Evaluation of demographic and health surveys

External evalation covers 9/84-12/87 and is based on a site visit, interview, and document review. The contract to implement 35 surveys over 5 years has required a torrid pace. In a short period of time, the contractor - the Westinghouse Institute for Resource Development (WIRD) - has developed survey instruments, supporting documents, and computer software, and has collected a remarkably large and complex volume of data. Originally 21 short-instrument, 10 long-instrument, and 4 in-depth surveys were planned. The current schedule includes 29 long-instrument, 2 in-depth, 2 male-respondent, and 2 experimental surveys. Of the 29, 12 are in sub-Saharan Africa, 10 in Latin America/Caribbean, and 7 in Asia and North Africa combined. Both experimental surveys were carried out in Latin America. One in-depth survey occurred in Nepal; another is planned for Egypt. As of now, final reports on 3 of the surveys have been published, and 2 are in press. Twelve standard surveys (plus 5 non-standard) are scheduled to have final reports completed sometime in 1988. Along with the remaining 12 (plus one in-depth), this means that 30 survey reports need completion before the project terminates in 9/89. The workload to be completed is clearly overwhelming. The 25 further analyses (under development primarily by the subcontractor - the Population Council) have of necessity been backed into the last 2 years of the project; 5 of these are near official completion. Due to the belated implementation of this activity, the analyses could not be evaluated here. Attention needs to be given to the timely completion of the remaining analyses. With A.I.D. cognizance, the course of project implementation has departed importantly, and most likely for the better, from the original contract plan. WIRD has undertaken more extensive data collection (e.g., a major and expansive set of child health questions) than planned. It has also adopted microcomputer technology exclusively for data processing and analysis, developed in-house software for this purpose, and distributed four times as many microcomputers as called for in the contract. The impressive and rapid pace of implementation, however, has been purchased at the cost of lost opportunities for timely analysis and reflection on the data gathered. The microcomputer technology developed may have contributed to improved data quality, but this remains to be assessed. A more structured opportunity within Phase II for consideration of the information benefits of WIRD"s massive data collection effort is recommended. (Author abstract, modified).