Performance of five two-sample location tests for skewed distributions with unequal variances

Ullevål Department of Research Administration, Oslo University Hospital, N-0407 Oslo, Norway.
Contemporary clinical trials (Impact Factor: 1.94). 08/2009; 30(5):490-6. DOI: 10.1016/j.cct.2009.06.007
Source: PubMed


Tests for comparing the locations of two independent populations are associated with different null hypotheses, but results are often interpreted as evidence for or against equality of means or medians. We examine the appropriateness of this practice by investigating the performance of five frequently used tests: the two-sample T test, the Welch U test, the Yuen-Welch test, the Wilcoxon-Mann-Whitney test, and the Brunner-Munzel test. Under combined violations of normality and variance homogeneity, the true significance level and power of the tests depend on a complex interplay of several factors. In a wide ranging simulation study, we consider scenarios differing in skewness, skewness heterogeneity, variance heterogeneity, sample size, and sample size ratio. We find that small differences in distribution properties can alter test performance markedly, thus confounding the effort to present simple test recommendations. Instead, we provide detailed recommendations in Appendix A. The Welch U test is recommended most frequently, but cannot be considered an omnibus test for this problem.

Download full-text


Available from: Morten Wang Fagerland,
  • Source
    • ", [27]–[31]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: A novel coherence estimation method for small data sets is presented for interferometric synthetic aperture radar (SAR) (InSAR) data processing and geoscience applications. The method selects homogeneous pixels in both the spatial and temporal spaces by means of local and nonlocal adaptive techniques. Reliable coherence estimation is carried out by using such pixels and by correcting the bias in the estimated coherence caused by the non-Gaussianity in high-resolution SAR scenes. As an example, the proposed method together with coherence decomposition is applied to extract the temporal decorrelation component over an area in Macao. The results show that the proposed algorithms work well over various types of land cover. Moreover, the coherence change with time can be more accurately detected compared to other conventional methods.
    IEEE Transactions on Geoscience and Remote Sensing 10/2014; 52(10):6584-6596. DOI:10.1109/TGRS.2014.2298408 · 3.51 Impact Factor
  • Source
    • "The significance of the data were analysed using the two-sample t-test and one way analyses of variance (ANOVA; [74]). A significant difference was considered to be p < 0.05. "
    [Show abstract] [Hide abstract]
    ABSTRACT: As an opportunistic human pathogen Pseudomonas aeruginosa is able to cause acute and chronic infections. The biofilm mode of life significantly contributes to the growth and persistence of P. aeruginosa during an infection process and mediates the pathogenicity of the bacterium. Within a biofilm mucoid strains of P. aeruginosa simultaneously produce and secrete several hydrolytic enzymes and the extracellular polysaccharide alginate. The focus of the current study was the interaction between extracellular lipase LipA and alginate, which may be physiologically relevant in biofilms of mucoid P. aeruginosa. Fluorescence microscopy of mucoid P. aeruginosa biofilms were performed using fluorogenic lipase substrates. It showed a localization of the extracellular enzyme near the cells. A microtiter plate-based binding assay revealed that the polyanion alginate is able to bind LipA. A molecular modeling approach showed that this binding is structurally based on electrostatic interactions between negatively charged residues of alginate and positively charged amino acids of the protein localized opposite of the catalytic centre. Moreover, we showed that the presence of alginate protected the lipase activity by protection from heat inactivation and from degradation by the endogenous, extracellular protease elastase LasB. This effect was influenced by the chemical properties of the alginate molecules and was enhanced by the presence of O-acetyl groups in the alginate chain. We demonstrate that the extracellular lipase LipA from P. aeruginosa interacts with the polysaccharide alginate in the self-produced extracellular biofilm matrix of P. aeruginosa via electrostatic interactions suggesting a role of this interaction for enzyme immobilization and accumulation within biofilms. This represents a physiological advantage for the cells. Especially in the biofilm lifestyle, the enzyme is retained near the cell surface, with the catalytic centre exposed towards the substrate and is protected from denaturation and proteolytic degradation.
    BMC Microbiology 07/2013; 13(1):159. DOI:10.1186/1471-2180-13-159 · 2.73 Impact Factor
  • Source
    • "It is fairly robust to deviations from normality [4], and—by the central limit theorem—increasingly so when the sample size increases. When the sample size of a study is 200, the t-test is robust even to heavily skewed distributions [5]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: During the last 30 years, the median sample size of research studies published in high-impact medical journals has increased manyfold, while the use of non-parametric tests has increased at the expense of t-tests. This paper explores this paradoxical practice and illustrates its consequences. A simulation study is used to compare the rejection rates of the Wilcoxon-Mann-Whitney (WMW) test and the two-sample t-test for increasing sample size. Samples are drawn from skewed distributions with equal means and medians but with a small difference in spread. A hypothetical case study is used for illustration and motivation. The WMW test produces, on average, smaller p-values than the t-test. This discrepancy increases with increasing sample size, skewness, and difference in spread. For heavily skewed data, the proportion of p<0.05 with the WMW test can be greater than 90% if the standard deviations differ by 10% and the number of observations is 1000 in each group. The high rejection rates of the WMW test should be interpreted as the power to detect that the probability that a random sample from one of the distributions is less than a random sample from the other distribution is greater than 50%. Non-parametric tests are most useful for small studies. Using non-parametric tests in large studies may provide answers to the wrong question, thus confusing readers. For studies with a large sample size, t-tests and their corresponding confidence intervals can and should be used even for heavily skewed data.
    BMC Medical Research Methodology 06/2012; 12(1):78. DOI:10.1186/1471-2288-12-78 · 2.27 Impact Factor
Show more