de Carvalho, Alexandre Xavier Ywata da Silva, Gabriela Drummond Marques de Almeida Júnior, Gilberto Rezende Albuquerque, Pedro Henrique de Melo
Year of Publication:
Texto para Discussão, Instituto de Pesquisa Econômica Aplicada (IPEA) 1662
The mapping of frequency rates of events is an important tool for guiding public policy. Nevertheless, this problem has not received adequate attention by most researchers. In many studies, the used ratios have serious problems when the geographic units have a small population at risk and when the observed number of cases is very low, as seen in Pringle (1996). This paper intends to review the use of Bayesian statistics in order to minimize data variability. The Bayesian statistics discussed is called empirical because it uses information for the composition of the sample prior distribution. The studied rates will be exemplified by analyzing the occurrence of homicides by throughout Brazil in 2008.