Human ecology, statistical analysis, and the logic of valid correlations
ResumoHuman health ecology is an expanding field of interdisciplinary research, combining knowledge drawn from distinct areas. Human ecology is essentially interdisciplinary. The search for the origin and distribution of diseases demands a combination of theories and methods from both natural and social sciences. There is a limit to the replication and repetition of observations and experiments in scientific research: hence the importance of statistics. As a tool, statistical analyses are usually based upon selected samples and samples must be representative of the whole. Irrelevant, erroneous, false, and misleading, information based on flawed interpretation of statistics has always plagued scientific publications. In some cases, not only professional ethics has been breached, but risks to well-being and health has been fostered. In this article I will not deal with fraud, but with bona fide errors of judgment and the use of erroneous statistical concepts. My objective in this article is to caution against a common practice adopted by researchers while using qualitative as well as quantitative methods of analysis to establish demographic categories.
Os artigos e demais trabalhos publicados em Política & Sociedade, periódico vinculado ao Programa de Pós-Graduação em Sociologia da UFSC, passam a ser propriedade da revista. Uma nova publicação do mesmo texto, de iniciativa de seu autor ou de terceiros, fica sujeita à expressa menção da precedência de sua publicação neste periódico, citando-se a edição e data dessa publicação.
Esta obra está licenciada sob uma Licença Creative Commons.