Revista Panamericana de Salud Pública
versión impresa ISSN 1020-4989
PERUGA, Armando et al. Analysis of participation in surveys in five countries: the importance for public health research. Rev Panam Salud Publica [online]. 2000, vol.7, n.4, pp. 249-254. ISSN 1020-4989. http://dx.doi.org/10.1590/S1020-49892000000400006.
This study compares participation rates and reasons for nonresponse in surveys conducted in five countries of Latin America and the Caribbean. The objective of the surveys was to measure the prevalence of risk behaviors affecting the transmission of human immunodeficiency virus. The surveys were based on probability samples of the population of both sexes between 15 and 49 years old, except in Mexico, where only men were included. Proportions of three components of participation were estimated: residences interviewed, interviewed residences with eligible persons, and eligible persons who completed the interview. In addition, an overall index that combined the three components was calculated. The overall response rate ranged from 35.6% in Mexico to 81.4% in Chile. The component with the greatest variability was the participation of eligible persons, which ranged from 50% in Mexico to 95% in Cuba. These values were lower than what had been expected, especially among men, and will serve to guide future surveys, since rejection rates higher than the ones expected in the protocol should be considered. The results make it possible to infer the validity of the prevalence estimates for the various observed risk behaviors. The results also establish a benchmark to calculate the sample size in future surveys and to improve research methodology.