“Para” means closely related and “metric” means measure. Nonparametric statistical method, as the name suggests, has a different approach from the parametric statistics analysis. Nonparametric tests can be used with both the nominative and ordinal parameters of measurement. Its major contrast with parametrical method lies in the fact that it does not follow the usual distribution of curve or require assumptions to be made about the data format.
When these assumptions are not true of the main data, it is possible to apply a compatible transformation. In the cases where even transformed data don’t comply with the assumptions, use of t-tests, or parametric analysis is becomes invalid. You need to consider the raw data you have collected, the type of comparison you wish to make in your dissertation before choosing nonparametric statistics.
This is perhaps the most widely used nonparametric methods of all time. It caters to a number of services, the most notable of them is its use a mode of comparison between a single sample with an assumed value. It is useful when paired or single t-tests might be generally applied. Aspects like the relative calculation of risk factor, detrimental effect and development of an issue are represented through the study.
Nonparametric tests are used for various services of different fields such as in behavioral science where there is no basis of assumption of certain types of distributions. As mentioned earlier, nonparametric test is applicable for both ordinal and nominal levels of measurement. Note that parametric tests should not be altered by nonparametric when the former is more suitable. The later should be exercised only when the assumptions of the parametric tests cannot be satisfied. It is a perfect method when the number of the data is minimal and when there is no basis of assumption for certain types of distributions.