The article featured today is from Statistics in Medicine with the full article now available to read here.
A distribution-free procedure for testing versatile alternative in medical multisample comparison studies. Statistics in Medicine. 2022; 41( 16): 2978– 3002. doi:10.1002/sim.9397, , .
The comparison of samples from several populations is an important practical problem, especially in medical studies, but also in various other scientific areas. The paper proposes a test for multisample comparison studies that can be applied without strict assumptions, especially when the underlying population distributions are far from normal. The test can detect differences not only in location or scale but also in shape parameters among parent population distributions. The research is motivated by numerous medical studies, where the variables are not normally distributed and may present more complex differences than simple differences in a particular aspect of underlying distributions, such as location or scale. In these situations, traditional ANOVA and Kruskal-Wallis tests are unreliable since the underlying assumptions are not valid. The proposed procedure also allows the researcher to determine which aspects are more responsible for a significant result. This is an important practical advantage over procedures testing for general differences among the distribution functions but cannot identify which aspects lead to significant results. Both large and small sample behaviours of the test are analyzed. The proposed procedure is illustrated with a multisample comparison study of a biomarker for liver damage in patients with hepatitis C. More precisely, 75 patients with a proven serological and histopathological diagnosis of hepatitis C were grouped into three groups according to the hepatic activity index. The proposed test is used to compare aspartate transaminase (AST) blood level among the three groups to see if it is significantly different as expected since the groups refer to varying levels of liver damage. Positive skewness and kurtosis larger than (normal) 3 of AST distributions are very well documented. It is shown that the proposed test has several practical advantages over common competing methods.More Details