Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
F Recommendations for the Demographic and Health Surveys Definition of urban The First Country Report that accompanies the release of each DHS survey should clearly state what the term "urban" means in the context of the survey. This would be particularly helpful for researchers concerned about the definition of towns. Even if the defining criteria are the same as those used by the national statistical office as the panel sus- pects they usually are the definition should be prominently displayed in the report. It would also be helpful to know how the answers given by respondents say, to a question on place of previous residence are con- verted into the urban classifications used in the published data. Remarkably little of this basic information can be found in the DHS reports. City size categories Current DHS practice in reporting city size categories must be revised. The current practice is to define "large and capital cities" as those cities of 1 million population and above, leaving a broad residual cat- egory of "small cities," which includes every city of between 50,000 and 1 million population. The smallest urban units, "towns," are those whose pop- ulation is below 50,000 but that meet country-specific criteria for defining a locality as urban. This coding yields too little information on the range and variety of urban locations. At a minimum, the number of city size classes used in the standard DHS location variables should be increased. If it is feasible to do so, these city size codings should be applied throughout each survey in, for example, questions referring to childhood and other previous residences. City population sizes The population sizes of all cities of 100,000 and above should be listed, city by city, in the First Country Report, organized accord- ing to the most disaggregated regional classification given in the dataset. 503
504 CITIES TRANSFORMED Ideally, the population counts would be the most recent available to the na- tional statistics office; a second-best solution would be to report the city populations given in Table 8 of the United Nations Demographic Yearbook or the appropriate table of World Urbanization Prospects. Confidentiality: Two levels of access Until recently, the DHS program did not release detailed geographic identifiers, evidently out of a desire to protect the confidentiality of the survey respondents. An authorization mechanism should be developed that would allow researchers to apply for permission to use geographically detailed identifiers for earlier DHS surveys, which are not covered in the new system. This mechanism might take the form of a two-level permission system whereby: · Some authorized researchers would have access to a data file linking sampling cluster codes to city names, thus permitting use of comple- mentary city-level data in their analyses. · Other authorized researchers would have access to a more detailed data file linking sampling cluster codes to geographic information sys- tem (GIS) codes, thus permitting use of both city-level and neighbor- hood-level data. Migration status for all women If possible, the household questionnaire should elicit the migration status of all women in the household, not just those of reproductive age who are selected for individual interviews. Migration status for men If possible, the household questionnaire should elicit the migration status of all men as well. Duration in current residence For women who participate in individual inter- views, duration of residence in the current location should be categorized more clearly. For migrants, duration of residence is required to properly allocate conceptions and births between current and former residences and likewise to apportion risks of child death. Ideally, information would be provided on the month of the most recent move in the 5 years before the survey. Presumably this information could be gathered even in cases in which it is not feasible to collect a full migration calendar. Water supply For urban residents, data on access to drinking water need to be supplemented with information on the variability of water supply. Weekly or biweekly recall periods would probably be feasible. Questions should also be included on the time (i.e., waiting to use a well) and money costs of water supply. Private access to sanitation Some countries include questions on private access to pit toilets and latrines, but this is definitely not the norm. Every effort
RECOMMENDATIONS FOR THE DHS 505 should be made to include such questions in the standard survey and to supplement them with questions on the time (i.e., waiting) and money costs of access to public or private latrines. Electricity For urban residents, data on access to electricity need to be sup- plemented with information on the variability and money costs of supply. Weekly and biweekly recall periods would appear to be feasible. Reproductive health services As discussed in Chapter 6, urban populations make far greater use of private-sector sources of reproductive health services than do rural populations. The DHS program should consider how to expand its measures of knowledge and access to include more information on private- sector sources of family planning and other reproductive health services. The time and money costs of access should also be included. It may prove useful as well to inquire into the location of services vis-a-vis place of work. Urban community modules At present, DHS community surveys and invento- ries of public services and clinics are fielded mainly in rural communities. Urban communities and neighborhoods are more difficult to define than ru- ral, and some believe that urban communities are so much better supplied with reproductive health (and other) services that there is little to be gained from fielding urban community surveys. Yet situation analyses as shown in Chapter 6 indicate that neither access to services nor service quality should be assumed to be superior in urban as compared with rural areas. Hence the panel urges the DHS program to begin to field urban community modules, working on an experimental basis in selected cities or countries, to gain experience in the construction of meaningful assessments for these communities.
506 CITIES TRANSFORMED TABLE E-2 Urban Migrant-Nonmigrant Differences in Poverty and Access to Services, by Type of Origin Area and Region I, DHS Surveys in Region Nosh Africa Surveys Disadvantage, city origin Disadvantage, town origin Disadvantage, rural origin Sub-Saharan Africa Surveys Disadvantage, city origin Disadvantage, town origin Disadvantage, rural origin Southeast Asia Surveys Disadvantage, city origin Disadvantage, town origin Disadvantage, rural origin South, Central, West Asia Surveys Disadvantage, city origin Disadvantage, town origin Disadvantage, rural origin Latin America Surveys Disadvantage, city origin Disadvantage, town origin Disadvantage, rural origin TOTALS Surveys Disadvantage, city origin Disadvantage, town origin Disadvantage, rural origin Access to Relatively Piped or Poor In-homeWater Access to Flush Access to Toilet Electricity 6 o 33 11 4 2 2 15 6 11 65 s 10 33 3 o o o 32 o 9 4 O O o o o o 29 o o 11 3 o o o 31 o 11 O O s O O O O o 12 13 1 0 0 0 1 1 5 7 4 3 o o 2 0 56 2 o 14 54 45 O O 2 15 20 NaITE: Estimates and tests from probit models, adjusted for city size and woman's age. 5