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
- /
- Joint Confidence Interval in SAS

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

05-02-2014 11:38 AM

Hello,

As part of the proc reg procedure, SAS has an option to report confidence intervals for each of the individual covariate coefficients. However, is there a way to report a joint confidence interval for the summation of two of the covariate coefficients?

For example, given the following regression model:

Wage = b0 + b1 Female + b2 Married + b3 Female x Married

I am interested in z = b1 + b3 since this gives me the average amount by which married females make more/less than single females. I would like to find a confidence interval for z to determine whether it is statistically significant.

Your help is appreciated. Thanks.

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

05-02-2014 12:48 PM

If you are looking for interaction effect. Try this one.

proc glm data=have;

class female married;

model wage = female married female*married;

run;

quit;

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

05-02-2014 12:56 PM

I think you'll need to move into a different proc, such as GLM with the estimate or means statement.

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

05-02-2014 01:07 PM

Thanks for the response, would these commands help me in testing H0: b1 + b3 = 0?

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

05-02-2014 01:18 PM

I believe so.

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

05-02-2014 01:20 PM

I see, thanks for the help. I'll look into it.

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

05-02-2014 01:22 PM

https://support.sas.com/resources/papers/proceedings11/351-2011.pdf

Good reference for these

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

05-02-2014 01:06 PM

Thank you for your responses, I have actually created a FemaleMarried (= Female * Married) interaction variable before using the proc reg command. From my understanding, the command:

test Female, FemaleMarried

would test H0: b1 = b3. However, I am interested in testing H0: b1 + b3 = 0. Is there a way to do this in SAS?

Also, I am running a similar model using the quantreg procedure, so I was wounding if I can apply the same test there?

Thanks again for your help.

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

05-08-2014 05:01 PM

I think you want b2+b3. The wage for single females under your model is b0+b1. For married females it is b0+b1+b2+b3. So the difference in wage is b2+b3.