Explore chapters and articles related to this topic
Nonparametric Methods
Published in Marcello Pagano, Kimberlee Gauvreau, Heather Mattie, Principles of Biostatistics, 2022
Marcello Pagano, Kimberlee Gauvreau, Heather Mattie
Although the sign test frees us from having to make any assumptions about the underlying distribution of differences, it also ignores some potentially useful information: the magnitude of these differences. For patients with cystic fibrosis and healthy individuals matched on age, sex, height, and weight, a difference in REE of 8 kcal/day is counted the same as a different of 472 kcal/day. As a result, the sign test is not often used in practice. Instead, the Wilcoxon signed-rank test can be used to compare two populations that are not independent. Like the sign test – and the paired t test – the signed-rank test does not consider the measurements sampled from the two populations separately. Instead, it focuses on the difference in values for each pair of observations. It does not require that the population of these differences be normally distributed. However, it does take into account the magnitudes of the differences as well as their signs. The Wilcoxon signed-rank test is used to evaluate the null hypothesis that in the underlying population of differences among pairs, the median difference is equal to 0. The alternative hypothesis is that the population median difference is not equal to 0.
Experimental Design, Evaluation Methods, Data Analysis, Publication, and Research Ethics
Published in Yuehuei H. An, Richard J. Friedman, Animal Models in Orthopaedic Research, 2020
Nonparametric tests are statistical techniques which can be applied when there is no assumption of a Gaussion distribution within the population. Thus they are often referred to as distribution-free tests. While not as powerful or flexible as their parametric counterparts (i.e. t-test and ANOVA), nonparametric tests can be applied in situations where parametric tests are not valid. There are a number of nonparametric tests. Chi-square is the most commonly used and it requires nominal data. Others, which require ordinal data, include the Mann-Whitney U test, the Wilcoxon Signed Rank test, and the Kruskal Wallis test. The Mann-Whitney U test is used to test the hypothesis that the distributions of two different sets of data are equal. It is the nonparametric equivalent of an unpaired t-test. Similarly, the Wilcoxon Signed Rank test is the nonparametric equivalent of the paired t-test. The Kruskal Wallis test is analogous to ANOVA and tests whether two or more sets of data come from the same distribution or from different distributions. Further discussion of their uses may be found in many statistical texts such as those by Abacus Concepts, Inc.,33 Munro et al.,12 and Forthofer and Lee.18
The effect of hypertonic saline irrigation vs baby shampoo in chronic rhinosinusitis
Published in Cut Adeya Adella, Stem Cell Oncology, 2018
D. Munir, I.N. Tobing, M. Hasibuan, A. Aboet, P.C. Eyanoer, S.V. Hutagalung
Statistical analysis was conducted to obtain the frequency distribution and mean value of each variable. The dependent t-test was used to compare the proportions. When data was not normally distributed, the Wilcoxon signed-rank test was used for analysis. P values lower than 0.05 were considered significant.
Effectiveness of polyene phosphatidylcholine and its combination with other drugs in patients with liver diseases based on real-world research
Published in Expert Review of Clinical Pharmacology, 2022
Ying Li, Anni Chen, Zhizhen Li, Xiuliang Cui, Guoqing Zhang
All statistical tests were conducted on a two-sided basis, and p value <0.05 was considered statistically significant (unless otherwise noted). Quantitative data were described by number, mean, standard deviation, median, minimum and maximum, and the upper and lower quartiles. Categorical data were described by the number and percentage of the cases. The main statistical methods were Wilcoxon signed-rank test, Chi-square test, and Mann–Whitney U test. The Wilcoxon signed-rank test tested the null hypothesis that two related paired samples come from the same distribution. In particular, it tested whether the distribution of the differences x – y was symmetric about zero. The Mann–Whitney U test was a nonparametric test of the null hypothesis that the distribution underlying sample x was the same as the distribution underlying sample y. It was often used as a test of difference in location between distributions. The Chi-square statistic and p-value were for the hypothesis test of independence of the observed frequencies in the contingency table observed. In phase III, propensity score matching (PSM) was applied to reduce the effects of selection bias and potential confounding factors, in order to achieve balance or comparability of treatment groups [17]. The controlling variables in PSM included age, gender, length of hospital stay, duration of medication, ALT interval before treatment, ALT interval after treatment, ALT categories before treatment, surgery (non-surgery; hepatobiliary surgery; and non-hepatobiliary surgery), and liver disease classification.
Consistent Iodine Status Assessment in Chinese Adults by Different Spot Urinary Iodine Concentrations in a Day Together With Corresponding Correction Coefficients
Published in Journal of the American College of Nutrition, 2019
Wendi Liu, Peng Zhang, Xin Zhao, Maocheng Sang, Xiaotong Liu, Lu Liu, Shiyan Liu, Haiyue Lin, Zhongna Sang
Excel 2007 was used for data entry and verification. All data analyses and statistics were carried out using SPSS 16.0 software (SPSS Inc.). The normality of the data was checked with the Kolmogorov–Smirnov test. Indicators were expressed as means ± standard deviation (SD) for normally distributed data, while other data were presented as a median (interquartile range). Bland–Altman method (22) was used to analyze the agreement of the UIC at different time points of the day and the UIC in the early morning. As commonly assumed (23), a Bland–Altman index of a maximum of 5% was interpreted as positive validation of the method of measurement. The correlation of the UICs was performed by Spearman rank correlation analysis. The non-normally distributed data of urinary iodine were converted to normally distributed data by ln-transformation (natural logarithm function), and then the ln-transformed values were used by linear mixed models for repeated measurement data analysis and pairwise comparisons. Wilcoxon signed-rank test was used to test whether there was a significant difference between two related pairs of the sample. p < 0.05 was considered to be statistically significant.
Relapse rates and predictors for relapse in a real-life cohort of IBD patients after discontinuation of anti-TNF therapy
Published in Scandinavian Journal of Gastroenterology, 2019
Steven J. Bots, Sabine Kuin, Cyriel Y. Ponsioen, Krisztina B. Gecse, Marjolijn Duijvestein, Geert R. D’Haens, Mark Löwenberg
Descriptive statistics were used to characterize the patient population. Results were provided as numbers (percentages) for discrete variables and median (range) for continuous variables and as frequencies and percentages for categorical variables. Differences in not normally distributed paired parameter were tested with the Wilcoxon signed rank test. Time to relapse and relapse rates were assessed using Kaplan–Meier analysis. Univariable Cox regression analysis was used to study potential predictors for relapse. Multivariable Cox proportional hazard analysis was not performed since data was collected retrospectively resulting in occasionally missing data. The proportional hazards assumption was tested by testing the product of time to relapse and each variable for interaction within the model with the same variable alone. A p-value of <.05 was considered to be statistically significant. All statistical analyses were performed using SPSS 24.0 software (IBM Corporation, Armonk, NY).