artifact_id;status_id;status_name;priority;submitter_id;submitter_name;assigned_to_id;assigned_to_name;open_date;close_date;last_modified_date;summary;details;"Hardware";"Product";"Operating System";"Component";"Version";"Severity";"Resolution";"URL"
422;1;"Open";3;903;"Ott Toomet";100;"Nobody";"2009-04-26 04:18";"";"2009-04-26 04:18";"tobit*: test multicollinearity";"Create a test for testing the invMillsRatio multicollinearity in tobit2 & tobit5 models (see changes in version 1008).";"None";"None";"None";"None";"None";"trivial";"None";""
1278;1;"Open";2;903;"Ott Toomet";903;"Ott Toomet";"2011-02-10 16:31";"";"2011-02-12 21:25";"Discuss Maddala/Powers/stata MoveStay differences";"(by Nigel Melville)
Movestay is widely used in the Stata community.
Here is a document from Dan Powers that provides clear explanation of switc=
hing models in stata:
http://umich.com.cn/teaching/download/20092/CDAcoursematerials/handouts/Sel=
ectRegQ.pdf
If you visit the document by Powers, you will see a simulation example on p=
age 12-13 of switching regression. The data generating process is given on =
page 9.
When I run the same model as that on page 12/13,
I get the following:
[cid:3379615321_13702523]
As you can see, the rho1 (selection 1 FALSE) is .3143.
In Stata movestay (widely used program) the FALSE selection is rho2 and is =
given by -.3143.
There is a different sign here.
My conclusions, based on the discussion in Powers on pages 8-9 on ""Interpre=
ting the Covariance Parameters"" is that
1. Powers starts with the positive-signed disturbance term in selection =
equation on page 6 (deviating from Maddala), so the rho2 must be changed in=
sign for correct interpretation.
2. Your sampleSelection package starts with the negative signed disturban=
ce (consistent with Maddala) and so the rho1 can be immediately interprete=
d without any sign change.
Anyhow, once again I appreciate this discussion and I think we are getting =
closer to enlightenment :-)
Best, Nigel
Hi Nigel
On 3 February 2011 23:15, Melville, Nigel wrote:
When you say ""all error terms are added (rather than subtracted)"" this does=
not seem to be consistent with Maddala's (1983) formulation which I have copied b=
elow. Note the negative sign on selection equation 9.62.
";"All";"None";"All";"None";"None";"minor";"None";""
1739;1;"Open";3;903;"Ott Toomet";100;"Nobody";"2012-01-11 06:46";"";"2012-01-11 06:46";"probit does not accept 'maxMethod'";"maxMethod argument gives an error with probit. Probably related to binaryChoice stuff.";"All";"None";"All";"None";"None";"minor";"None";""
1741;1;"Open";3;903;"Ott Toomet";100;"Nobody";"2012-01-12 06:02";"";"2012-01-12 06:02";"probit: length of the 'start' argument";"Probit (binaryChoice) should check the length of 'start' argument. Currently get fuzzy error messages.";"None";"None";"None";"None";"None";"None";"None";""
1762;1;"Open";3;903;"Ott Toomet";100;"Nobody";"2012-01-24 14:18";"";"2012-01-24 14:18";"intReg: test for missing $estimate from maxLik";"maxLik may return object with $estimate = NULL.
Test handling of it.";"None";"None";"None";"None";"None";"None";"None";""
6017;1;"Open";3;903;"Ott Toomet";100;"Nobody";"2015-01-20 04:58";"";"2015-01-20 04:58";"2-step estimator in tobit5 is wrong";"Found piece of code in heckit5fit.R:
sigma1 <- sqrt( se1^2 + ( betaL1*delta1)^2)
sigma2 <- sqrt( se2^2 + ( betaL2*delta2)^2)
It should be
sigma1 <- sqrt( se1^2 + betaL1^2*delta1)
sigma2 <- sqrt( se2^2 + betaL2^2*delta2)
heckit2fit should also be checked.";"None";"None";"None";"None";"None";"minor";"None";""
6031;1;"Open";3;23752;"florian oswald";100;"Nobody";"2015-02-24 14:13";"";"2015-02-24 14:13";"standard error of prediction and weights";"hi,
just wanted to submit a feature request: it would be great to have a standard error on the prediction of the expected value of the outcome equation. in terms of your notation, if this is the expected outcome:
E[y|X,Z,w=1] = X %*% beta + rho * sigma * dnorm( Z %*% gamma ) / pnorm( Z %*% gamma )
then I want
Var( E[y|X,Z,w=1] ). Stata uses the delta method to estimate Var(rho*sigma). not sure if implementing this is more or less work than just bootstrapping the whole thing.
http://www.stata.com/manuals13/rheckman.pdf (page 16)
in that document, there's also a simple way to incorporate sampling weights. this would be very useful when working with survey data, which always comes with sampling weights.
thanks!";"Macintosh";"None";"None";"None";"None";"None";"None";""
6599;1;"Open";3;318857;"Cedric Edano";100;"Nobody";"2018-09-17 08:13";"";"2018-11-17 12:35";"Issues after estimating mvProbit ";"Hello,
I have issues after estimating multivariat probit using mvProbit package on R. In fact, i am working on decisions of marital statuts so i have five responses.
After estimation, when i am using command summarize, i get this result:
> summary(estimat)
Call:
mvProbit(formula = cbind(single, free_union, married, separated,
divorced) ~ age + age2 + location + living_child + no_education +
no_religion + age_frist_sex + poor + gour + akan, data = newom,
method = ""BHHH"", algorithm = ""GHK"", nGHK = 1000)
Coefficients:
Error in printCoefmat(coef(x), digits = digits) :
'x' must be coefficient matrix/data frame
When i use command print:
print(estimat)
Call:
mvProbit(formula = cbind(single, free_union, married, separated,
divorced) ~ age + age2 + location + living_child + no_education +
no_religion + age_frist_sex + poor + gour + akan, data = newom,
start = NULL, method = ""BHHH"", algorithm = ""GHK"", nGHK = 1000)
Coefficients:
NULL
I want help for this; how can make estimation and after get results.
Best regards.";"Macintosh";"None";"MacOS X";"None";"None";"critical";"None";""