turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Proc univariate/severity

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-05-2016 12:29 PM

I use Proc severity and proc univariate to estimate my parameters of continuous probability function.

I don't understand why statistics of Kolmogorov Smirnov is different with the two proc.

At the same time, statistics have the same definition in the documentation :

Other statistics like Anderson Darling and Cramer-von Mises are the same.

Accepted Solutions

Solution

09-06-2016
05:05 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-05-2016 03:34 PM

I'm away from the office so I can't check D’Agostino, R. and Stephens, M. (1986), *Goodness-of-Fit Techniques.*

However, my **guess** is that SEVERITY is using an asymptotic test statistic. Under the null hypothesis, sqrt(N)*D converges in distribution to another distribution (called the Brownian bridge). In contrast, UNIVARIATE is probably using table and formulas that describe the distribution of D itself.

For large samples, I would think that the p-values would be similar.

All Replies

Solution

09-06-2016
05:05 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-05-2016 03:34 PM

I'm away from the office so I can't check D’Agostino, R. and Stephens, M. (1986), *Goodness-of-Fit Techniques.*

However, my **guess** is that SEVERITY is using an asymptotic test statistic. Under the null hypothesis, sqrt(N)*D converges in distribution to another distribution (called the Brownian bridge). In contrast, UNIVARIATE is probably using table and formulas that describe the distribution of D itself.

For large samples, I would think that the p-values would be similar.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-05-2016 03:42 PM

You didn't post your code. According to the documentation, the two procedures will be equivalent only when edf=STANDARD is chosen in proc SEVERITY, in the absence of censoring.

PG

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-06-2016 09:25 AM

To follow up, the main difference seems to be that UNIVARIATE, which does not support censoring, uses one modification of the D statistic, whereas SEVERITY, which does support censoring, uses a slightly different modification.

The PROC SEVERITY formula

D*sqrt(n) + 0.19/sqrt(n))

appears on p. 113 in D'Agostino and Stephens. The formula handles the possibility of censored data and converges to an asymptotic distribution, which is tabulated in Table 4.4 and credited to Koztol and Byar (1975).

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

09-06-2016 09:27 AM

Thank you very much