2.13.2.3 Nonparametric Tests

Hypothesis tests are parametric tests when they assume the population follows some specific distribution (such as normal) with a set of parameters. If you don't know whether your data follows normal distribution or you have confirmed that your data do not follow normal distribution, you should use nonparametric tests.

Origin provides support for the following X-Functions for non-parametric analysis, they are available in OriginPro

Name Brief Description
signrank1 Test whether the location (median) of a population distribution is the same with a specified value
signrank2/sign2 Test whether or not the medians of the paired populations are equal. Input data should be in raw format.
mwtest/kstest2 Test whether the two samples have identical distribution. Input data should be Indexed.
kwanova/mediantest Test whether different samples' medians are equal, Input data should be arranged in index mode.
friedman Compares three or more paired groups. Input data should be arranged in index.

As an example, we want to compare the height of boys and girls in high school. 

//import a sample data
newbook;
fname$ = system.path.program$ + "Samples\Statistics\body.dat";
impasc;

//Mann-Whitney Test for Two Sample
//output result to a new sheet named mynw
mwtest irng:=(col(c), col(d)) tail:=two rt:=<new name:=mynw>;

//get result from output result sheet
page.active$="mynw";

getresults tr:=mynw;

//Use the result to draw conclusion
if (mynw.Stats.Stats.C3 <= 0.05); //if probability is less than 0.05
{
	type "At 0.05 level, height of boys and girls are differnt.";
	//if median of girls height is larger than median of boy's height
	if (mynw.DescStats.R1.Median >= mynw.DescStats.R2.Median) 
		type "girls are taller than boys.";
	else
		type "boys are taller than girls."
}
else
{
	type "The girls are as tall as the boys."
}
English | Deutsch | 日本語