Premium
Tutorials in clinical research: VII. Understanding comparative statistics (contrast)—part B: Application of T‐test, Mann‐Whitney U, and Chi‐Square
Author(s) -
Neely J. Gail,
Hartman James M.,
Forsen James W.,
Wallace Mark S.
Publication year - 2003
Publication title -
the laryngoscope
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.181
H-Index - 148
eISSN - 1531-4995
pISSN - 0023-852X
DOI - 10.1097/00005537-200310000-00011
Subject(s) - statistics , contrast (vision) , null hypothesis , sample size determination , statistic , mann–whitney u test , statistical hypothesis testing , chi square test , p value , statistical power , statistical significance , confidence interval , test statistic , computer science , sample (material) , test (biology) , mathematics , artificial intelligence , paleontology , chemistry , chromatography , biology
Objective : This tutorial on comparative statistics has been written in two complementary segments. The first paper (part A) focused on explaining the general concepts of the null hypothesis and statistical significance. This second article (part B) addresses the application of three specific statistical tests. These two articles should be read sequentially and the first article should be available for reference while one reads the second. Study Design : Tutorial. Methods : The authors met weekly for 10 months to discuss clinical research articles and the applied statistics. The difficulty was not the material but the effort to make it easy to read and as short as possible. Results : The article discusses the application of three common statistical indexes of contrast, χ 2 , Mann‐Whitney U , and Student t ‐test and other concepts, such as sample size, degrees of freedom, errors, power, and confidence intervals. Conclusions : Statistical tests generate a number known as a statistic (χ 2 , U, t ), which is sometimes called a “critical ratio” because it helps us to make a decision. This number is then associated with a probability, or P value. Sample size is a crucial element in the initial design of a research project and in the subsequent ability of the results to show statistical significance if the difference is clinically important. The example data used in this paper demonstrate the application of the three specific tests and illustrate the effect of sample size on the results.