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
- /
- General Programming
- /
- P-Value for McNemar's Test

- 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

05-19-2016 10:34 PM

Hi,

I have calculated the McNemar's Chi-Square (ChiSq variable) for my data, however when I calculate the associated P-value using formula : p= 1 - cdf('CHISQUARE',ChiSq,1) the value seems to be off ( about half lower) from the other statistical packages.

Thank you for your help.

Accepted Solutions

Solution

05-20-2016
07:39 PM

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

05-19-2016 11:12 PM

McNemar's test is available (asymptotic and exact versions) in proc freq when you specify option AGREE.

PG

All Replies

Solution

05-20-2016
07:39 PM

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

05-19-2016 11:12 PM

McNemar's test is available (asymptotic and exact versions) in proc freq when you specify option AGREE.

PG

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

05-19-2016 11:16 PM

Thank you!

Can I just use a P - value formula though?

Thank you.

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

05-19-2016 11:36 PM

Just checked - Proc freq works fine, the result is the same as my formula for p value though. Thank you for your help.

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

05-20-2016 12:20 PM

Learner25 wrote:

Hi,

I have calculated the McNemar's Chi-Square (ChiSq variable) for my data, however when I calculate the associated P-value using formula : p= 1 - cdf('CHISQUARE',ChiSq,1) the value seems to be off

( about half lower) from the other statistical packages.

Thank you for your help.

Consider whether SAS is doing a TWO tail test and your other package or you are calculating for a one-sided test.

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

05-20-2016 07:39 PM

Thank you. The other test uses binomial distribution.