Parametric Tests
Uploaded by
ScribdTranslationsParametric Tests
Uploaded by
ScribdTranslationsWhat are parametric tests?
Parametric tests are a statistical tool used for analyzing population factors.
This sample must meet certain requirements such as size, as the larger it is, the more accurate the calculation will be.
This method requires specifying the distribution shape of the studied maternal population. It may be about,
example, of a normal distribution, as generally occurs when dealing with large sample sizes. In general,
these tests can only be applied to numerical variables.
Parametric tests are based on the distribution law of the variable being studied. Although there are
many types of distribution laws, these are based on the normals, which have two parameters: the mean and the deviation
standard. Enough to know the probability.
Conditions that must be met by parametric tests
A parametric test must meet the following elements:
Normality: The analysis and observations obtained from the samples should be considered normal. For this
goodness of fit tests must be performed where it describes how well the
observations and how they differ from the expected values.
Homoscedasticity: The groups must present uniform variables, that is, that they are homogeneous.
Errors: The errors that occur must be independent. This only happens when the subjects are
assigned randomly and are distributed normally within the group.
Types of parametric tests:
Z value test of the normal distribution
Student's T-test for related data (dependent samples)
Student's T-test for unrelated data (independent samples)
Student-Welch T test for two independent samples with non-homogeneous variances
Bartlett's Chi-Squared Test to demonstrate the homogeneity of variances
F test (analysis of variance or ANOVA).
Advantages and disadvantages of parametric tests
Some of the advantages of parametric tests are:
They are more efficient.
They are perceptible to the characteristics of the information obtained.
Errors are very unlikely
Probabilistic calculations are very accurate.
The disadvantages of parametric tests are:
The calculations are difficult to make.
The data that can be observed is limited.
Parametric tests are a useful tool for multiple situations, calculations, and interpretations.
Thanks to their common use, it is possible to observe the results obtained through an analysis. They are a method
very powerful if the conditions for its application are met. However, researchers must keep in mind that
If the variables being studied do not follow a normal law, they cannot be chosen.
You might be interested in reading: Marketing Strategies: The Chi-squared, the Corrected Residuals, and the Perceptual Maps
they mean sales.
Yáñez Cortés Pedro David 1353
What are non-parametric tests?
Non-parametric tests, also known as distribution-free tests, are those that are based on certain
hypothesis, but the observed data do not have a normal organization. Generally, non-parametric tests
they contain statistical results that come from their arrangement, which makes them easier to understand.
Non-parametric tests have some limitations, one of which is that they are not strong enough.
when a normal hypothesis is fulfilled. This can cause it not to be rejected, even though it is false. Another of its
Limitations are 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 non-parametric tests are:
It is a measurement method that is difficult to apply.
It is necessary to conduct hypothesis testing.
The hypotheses are strict.
The observations must be independent.
Types of non-parametric tests and their applications
The types of non-parametric tests are:
Sign test of a sample
Wilcoxon signed-rank test
Mann-Whitney U test
Kruskal-Wallis test
Mood median test
Friedman test
Advantages of non-parametric tests
The advantages of non-parametric tests are:
They can be used in different situations, as they do not have to meet strict parameters.
Generally, their methods are simpler, which makes them easier to understand.
They can be applied to non-numeric data.
It facilitates the acquisition of the most important and appropriate specific information for the research process.
Disadvantages of non-parametric tests
The disadvantages of non-parametric tests are:
They are not systematic tests.
The distribution varies, which complicates selecting the right choice.
The application formats are different and cause confusion.
It is possible that information may be lost because the collected data becomes qualitative information.
It may be necessary to have sources and a backup with more weight.
This type of statistics can be used without the sample size or the estimation of any related parameter.
of which there is no information. Since the assumptions are fewer, they can be applied in multiple ways.
Yáñez Cortés Pedro David 1353
Differences between non-parametric tests and parametric tests
Non-parametric tests are those that analyze data that do not have a specific distribution and
They are based on a hypothesis, but the data is not organized in a normal manner. Although they have some limitations, they account for...
with ordered statistical results that facilitate their understanding.
Parametric tests, on the other hand, are based on the laws of normal distribution to analyze the elements of a
sample. Generally, they only apply to numerical variables and for their analysis a large population must be maintained, since
which allows the calculation to be more accurate.
Differences between non-parametric tests and parametric tests
Non-parametric tests Parametric tests
Greater statistical power. Lower statistical power.
They are applied to categorical variables. They apply to normal or interval variables.
They are used for small samples. They are used for large samples.
The form of data distribution is not known. Its data distribution is normal.
They don't make many assumptions. They make many assumptions.
They demand a lower validity condition. They demand a higher validity condition.
Higher probability of errors. Lower probability of errors.
The calculation is easier to do. The calculation is complicated to make.
The hypotheses are based on ranges, median, and frequency. The hypotheses are based on numerical data.
of data.
The calculations are not exact. The calculations are too precise.
Consider the missing values to obtain information. Do not take into account the missing values to obtain
information.
Greater statistical power. Lower statistical power.
They are applied to categorical variables. They apply to normal or interval variables.
Before applying non-parametric tests or parametric tests, it is important to understand aspects such as the objective.
from the research, the size of the population and the scale that will be used to measure the data.
It is likely that the data does not meet the requirements needed for a parametric test and a non-parametric one has to be chosen.
non-parametric, that is, when the sample size is small or the distribution is not normal.
Another factor that is necessary to consider is that parametric tests can use an abnormal distribution, but
A non-parametric one has extremely strict assumptions that cannot be ignored.
Finally, if the sample size is small, it is most likely that the results will not be obtained if a
non-parametric test. When the population is not really large, the probabilities of identifying an effect
significant are smaller.