| T O P I C R E V I E W |
| luciana12 |
Posted - 06/25/2014 : 3:34:22 PM
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
Hi all,
I performed the KS test for the first time in OriginPro8.using two independant samples, in the statistics window.
I got these data. D=0.102 Z=0.0031 Prob>(D)= 3,401 E-10
please if you can help me, to know whether the 2 groups I compared are different.
|
| 3 L A T E S T R E P L I E S (Newest First) |
| Echo_Chu |
Posted - 06/27/2014 : 02:35:44 AM
Hi,
The K-S test is performed based on the empirical distribution function while t-test is to compare the means. Base on the result you get I assume the mean of your data are similar but if you plot a histogram from your data, you will find the shape of histograms are different.
Perhaps you can try Mann-Whiney test, which can be access from Statistics: Nonparametirc Tests: Mann-Whitney Test. . It is the nonparametric counterpart to the two-sample t test
Echo
OriginLab Technical Service |
| luciana12 |
Posted - 06/26/2014 : 10:56:45 AM
Indeed, I want to compare the distribution of cumulative frequency between two groups. in OriginPro, i made 2 columns, each column represents individual events. aroung 2100 events per each group.
I got a footnote as you illustrated in your reply. according to the sentence in below of the table, it is written. at 0.05 level, the two distributions are significantly different.
p.s: When I compared the response of the two groups using T.student test, no statistical difference is observed, therefore, I was suggested to use KS test to compare the distribution.
Thank you
HTML is OFF
Forum Code is on |
| Shirley_GZ |
Posted - 06/26/2014 : 04:30:20 AM
Hi,
Two-Sample Kolmogorov-Smirnov Test is to perform the test based on the empirical distribution function, which tests whether the distribution of the response is the same between the two sample.
Usually there should be a footnote to indicate whether the two distribution are significantly different.
Anyway, based on the P value you have, Prob>(D)= 3,401 E-10, that means the two distributions are significantly different at the 0.05 level.
You said you want to know if 2 groups you compared are different. It depends on what the 2 groups are.
If you want compare the median values of two groups, I will suggest you to use the mann-whitney test.
Originlab Technical Service Team |
|
|
