What is a non -parametric test?
non -parametric test is a type of statistical testing of hypotheses that does not assume normal distribution. For this reason, non -parametric tests are sometimes referred to as distribution. The non -parametric test is more robust than a standard test, generally requires smaller samples, less likely to be influenced by remote observations and can be applied with a fewer assumptions. On the other hand, non -parametric tests may be less effective than their standard counterparts, especially if the population is actually distributed. Non -parametric testing is particularly effective for frequency and proportions. If there is a difference between a sample or parameter parameter - usually a diameter and/or scattering - a sufficiently large, the pact -up sample can be assessed as different from the control population. Such parametric testing requires that the parameters come from normal distribution.
It has been shown that sample size 30 or more will behave approximately as normal distribution, so thisThe requirement is generally assumed. However, if the assumption is not justified, the test results may not be valid. Non -parametric testing avoids this assumption.
Instead, non -parametric hypothesis testing normally examines data either by categorization or by ordering. If the populations of samples and controls are the same and if the data has been collected correctly, the results of chance are strictly differences between their categories or rankings. If there is a probability that these differences may have occurred by accident, also called p-value, it is less than the selected significant probability, usually either 5 percent or 1 percent, then the tester rejects the hypothesis that the population of the sample and controls is the same and concludes that they differ.
One common non-parametric test is the test of the chi-quadrate, which is used to compare the observed frequencies or proportions. When only poison is examinedA set of frequencies, often called it a test of goodies and is used to determine whether the frequency observed fits into the extent expected. For example, a good adaptation test could be used to determine whether the roulette table was manipulated by comparing the results of the results that the theory of probability predicts or to determine whether a cure for headache was effective by comparing people whose headache improved in medicine with the proportion of people whose headache improved when placebo. If two frequencies are examined, a non-parametric test of the chi-quadrates can be used to test correlation or independence. Political votes often seek correlation between social, economic or demographic factors and political beliefs, such as seeing whether there is a correlation between human education and whether it approves how the elected official performs.
Another non -parametric test is the Wilcoxon Rank Sum test that is generally used in the same situationsas standard testing of parametric hypothesis hypothesis. However, instead of examining the diameter of each sample, the Wilcoxon test examines the rank of each value if both samples are ordered from the least to the largest. If both samples are the same, each group should be scattered evenly through the evaluation. If one group is grouped at the lower or upper end of the evaluation, it means that both groups are different.
For example, assume that someone wanted to determine whether animated films are longer or shorter than non -nanimized films. For a standard test, the duration for an animated films sample and a sample of live action movies and compare the difference with the sample scattering would be determined. For the non -parametric test Wilcoxon, the film times are listed in order from the least to the largest and a number of animated film times are summarized.
a person could calculate the probability that the sum of the order would be such a size or a minor determination of the number of possible arrangements with a given valueAnd the total number of possible arrangements, which is simple due to sufficient calculation of rough force. With two small samples of six films there are already 924 possible arrangements of charts, a number that grows rapidly when movies are added. Alternatively, there are published tables that provide probabilities corresponding to the serial amounts for a given sample size. These can be found in statistical texts or online.
non -parametric testing is a growing field. It can be used in any way, in which conventional statistics were also used. However, applications are particularly common in social sciences and medicine, especially if normal distribution cannot be used.