How to get R-squared for a scobit (skewed regression) model?

[Som Nwegbu]

Hello,

I'm working with Stata 13 and complex survey data. I'm using scobit regression for my skewed binary outcome (experience of intimate partner violence: Yes/No). Running my model does not return all the usual statistics one would get with a logit model, so I'm wondering how to get them; specifically R-squared, AIC and BIC.

Thanks in advance for your help.
Som

[Marcos Almeida]

Hello Som,

According to the Stata Manual

. estat ic

Will present the AIC and BIC.

With regards to getting the r-squared when Stata "refuses" to provide it, I wish to share with you this masterclass , aptly entitled "How can I get an R-squared value when a Stata command does not supply one?")

Hopefully that helps.

[Rich Goldstein]

I think that the issue here is the "complex survey data" - at least for AIC and BIC; you don't show us your actual command but if it is something like "svy: scobit ..." then it is probably not an MLE and therefore estat ic won't work and you can't get AIC/BIC as they are for MLE's - Nick's faq may well help but you need to be careful with r-squared for these types of models; e.g., if I remember correctly, there are at least 6 definitions of r-squared just for logistic - the situation may be simpler, or more complicated for scobit

[Marcos Almeida]

Rich Goldstein Thanks for pointing out the "svy:" issue. Absolutely right. I confess I overlooked that when reading #1, and I assumed the command was just "scobit".

[Richard Williams]

Yes, the scobit is a red herring here. You don't get pseudo R^2, BIC, or AIC after svy: logit or svy: probit either:

webuse nhanes2f, clear
svy: probit diabetes i.female i.black weight age
estat ic
svy: logit diabetes i.female i.black weight age
estat ic
svy: scobit diabetes i.female i.black weight age
estat ic
testparm i.female i.black weight age

The only real difference with scobit that I can see is that for some reason it doesn't give you an overall F test either, but you can get that with a testparm command.

Some of the key differences between svy and non-svy analysis are summarized in

http://www3.nd.edu/~rwilliam/xsoc73994/SvyCautionsX.pdf

[Som Nwegbu]

Rich Goldstein you're exactly right. Stata keeps telling me I can't use the estat ic after I run my model. I'll check out Nick's FAQs as you suggested. Thanks!

[Marcos Almeida]

Som Nwegbu : indeed, the link to Nick's FAQ was shared in #2.


Below, an excerpt, but you may consider as take-home message for similar cases:

Why is R-squared not supplied?

If Stata refuses to give you an R-squared, there may be a good explanation other than that the developers never got around to implementing it. Perhaps the R-squared does not seem to be a good measure for this model, on some technical grounds. You have to consult the literature or an expert to take this further, unless you are an expert, in which case you probably disagree with the other experts.

[Som Nwegbu]

Thanks Marcos. I did check it out, as was suggested, and found it helpful.

Som