Free Other Non-Parametric Tests Essay Sample

Question #1

Nonparametric methods are designed for those situations where the researcher does not know anything about the parameters of the study population (hence the name of the method - non-parametric). In more technical terms, nonparametric methods are not based on the estimation of parameters (such as the mean or standard deviation) describing the distribution of quantities of interest.

Therefore, these methods are sometimes also called parameter-free or free distribution.

Nonparametric methods allow manipulating the data "low quality" from small samples with variables, about the distribution of which little or nothing is known.
Usually, when there are two samples (e.g., men and women) and you want to compare the mean value of a variable of interest, you are using a t-test for independent samples. Nonparametric alternatives parametric test for two independent groups are:

U Mann-Whitney test

Criterion series of Wald-Wolfowitz
Two-sample Kolmogorov-Smirnov
If you have multiple groups, you can use the analysis of variance (ANOVA). Its nonparametric analogs are:
Ranked ANOVA Kruskal-Wallis test
The median test
If you want to compare two variables related to the same sample (e.g., mathematical students' progress at the beginning and end of the term), it is usually used a t-test for dependent samples.

Alternative non-parametric tests are:

Wilcoxon paired comparisons
sign test

This is three examples when we should use non-parametric tests instead of parametric.

Question #2
Each nonparametric procedure in the module has its own advantages and disadvantages. For example, two-sample Kolmogorov-Smirnov test is sensitive not only to the difference in the position of the two distributions, such as differences in average, but is also sensitive to the form of distribution. Wilcoxon paired comparison suggests that it is possible to rank the differences between the compared observations. If it is not, it is better to use the sign test. In general, if the result of this study is important (for example, providing assistance to people whether certain very expensive and painful therapy?), It is always advisable to use a variety of non-parametric tests. Perhaps the test results (different tests) will be different. In this case, you should try to understand why different tests gave different results. On the other hand, nonparametric tests have less statistical power (less sensitive) than their parametric counterparts, and if it is important to detect even slight deviations (for example, whether a given food additive dangerous to humans) should be especially careful to choose the test statistic.

Question #3

Mann-Whitney – 3. Independent sample t-test
Kruskal-Wallis – 1. One-way ANOVA
Wilcoxon Signed Rank – 1. Paired sample t-test
Friedman – 5. Repeated measure ANOVA
Spearman’s Rank – 4. Pearson correlation
References