Nonparametric statistics is a branch of statistical inference whose calculations and procedures are based on unknown distributions.
Nonparametric statistics are not very popular. However, there is a very extensive literature on it. The problem that nonparametric statistics aims to solve is the lack of knowledge of the probability distribution.
In other words, nonparametric statistics try to find out the nature of a random variable . For, once you know how it behaves, perform calculations and metrics that characterize it. What is nonparametric statistics?
Nonparametric tests, also known as free distribution tests, are those that are based on certain hypotheses, but the observed data do not have a normal organization. Generally, nonparametric tests contain statistical results that come from their ordering, which makes them easier to understand.
Nonparametric tests have some limitations, among them is that they are not strong enough when a normal hypothesis is fulfilled. This can cause it not to be rejected even if it is false. Another of their limitations is that they need the hypothesis to be changed when the test does not correspond to the question of the procedure if the sample is not proportional.
Some of the characteristics of nonparametric tests are:
- It is a difficult measurement method to apply.
- Hypothesis testing is necessary.
- The hypotheses are strict.
- The observations must be independent.
Objective of nonparametric statistics
There are different types of probability distributions that parametric statistics work on . Now, when we do not know what type of probability distribution a variable corresponds to, what calculations do we use? What is nonparametric statistics?
That is, when we do not know the probability distribution of a data set, we must make statistical inferences with non-parametric procedures.
Of course, if we don’t know how a random phenomenon is distributed, what should we do? Very easy. Our mission will be to try to know how it is distributed. To try to find out what type of distribution a certain phenomenon has, we have a series of tests available to help us do so. Among the most popular non-parametric tests are: What is nonparametric statistics?
- Binomial test
- Anderson-Darling test
- Cochran test
- Cohen kappa test
- Fisher test
- Friedman test
- Kendall’s test
- Kolmogórov-Smirnov test
- Kuiper test
- Mann-Whitney test or Wilcoxon test
- McNemar test
- Median test
- Siegel-Tukey test
- Signs test
- Spearman’s correlation coefficient
- Wald-Wolfowitz test
- Wilcoxon signed rank test
All these tests are intended to tell us if a random variable is distributed in one way or another. For example, a possible result could be: the random variable X is distributed at the rate of a normal distribution. What is nonparametric statistics?
All being said, the results are not infallible. To perform non-parametric tests we must have statistical samples . Therefore, the results can be reliable but they do not have to be 100% perfect.
Advantages of nonparametric tests
The advantages of nonparametric tests are:
- They can be used in different situations, since they do not have to comply with strict parameters.
- Their methods are generally simpler, which makes them easier to understand.
- They can be applied on non-numeric data.
- It facilitates obtaining the most important and appropriate particular information for the investigation process. What is nonparametric statistics?
Disadvantages of nonparametric tests
The disadvantages of nonparametric tests are:
- They are not systematic tests.
- The distribution varies, making it difficult to select the correct choice.
- The application formats are different and causes confusion.
- Information may be lost because the collected data is converted into qualitative information.
- Heavier fonts and backing may be required.
These types of statistics can be used without sample size or estimation of any related parameter for which no information is available. Since the assumptions are minor, they can be applied in multiple ways. What is nonparametric statistics?