Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Outlier diagnostics for tobit (postestimation)

From   Maarten Buis <>
Subject   Re: st: Outlier diagnostics for tobit (postestimation)
Date   Fri, 19 Oct 2012 10:28:09 +0200

On Thu, Oct 18, 2012 at 9:56 PM, Nick Cox wrote:
> This is very similar to questions asked within the last week in
> various threads started by Ebru Ozturk.
> One short answer is
> 1. Not much is obviously available, as the idea of residuals is
> problematic if the response is bounded by definition.
> 2. You can always try simulating to see what would happen with what
> you regard as a reasonable data generation process.

See: <>
for an example on how to do that.

> 3. You should be clear that you really do have a tobit problem and not
> one analogous to a logit or probit problem.

You can compute the dfbetas using -jackknife-, see the post by Steve
Samuels: <>

Applying that to -tobit-:

*----------------- begin example -----------------
// prepare data
sysuse auto, clear
gen wgt = weight/2000
label var wgt "weight in short tons"
replace mpg = 17 if mpg <= 17

// compute dfbetas
local xvars wgt foreign price

jackknife _b, keep: tobit mpg `xvars', ll
tobit mpg `xvars', ll

foreach z of local xvars {
di e(N_sub)
	gen dfb_`z' = ///
	(1/_se[`z'])* (model_b_`z'-_b[`z'])/(e(N)-1)
sum dfb*
*------------------ end example ------------------
(For more on examples I sent to the Statalist see: )

-- Maarten

Maarten L. Buis
Reichpietschufer 50
10785 Berlin
*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index