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
- /
- Poisson distribution - Is there a Test for normali...

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
- Permalink
- Email to a Friend
- Report Inappropriate Content

01-07-2016 12:58 PM

Hi there:

I am looking for any test to evaluate the normality for a data set in Possion..

Thanks

Accepted Solutions

Solution

01-07-2016
02:00 PM

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

Posted in reply to jonatan_velarde

01-07-2016 01:28 PM

Do you mean "Is there a test to determine whether data are Poisson distributed?" (You can't check for "normality" unless you think the data are normally distributed.)

If so, the answer is yes. You can fit a Poisson distribution to the data and examine the goodness of fit statistics.

All Replies

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

Posted in reply to jonatan_velarde

01-07-2016 01:27 PM

I don't understand the question. If the data are Poisson, then why test for normality? Maybe you are wondering if you counts are large enough that the Poisson is adequately approximated by a normal distribution? If so, then use PROC UNIVARIATE and look a the output for tests for normality.

proc univariate data=... normal;

var y;

run;

Solution

01-07-2016
02:00 PM

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

Posted in reply to jonatan_velarde

01-07-2016 01:28 PM

Do you mean "Is there a test to determine whether data are Poisson distributed?" (You can't check for "normality" unless you think the data are normally distributed.)

If so, the answer is yes. You can fit a Poisson distribution to the data and examine the goodness of fit statistics.