FUNCTION OF NON-PARAMETRIC STATISTICS
The function of sample that does not depend upon the parameter
and secondly, it refers to the statistical tests that are used for testing
various non-parametric hypotheses
Non-parametric statistics are used to define the characteristics of the
population when the data is measured on either nominal or ordinal scale.
For instance, if the sample observations are obtained in terms of ranks,
then the median is computed to know the average of the sampled data.
In this case, mean cannot be computed as a measure of central tendency
because the data has neither been measured on interval nor ratio scale.
In such data set, instead of standard deviation, quartile deviation is used
as a measure of variability. Thus, where median is used as a measure
of central tendency, the quartile deviation is used as a measure of
variability.
The median and quartile deviation are known as non-parametric statistics.
Similarly, other statistics like highest score, lowest score, and ordered set
are also known as non-parametric statistics.
As it is now clear that the non-parametric statistics also refers to the
statistical tests for testing the non-parametric hypothesis, the distinction
between non-parametric and parametric statistical tests shall be discussed
first.
The non-parametric statistical tests do not assume anything about the
form of the distribution of the population from which the sample is drawn
for testing the non- parametric hypothesis.
Therefore, the non-parametric statistical tests are known as
distribution free methods also.
While using most of the parametric statistical tests, one assumes that the
distribution of the population is normal.
Besides this, several other assumptions also need to be satisfied before
using it.
If these assumptions are violated, then non-parametric tests can
efficiently he used to draw conclusions.
Corresponding to most of the parametric statistical test, there exists a
non- parametric test. For example, corresponding to t-test and one-way
analysis of variance (ANOVA) in parametric one has an equivalent test
like Mann-Whitney test and Kruskal wallis test.
CHI-SQUARE TEST
Developed by Karl Pearson
Also known as general purpose statistics.
The chi sqaure test is used when the data are expressed in terms of
frequencies or percentages.
BASIC IDEA BEHIND USING CHI-SQAURE TEST
1) (a) we have a theory of any kind concerning how our cases should be
distributed,
(b) we have a sample which shows how the cases actually are distributed.
(c) we want to know whether or not the differences between theoretical
frequencies and observed frequencies are of such size that these
differences might reasonably be attributed to the chance factor.
Use of chi-square test
1) First, chi-square may be used as a test of equal probability hypothesis.
By equal probabilay hypothesis, we mean the probability of having the
frequencies in all the given categories a equal.
Suppose, for example, 100 students answer an item in an attitude scale.
The item has five categories of response options strongly agree, agree,
neutral, disagree and strongly disagree.
According to the equal probability hypothesis, the expected frequency of
responses given by 100 students would be 20 in each.
The chi-square test would test whether or not the equal probably becomes
tenable. If the value of the chi-square test is significant, the equal
probability hypothes becomes untenable and if the value of the chi-square
is not significant, the equal probabilty hypothesis becomes tenable.
In testing equal occurrence hypothesis, one is interested to test whether
all the events are equally probable in the population. Consider a situation
where one is interested to know whether the three sports, say, cricket,
football and basketball are equally popular in a college. In this case, one
shall test the deviation of the observed responses of the students (in terms
of observed frequencies) with that of the theoretical (expected)
frequencies under the assumption that all the three sports are equally
popular.
2) TESTING THE SIGNIFICANCE OF INDEPENDENCE
HYPOTHESIS/SIGNIFICANCE OF ASSOCIATION BETWEEN
TWO ATTRIBUTES.
By independence hypothesis is meant that one variable is not affected by
or related to another varible and hence these two varibles are independ
