- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
i have more than 100 variables and use npar1way to estimate KS.
My aim is each variable to be tested based on the class variable
proc npar1way edf data=mydata noprint;
class source;
var &val;
output out=stat ks;
run;
Below is an example of the outcome for one variable. My question is, do i need to use the KS or the D statistic?
Kolmogorov-Smirnov Two-Sample Test (Asymptotic) | |||
KS | 0.038621 | D | 0.1039 |
KSa | 1.07377 | Pr > KSa | 0.1991 |
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
For comparing two different distributions, look at the graphic at Wikipedia, D is the size vertically of the black arrow, it is the maximum vertical difference between the two distribution. That's the one you want.
The documentation explains the difference: https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.4/statug/statug_npar1way_details24.htm
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
For comparing two different distributions, look at the graphic at Wikipedia, D is the size vertically of the black arrow, it is the maximum vertical difference between the two distribution. That's the one you want.
The documentation explains the difference: https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.4/statug/statug_npar1way_details24.htm
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content