Stcox value missing

[Gonçalo Gonçalves]

Hello stata friends!

So, I was doing an stcox test on survivability and a problem came in.

As I saw my test output through outreg2, one of the variables called "aspirations" appeared as "o.aspirations" with no value. In my data editor that variable shows lots of values. I don't know what happened.

If you know how to solve this problem I would be very thankful!

Best regards
Gonçalo

[Svend Juul]

Please show exactly what you typed, and exactly what Stata responded. What was Stata's direct output (as opposed to the output produced by outreg)?

[Gonçalo Gonçalves]

I typed the following:

stcox i.industry industrysalesgrowth_pre Ifirm Iind lagROA aspirationsfirm aspirationsindustry lagdfb lagslackindex, robust
outreg2 using reg_stcox, tstat dec(4) ctitle(Model 1) excel append

Stata gave me what I just attached. Hope it's what you asked.

[Svend Juul]

I don't see "o.aspirations" anywhere in the attached output.

[Gonçalo Gonçalves]

Not in the stata output, but in the outreg2 it appeared o.aspirationsfirm.

In the stata output it appears "omitted" in "aspirationsfirm".

I don't know why it is omitted, I have plenty of values in that variable.

If you can help

[Marcos Almeida]

Hi, Gonçalo,

I wonder if this text would please you: http://www.stata.com/support/faqs/st...ted-variables/

It was written by James Hardin and has a rather suggestive title:"Why do estimation commands sometimes omit variables?"

Hopefully that helps.

Best,

Marcos

[Rich Goldstein]

I'm confused; according to #3 above, the first part of the command you typed is

stcox i.industry

but this implies that the outcome variable is categorical which makes no sense - please clarify

[Marcos Almeida]

Under survival analysis, - stset - provides time and "failure" variables. That leaves (grosso modo) only the covariates to be inserted after - stcox -.

Worth mentioning in this particular thread, -outreg2 - is an user-written command.

I checked it out and we can perform the commands after survival analysis.

According to the help file (http://repec.org/bocode/o/outreg2.html) , in order to get the HR you are supposed to insert "eform" among the options:

eform specifies coefficients to be reported. This corresponds to or option for logit, clogit, and glogit estimation, irr for poisson estimation, rrr for mlogit, hr for cox and stcox hazard models, and eform for xtgee. The default form of stcox is the hazard rate form; use the eform to report it.

Best,

Marcos

[Gonçalo Gonçalves]

Thank you very much, but I still couldn't solve my problem.

Can it be a multicollinearity problem?

[Marcos Almeida]

Hello Gonçalo,

Have you accessed - and read - the text I provided in #6, exactly on "Why do estimation commands sometimes omit variables"?

Best,

Marcos