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
- /
- What is SAS equivalent of ivprovit in STATA?

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

03-08-2015 09:39 AM

***Apologies for cross-posting***

Hi,

I am interested in running the SAS equivalent of ivprobit in STATA. Let’s say I want to run a model with a binary outcome (Y), two continuous endogenous explanatory variables (p1 p2), and one continuous exogenous explanatory variable (p3). I have three instruments (z1 z2 z3).

proc qlim data=a;

model Y = p1 p2 p3 / discrete (dist=logit);

model p1 p2 = p3 z1 z2 z3;

run;

- Are the SAS codes above correct?
- I want to test the null hypothesis that the two endogenous explanatory variables (EEV) are in fact exogenous. I understand that when there are two or more EEVs, the test becomes the joint likelihood ratio test of whether corresponding correlations are 0 or not. How do I modify the SAS codes above to run this test?
- Recall I have three instruments (z1 z2 z3). I want to test the validity of all three of my instruments by specifying the OVERID option. How do I modify the SAS codes above to incorporate the OVERID option?
- I want to test the strength of all three instruments used. I understand that in 2SLS one can use the F statistic based on the Stock-Yogo recommendation. Can the F statistic be applied in this case (with binary outcome) as a test of weak instruments? If so, how do I modify the SAS codes above?

Thanks in advance for any help you can render!

Best,

Elizabeth