The purpose of the survey was to (a) monitor the performances of the programs on child survival protection and development at the sub-national level (that is, at the district level) and (b) to ascertain the achievement of the national-level goals of such programs. Key information collected in the MICS were about prevalence and treatment of diarrhea, breastfeeding, immunization, malnutrition, enrolment in school, maternal and child health, acute respiratory infection, use of iodized salt, water and sanitary conditions.
For the MICS, Bangladesh was divided into rural and urban universes. The urban universe consists of the four Statistical Metropolitan Areas (SMAs) of Dhaka, Chittagong, Khulna and Rajshahi plus the urban areas under the seventy one other municipality (pourasavas). The rural universe was made up of the areas not included in the urban universe.
For generating estimates at different sub-national levels (such as district and divisional levels), the sample was drawn splitting the urban universe into 12 domains and the rural universe into 64 domains. The domains were also referred to as the estimation areas, since the MICS was designed to provide, among others, separate estimates for each domain. An urban domain may spread over one or more than one district. But every rural domain was constructed with the areas taken from only one district. Thus, the rural domains were often referred to as rural districts.
The MICS sample was a two stage sample. The first stage sample consists of 13/16 rural ‘Mouzas’ or urban ‘Mahallas’ selected with Probability Proportional to Size (PPS) from each of the 76 (64 rural and 12 urban) estimation areas. Each selected rural ‘Mouza’ or urban ‘Mahalla’ was divided into clusters of 50 households, on average. The 1991 population census counts of households by Mouzas/Mahallas was used as the frame for the selections of Mouzas/Mahallas and for dividing a Mouza/Mahalla into the clusters of 50 households. The number of clusters created in a rural ‘Mouza’ or urban ‘Mahalla’ was equal to the “census counts of households in the rural ‘Mouza’ or urban ‘Mahalla’ divided by 50”, rounded to the nearest integer.
The second stage sample was constructed by randomly picking one cluster from each of the selected rural ‘Mouza’/urban ‘Mahalla’. Data were collected from all households in a selected cluster, covering all persons in the different target groups. In a household, interviews were usually conducted with mothers having under-five children.
According to the modified design, the MICS sample for a year was constructed by including the old clusters from 75 percent of the estimation areas (domains) and selecting new clusters from the remaining 25 percent of the estimation areas. Clusters covered during the previous year were defined as the old clusters. In new sampling, 16 clusters were selected from an estimation area, yielding an ultimate sample of 800 households from the estimation area. It was thus calculated that the 1997 MICS was a sample of 52,100 households spread over 754 old clusters in 58 estimation areas (48 rural and 10 urban) and 288 new clusters in 18 estimation areas (16 rural and 2 urban).
The specific activities were carried out for the collection of data including participation in a five day TOT for the coordinators. This training were organized jointly by BBS and UNICEF, recruitment of the enumerators and other field personnel, submission of the enumerator’s training program schedule to BBS and UNICEF before starting the enumerator’s training, conducting training of the enumerators and other field personnel on activities relating to the field work, and carrying out field testing of the questionnaire, submission of the schedule of field work for data collection to the BBS and UNICEF before sending the enumerators to the field, conducting the field work for collection of data from the selected clusters, the cluster lists and household lists were provided by BBS and UNICEF, conducting the editing work for the filled-in questionnaires, submission of all filled-in questionnaires to the Project Director, BBS, Goals Monitoring Projects and submission of a brief field report on the constraints faced, including recommendations.
The survey contract was jointly awarded by the UNICEF and the Bangladesh Bureau of Statistics (BBS).