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
- /
- SAS Programming
- /
- Base SAS Programming
- /
- when outcome variable is not normally distributed,...

Topic Options

- 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
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

06-07-2016 12:19 PM

I previously used regression analysis and there is a significant difference. Then I realize my dependent variable is not normally distirbuted, it is very skewed.

So I use proc genmod. Anyother analysises, please give advice.

below is the error message

1523 proc genmod data = have;

1524 model abb= mbb /dist=negbin;

1525 run;

NOTE: Non-integer response values have been detected for the

Negative Binomial distribution.

WARNING: No output destinations active.

NOTE: Algorithm converged.

NOTE: The negative binomial dispersion parameter was estimated

by maximum likelihood.

NOTE: PROCEDURE GENMOD used (Total

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

Posted in reply to wenling

06-07-2016 12:52 PM

"dist=negbin" requests maximum likelihood parameter estimation to explain things that can be counted. Counts are integers. What is *abb*?

PG

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

Posted in reply to PGStats

06-07-2016 12:56 PM

abb is my dependent variable, mbb is my independent variable

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

Posted in reply to wenling

06-07-2016 01:04 PM

Advice: Take an Introductory course in Probability. Learn the difference between discrete and continuous distributions.

PG

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

Posted in reply to wenling

06-07-2016 01:49 PM

You didn't mention which type of regression you initially performed. Many types do not require the data to be normally distributed but that the errors are normally distributed.

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

Posted in reply to ballardw

06-08-2016 10:14 AM

it is linear regression