ABSTRACT
Non-parametric tests are statistical methods that do not rely on assumptions about the underlying distribution of the data, making them robust alternatives to traditional parametric tests. This project investigates three widely used nonparametric tests: the median test, the sign test, and the Wilcoxon rank-sum test. The aim is to provide a comprehensive understanding of these tests, their applications, and their significance in data analysis. The study begins by exploring the background and principles of non-parametric tests, highlighting their advantages and limitations compared to parametric techniques. A critical review of relevant literature is conducted, covering topics such as the robustness of non-parametric tests, their performance in multivariate analysis, and their applications across various fields. The methodology section outlines the procedures for conducting the median test, sign test, and Wilcoxon rank-sum test, including hypothesis formulation, data preparation, test statistic calculation, and result interpretation. Real-world examples and data sets are analyzed using these non-parametric tests, demonstrating their practical applications and the interpretation of results. x The analysis chapter delves into the findings, discussing the implications of the non-parametric tests' performance and their suitability for different research scenarios. Comparative studies with parametric counterparts are presented, shedding light on the strengths and limitations of non-parametric approaches. Finally, the project provides recommendations for future research, such as conducting comprehensive comparative studies, exploring applications across diverse fields, investigating small sample size performance, and extending nonparametric tests to multivariate analysis. The study contributes to a deeper understanding of non-parametric tests and their potential applications in empirical research, particularly when dealing with non-normal or skewed data distributions.
