[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Problem with gllamm adaptive quadrature

From   Maarten buis <[email protected]>
To   [email protected]
Subject   Re: st: Problem with gllamm adaptive quadrature
Date   Sat, 1 Dec 2007 09:02:29 +0000 (GMT)

--- Barth Riley <[email protected]> wrote:
> I am running a multilevel logistic model using gllamm. After initial
> estimation of the fixed effects, gllamm then proceeds to perform
> adaptive quadrature. It appears that gllamm never proceeds past
> iteration 0. After a very long (hours) delay, I get the following
> error message: "initial vector: matrix must be dimension 28"
> This is the syntax for gllamm I am using:
> gllamm x d1_1-d17_1, i(id Section) nip(8) /// 
> link(mlogit) expand(obs choseno) weight(wt) /// 
> nocons adapt 
> There are 28 terms in the initial set of coefficients, so why would I
> get this error message? 

Does the set of dummies contain the reference category, i.e. do you
have one dummy for every question, instead of leaving one of the
dummies out as the reference category? If you included the reference
category than Stata is likely to drop one of the dummies when
calculating initial values. The vector of initial values will than
contain the correct number of 27 coefficients, while your model
statement contains the incorrect number of 28 coefficients.

-- Maarten

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Support the World Aids Awareness campaign this month with Yahoo! For Good
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index