Getting Wald Chi-Square test statistics instead of Z test statistics when using probit

[Nick Hallman]

Hello All,

This is my first post to statalist so please forgive me if I make a mistake of etiquette.

First let me state that I am using STATA version 12.1.

I am currently attempting to replicate a study that uses probit regressions. In the study's reported results, test statistics are listed as "Wald Chi-Square" for each individual coefficient. In my replication, my coefficient estimates and p-values are similar, which seems to suggest my analysis is correct. However the test statistics output by STATA for the probit command are "Z statistics" and are uniformly much lower in my replication than in the study I am attempting to replicate. Does anyone know if there is a way to have STATA output Chi-Square test statistics for each coefficient (I know that it provides a summary Chi-Square statistic to test the null that none of the coefficients are significant).

Thanks in advance for any help you are able to provide.

[Phil Clayton]

Can't you just square the z-statistics?

[Joe Canner]

Nick,

If you prefer a command that produces Wald Chi-square values directly, try the following:

probit y x1 x2 ....
test x1
test x2
...

See here for more information: http://www.ats.ucla.edu/stat/stata/faq/nested_tests.htm

Regards,
Joe

[Steve Samuels]

Welcome to Statalist, Nick. Before your next post, I suggest that you read http://www.statalist.org/forums/help from the bottom up.


Steve

[Phil Clayton]

If you prefer a command that produces Wald Chi-square values directly, try the following:

probit y x1 x2 ....
test x1
test x2
...

Nice one. In fact you can run all the tests with one command:

probit y x1 x2 ...
test x1 x2 ..., mtest

[Maarten Buis]

In fact, squaring the z-statistic and using test are equivalent as you can see in the example below

// create an example model
sysuse auto, clear
probit foreign price mpg

// use -test-
test price mpg, mtest

// square z-statistics
di as txt "Chi^2 for price " ///
   as result (_b[price]/_se[price])^2
di as txt "Chi^2 for mpg " ///
   as result (_b[mpg]/_se[mpg])^2

Here I use tricks from the Stata tip (Buis 2007) to get at the z-statistics. Using tricks from the same Stata tip you can compute the Chi square values for all parameters in one go using Mata:

// use Mata to do this for all
//parameters in one go
mata
b = st_matrix("e(b)")'
V = st_matrix("e(V)")
se = diagonal(cholesky(diag(V)))
(b :/ se ):^2

// or equivalently
(b:^2) :/ diagonal(V)
end

Maarten Buis (2007) Stata tip 53: Where did my p-values go? The Stata Journal, 7(4):584-586.
http://www.stata-journal.com/article...article=st0137