Back to Home page
19 REVISTA BRASILEIRA DE ESTATISTICA, July-December 1997, Vol. 58, N° 210
00.19.1 - LIMA DE FARIAS, Ana Maria ; SERRANO BARBOSA, Maria Tereza
Space analysis indicators of quality of life in Rio de Janeiro [Análise espacial de indicadores de qualidade de vida para o município do Rio de Janeiro]
In this article the authors define variables which allow them to classify the administrative sectors of the city of Rio de Janeiro in homogeneous groups concerning demographic, physical, economic and security characteristics. Factorial analysis is first applied to reduce the dimension of the set of the variables; then K-means method of clustering analysis is applied on grades defined over factorial scores to identify the clusters. Results show the diversity of the groups for the four dimensions of the study, emphasizing the importance of determining public priorities by administrative sector.
Portuguese- pp. 7-44
A. M. Lima de Farias, Rua Pires de Almeida, 49/201, Laranjeiras, RJ 22240-150, Brazil.
amlima@netgate.com.br ; tereza@malaria.procc.fiocruz.br.
(BRAZIL, METROPOLIS, NEIGHBOURHOODS, QUALITY OF LIFE, INDICATORS, CLASSIFICATION.)
*****
00.19.2 - PESSOA, Djalma Galvão Carneiro ; SILVA, Pedro Luis do Nascimento ; DUARTE, Renata Pacheco Nogueira
Statistical analysis of sample survey data: Problems link to the cluster and stratified samplings [Análise estatística de dados de pesquisas por amostragem: Problemas no uso de pacotes-padrões]
When performing descriptive analysis of estimates of finite population quantities of interest, such as means, totals, ratios and proportions, statistical agencies consider the weights and sample design adopted to obtain the data. However, users of such data outside the statistical agencies frequently carry out modelling and analysis without considering the weights, and principally the sample design. Statistical software usually assumes that the observations are IID (independent and identically distributed), that is, that they were generated by simple random sampling with replacement. Sample surveys, in general, use more complex designs, and this can affect the estimates, particularly of variances of point estimates. In this report the authors study the effect of ignoring sample weights and design in the analysis carried out by Leote (1996), who used data from the supplement on labour of the 1990 PNAD (a national household sample survey) for the State of Rio de Janeiro. She fitted logistic models to data on participation of the informal economy using standard statistical software. Here a similar model is fitted using software that incorporates both the complex survey design and its weights, and results of both modelling exercises are compared. The impact of ignoring important aspects of sample design, such as stratification, clustering and unequal weights is discussed. The difficulties faced by external users of sample survey data (such as those from PNAD) for modelling and analysis are also highlighted, because they often ignore such important features of the design.
Portuguese- pp. 53-75
D. Pessoa, P. L. Silva and R. Duarte, IBGE, Av. República do Chile, 500, 10° andar, Centro, 20031-170, Brazil
pedrosilva@ibge.gov.br
(METHODOLOGY, STATISTICAL ANALYSIS, SAMPLING, CLUSTER SAMPLING, STRATIFIED RANDOM SAMPLING, ERRORS.)
*****
19 REVISTA BRASILEIRA DE ESTATISTICA, January-June 1998, Vol. 59, N° 211
Bio-environment and quality of life: Biostatistical perspectives [Bioambiente e qualidade de vida: Perspectivas bioestatísticas]
The quality of human life in a community or society, judged from a broader perspective, depends not only on the standard of living in a conventional monetary sense but also on the bio-environment that governs a battery of social, economic, religious, cultural, health and psychological, ecological and environmental factors. These perspectives are appraised in the formulation and interpretation of quality of life, and, as presented in an article written by the same author (1996), due emphasis on biostatistial foundations is paid in this context.
Portuguese- pp. 7-46
P. K. Sen, University of North Carolina, Chapel Hill, U.S.A.
(QUALITY OF LIFE, THEORY, ENVIRONMENT, BIOSTATISTICS.)
*****
Back to Home page