| Author |
 Topic  |
|
|
msb
USA
3 Posts |
 Posted - 09/14/2011 : 3:34:26 PM
|
Hello,
I used Origin Pro to perform K-S test on 2 set of samples. I have D and Z values for both sets. In one case I get D=0.42 and Z = 0.02 and on teh second set D= 0.32 and Z=0.02 (both cases - result is : at 0.05 teh distributionsare significantly different). My question is - How can I interepret the diffrence in value i.e. Is D=0.32 much better than 0.42 or is there any meaning I can derive from it?
thanks for any suggestions..
MSB |
|
|
Echo_Chu
China
Posts |
Posted - 09/16/2011 : 03:25:34 AM
|
Hi, Mahabir
Z is the standardized statistic of D, which is equal to sqrt(n1+n2/n1n2)*D. To judge which test is better, I would suggest you to use p-value. It is the probalility for you to accept the null hypothesis. The smaller the p-value is, the less probablity for you to accept the null hypothesis. Use D is also fine, the larger the D value is, the less probbability for you to accept the nullhypothesis
Echo
OriginLab Technical Support
|
 |
|
|
msb
USA
3 Posts |
Posted - 10/13/2011 : 4:20:49 PM
|
Thank you, Echo. I was comparing my measured data with separate sets. So, I carried out two k-s tests (one with measured data and say set and another with measured data and set 2). If I understand correctly, I can say that since (e.g.) set 2 has higher value D, therefore set 1 is a better match with measured values. Is this a correct assumption?
Also, what does Prob>|D| signify?
Thanks
MB
quote:
Originally posted by Echo_Chu
Hi, Mahabir
Z is the standardized statistic of D, which is equal to sqrt(n1+n2/n1n2)*D. To judge which test is better, I would suggest you to use p-value. It is the probalility for you to accept the null hypothesis. The smaller the p-value is, the less probablity for you to accept the null hypothesis. Use D is also fine, the larger the D value is, the less probbability for you to accept the nullhypothesis
Echo
OriginLab Technical Support
|
|
|
| |
Topic |
|
|
|
help on K-S test output