RESUMEN
Um problema bastante comum, cuja definição é fundamental na etapa de criação de um projeto de pesquisa é o cálculo do tamanho da amostra. A partir deste cálculo, será definido o cronograma de coleta de dados, ou mesmo a viabilidade do projeto. O objetivo deste artigo é apresentar o cálculo de tamanho de amostra para a estimação de uma proporção (prevalência ou incidência) e para a comparação de duas proporções de grupos independentes, através de exemplos práticos. Verifica-se que o tamanho da amostra para estimação de uma proporção aumenta, quando aumentamos o nível de confiança do intervalo ou quando diminuímos a margem de erro. Quando o objetivo é comparar proporções, o tamanho da amostra aumenta, quando diminuímos o nível de significância ou quando aumentamos o poder do teste, ou quando diminuímos a diferença mínima clinicamente significativa que desejamos detectar entre as proporções.
Sample size calculation is a fairly common problem that has to be faced when designing a research project. The following aspects of the project will be defined based on this calculation: budget, schedule of data collection, existence (or not) of research subjects, i.e., viability of the project. The objective of the present article was to present the sample size calculation for estimation of a proportion (prevalence or incidence) and for the comparison of two proportions of independent groups using practical examples. We demonstrated that the sample size for estimation of a proportion increases as the confidence interval increases or as the margin of error decreases. When the objective is to compare proportions, the sample size increases as the level of significance decreases or as the power of the test increases, or even as the minimum clinical difference between the proportions is reduced.
Asunto(s)
Humanos , Masculino , Femenino , Proyectos de Investigación/estadística & datos numéricos , Tamaño de la Muestra , Incidencia , PrevalenciaRESUMEN
In a recent paper [4], Efron pointed out that an important issue in large-scale multiple hypothesis testing is that the null distribution may be unknown and need to be estimated. Consider a Gaussian mixture model, where the null distribution is known to be normal but both null parameters-the mean and the variance-are unknown. We address the problem with a method based on Fourier transformation. The Fourier approach was first studied by Jin and Cai [9], which focuses on the scenario where any non-null effect has either the same or a larger variance than that of the null effects. In this paper, we review the main ideas in [9], and propose a generalized Fourier approach to tackle the problem under another scenario: any non-null effect has a larger mean than that of the null effects, but no constraint is imposed on the variance. This approach and that in [9] complement with each other: each approach is successful in a wide class of situations where the other fails. Also, we extend the Fourier approach to estimate the proportion of non-null effects. The proposed procedures perform well both in theory and on simulated data.