Quantitative Service Delivery Survey in Health 2002

The description of the sampling procedure below is taken from "Initial Project Description: Survey of Education and Health Providers." This document is available in external resources.

"For the health sector, we plan to sample using a facility-based rather than population-based approach. Where it is available, we will acquire a list of all the Primary Health Centers (PHCs) within each district. (With the schools, this may not be possible, because there may be no central lists of private schools). Facilities will then be chosen randomly for visits, after stratification by rural/urban location. Lists would ideally be obtained from both central government and district authorities.

To reduce travel and transportation costs, it may sometimes be necessary to cluster villages/towns or facilities. Under the facility-based selection approach, for example, five areas may be randomly chosen within each district, and two schools in that area will be selected, rather than choosing a random sample of ten areas. During data analysis, we will adjust standard errors to account for clustered sampling.

At the facility level, we will also obtain a roster of teachers in the school or health care workers assigned to the clinic. If the facility is large (for example, if there are more than 25 teachers in a school), we will interview a random sample of the teachers to keep the size of the survey manageable.

This survey is focused on basic education and health care. Given time and personnel constraints, it will therefore focus only on primary schools and primary health care clinics, and not secondary schools or hospitals.

In each jurisdiction (nation or Indian state), we will survey 10 districts, with at least two visits each to a representative sample of at least 10 health facilities and 10 or more primary schools within each; if the average village has 1.5 schools, the sample will actually be 15 schools per district. This means detailed and representative provider- and facility-level results from perhaps 150 schools and some 100 health centers for each jurisdiction. In addition, there will be a third visit to some smaller sub-sample of the schools and to all of the health centers, as a check and to provide additional data on long-term absence. With these repeated visits, we expect to carry out some 300 school visits and 300 health center visits in each jurisdiction, which should provide several thousand observations of presence/absence for individual providers and all of the necessary facility-level correlates."