The Study of Cardiovascular Risk in Adolescents (ERICA) aims to estimate the prevalence of cardiovascular risk factors and metabolic syndrome in adolescents (12-17 years) enrolled in public and private schools of the 273 municipalities with over 100,000 inhabitants in Brazil. The study population was stratified into 32 geographical strata (27 capitals and five sets with other municipalities in each macro-region of the country) and a sample of 1,251 schools was selected with probability proportional to size. In each school three combinations of shift (morning and afternoon) and grade were selected, and within each of these combinations, one class was selected. All eligible students in the selected classes were included in the study. The design sampling weights were calculated by the product of the reciprocals of the inclusion probabilities in each sampling stage, and were later calibrated considering the projections of the numbers of adolescents enrolled in schools located in the geographical strata by sex and age.

This paper describes the sample design for the National Survey into Labor and Birth in Brazil. The hospitals with 500 or more live births in 2007 were stratified into: the five Brazilian regions; state capital or not; and type of governance. They were then selected with probability proportional to the number of live births in 2007. An inverse sampling method was used to select as many days (minimum of 7) as necessary to reach 90 interviews in the hospital. Postnatal women were sampled with equal probability from the set of eligible women, who had entered the hospital in the sampled days. Initial sample weights were computed as the reciprocals of the sample inclusion probabilities and were calibrated to ensure that total estimates of the number of live births from the survey matched the known figures obtained from the Brazilian System of Information on Live Births. For the two telephone follow-up waves (6 and 12 months later), the postnatal woman's response probability was modelled using baseline covariate information in order to adjust the sample weights for nonresponse in each follow-up wave.

Este artigo descreve como podem ser considerados na análise dos dados da Pesquisa Nacional por Amostra de Domicílios (PNAD) do IBGE os diversos aspectos de seu plano amostral complexo: estratificação, conglomeração, probabilidades desiguais de seleção e ajustes dos pesos para calibração. Para isso, inclui: uma descrição resumida porém completa do plano amostral dessa pesquisa; indicação de como seus dados podem ser usados para estimar totais; e também uma descrição resumida dos métodos essenciais para ajustar modelos paramétricos regulares com os dados da pesquisa levando em conta os aspectos de amostragem complexa. Apresenta ainda os resultados de algumas estimativas para características de pessoas e domicílios calculadas com base nos dados da PNAD/1998, para as quais são apresentadas estimativas dos respectivos desvios padrão e dos efeitos do plano amostral. Conclui com uma indicação dos cuidados que os usuários devem ter ao analisar tais dados em sua prática de pesquisa.

This paper describes the sample design used in the Brazilian application of the World Health Survey. The sample was selected in three stages. First, the census tracts were allocated in six strata defined by their urban/rural situation and population groups of the municipalities (counties). The tracts were selected using probabilities proportional to the respective number of households. In the second stage, households were selected with equiprobability using an inverse sample design to ensure 20 households interviewed per tract. In the last stage, one adult (18 years or older) per household was selected with equiprobability to answer the majority of the questionnaire. Sample weights were based on the inverse of the inclusion probabilities in the sample. To reduce bias in regional estimates, a household weighting calibration procedure was used to reduce sample bias in relation to income, sex, and age group.

We consider a model dependent approach for multi-level modelling that accounts for informative probability sampling, and compare it with the use of probability weighting as proposed by Pfeffermann et al. (1998a). The new modelling approach consists of first extracting the hierarchical model holding for the sample data as a function of the corresponding population model and the first and higher level sample selection probabilities, and then fitting the resulting sample model using Bayesian methods. An important implication of the use of this approach is that the sample selection probabilities feature in the analysis as additional outcome values that strengthen the estimators. A simulation experiment is carried out in order to study and compare the performance of the two approaches. The simulation study indicates that both approaches perform generally equally well in terms of point estimation, but the model dependent approach yields confidence (credibility) intervals with better coverage properties. A robustness simulation study is performed, which allows to assess the impact of misspecification of the models assumed for the sample selection probabilities under informative sampling schemes. SUMMARYWe consider a model dependent approach for multi-level modelling that accounts for informative probability sampling, and compare it with the use of probability weighting as proposed by Pfeffermann et al. (1998a). The new modelling approach consists of first extracting the hierarchical model holding for the sample data as a function of the corresponding population model and the first and higher level sample selection probabilities, and then fitting the resulting sample model using Bayesian methods. An important implication of the use of this approach is that the sample selection probabilities feature in the analysis as additional outcome values that strengthen the estimators. A simulation experiment is carried out in order to study and compare the performance of the two approaches. The simulation study indicates that both approaches perform generally equally well in terms of point estimation, but the model dependent approach yields confidence (credibility) intervals with better coverage properties. A robustness simulation study is performed, which allows to assess the impact of misspecification of the models assumed for the sample selection probabilities under informative sampling schemes.