The adequacy of different robust statistical tests （不同的健壮的统计测试的充分性）.pdfVIP

Psicológica (2013), 34 , 407-424 . The adequacy of different robust statistical tests in comparing two independent groups * Maribel Peró-Cebollero and Joan Guàrdia-Olmos Universitat de Barcelona In the current study, we evaluated various robust statistical methods for comparing two independent groups. Two scenarios for simulation were generated: one of equality and another of population mean differences. In each of the scenarios, 33 experimental conditions were used as a function of sample size, standard deviation and asymmetry. For each condition, 5000 replications per group were generated. The results obtained by this study show an adequate type error I rate but not a high power for the confidence intervals. In general, for the two scenarios studied (mean population differences and not mean population differences) in the different conditions analysed, the Mann-Whitney U-test demonstrated strong performance, and a little worse the t-test of Yuen-Welch. In social sciences, and particularly in psychology, many of the applied research studies use parametric statistical tests to evaluate their expectations or hypotheses. However, in most cases, the adequacy of the use of those tests is not assessed, and the use of those tests is often of dubious validity because the assumptions of the statistical test are violated. A clear example is the assumption of normal distribution, which is often assumed, although observed distributions do not usually follow a normal distribution. In recent years, there have been increasing numbers of studies that pay attention to the assu