contribute to the improvement of migration and remittances impact on development in sub-Saharan Africa. Some specific objectives are to provide better understandings of migration and remittances through household surveys. By enhancing the knowledge regarding migrant remittances sending or receiving, these data will be a reliable input for policy-makers. The survey will provide information on migrant and non-migrant households' socioeconomic characteristics and how they send remittances back home. Appropriate approach and methodology will be applied to carry out the survey activities for high quality data.
Kind of data
Sample survey data [ssd]
Unit of analysis
Ten most intensive migration provinces in Burkina Faso: Banwa, Boulgou, Boulkiemdé, Kadiogo, Namentenga, Passoré , Sanmatenga, Sourou, Tuy and Yatenga.
Unit of analysis
Households from the geographic coverage
Producers and sponsors
University of Ouagadougou
The sampling approach was chosen to focus on international migration and remittances. It has ensured an adequate representation of migration density in Burkina Faso. First we used the Burkina Faso Census 2006 that includes questions on demographic and migration. Then we calculated the incidence of international migration for each of the 45 provinces in Burkina Faso. We chose the largest 10 provinces which have the highest international migration incidence. The universe for this survey is the households that are placed in these provinces. After chosen the universe of households, we have adopted a two stages sampling: i) village/town sector and ii) household. Migration scope is different depending on the province or rural/urban area in Burkina Faso. We determine the number of village to be selected in each province according to the international migration incidence of each province. In total we selected 78 villages out of 2,273 villages for the 10 provinces.
Village town sector selection
The first stage of sampling consisted of villages and towns selection. We randomly selected the village/town sector within each province with equal probability. We have considered a village/town sector sample size of 27 households. We distributed the 27 households according to the following rule: 9 households with international migrants, 9 households with internal migrants and 9 households with no migrants. Thus, the whole sample is composed of 78 villages/towns sectors for 2,106 households. Considering migration density in each province, a number of villages/town sectors were randomly selected by province. In big town as Ouagadougou, each sector was considered as a village.
The second sampling stage was household selection. Households were randomly selected after doing an exhaustive and quick household census in the selected village. For the villages in rural areas, we listed all the households in the village. The listing survey was not done for all households in town: we have randomly selected about 120 households in each selected town sector around a random start point. This number is extended if needed in order to attend the household type quotas. We have determined this point by using the town addressing codes. The process consisted of selecting randomly a street code and an avenue code which constituted the two elements of the start point. The main objective of the census is the household sampling. It provides information on the head household name, household size and household category (households without migrants, households with internal migrants and households with international migrants). The sub sample size for each household category is 9 in selected village/town sector. These households are randomly selected among the total households of the category from the quick census.
Most of the questionnaires (99%) have been completed at the first visit. The non response rate is very low: 4 households out of 2106 have not responded. This score could be explained by the census process during which the high experienced team prepared the household for interview if it coma to be selected.
Following the sampling frame, weights are calculated at three levels: province, village and household. The province is weighted considering its population and the total population of the survey area (10 provinces). The village is weighted in the same way, according to its population, the total population of the corresponding province and the total population of the 10 provinces of the survey area. As households have been selected in three categories (households with no migrant, households with internal migrant and households with international migrant), the weight of each household is calculated taking into account the number of households in its category. It also includes the weights of its village and province. For any level, the weight is the inverse of the level unit selection probability. Further calculation details are in the STATA do-file programs while weights are in the database from the survey.
We have adopted a probabilistic two stage sampling method. It is necessary to correct the inequality in probability for the observed unit (household) to be selected in the survey universe. The standard correction method consists of weighting the statistic outcomes.
Dates of collection
Mode of data collection
Data collection supervision
The supervision is at the core of survey data quality. It was assured firstly by supervisor at the team level. The supervisor is the field team leader and his main task is to verify and insure that data are well collected by interviewers. He is the main technical assistance provider in the survey field and then, he is responsible for household sampling in the village. This random household sampling was conducted at the presence of population in the village or village representative persons. The supervisor has presented his team to local representative persons. He organised his team to be efficient in survey field. He has also managed any problem within team or between team and households. He has informed the survey staff about problems he could not manage. This report has informed us about problems his team has encountered, how it has been solved, and what consequences for data quality and analysis.
A global supervision for all the teams was done by the survey staff. It consisted of improvised team visiting. Each team was visited 2 times. Other supervision visits was needed for urgent situations as those encountered in the Boulgou Province. The main objective was to insure the correct survey ways for all teams in the work field and to share efficient astuteness found by any team for best quality data. We have focused a particular attention on the first global supervision in order to correct some possible survey errors at the earliest.
8 section questionnaire designed in English by the World Bank Migration and Remittance Team with the collaboration of the survey implementation team staff. This questionnaire has been translated into French by the survey team staff and submitted to pre-test and pilot in area with similar migration and remittance characteristics to the survey zone.
Data have been cleaned using STATA software. The double entry process was the first level of dataset quality control. Several consistency and cross checking rounds have been done in order to ensure a high dataset quality. This process often needed to check error using the hard completed questionnaires. Very few errors have been unsolved. We made no imputation for missing data in the dataset.
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including acronym and year of implementation)
- the survey reference number
- the source and date of download
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.
Development Economics Data Group
The World Bank
Generation of the DDI
Version 02 (September 2017)
Changed survey year from 2010 to 2009 according to data collection dates