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

From |
Sean <asmileguo@gmail.com> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
Re: st: RE: random effects, fixed effects and mixed effects models |

Date |
Sat, 15 Oct 2005 11:37:00 -0500 |

Thanks, I am just thinking about to read these two articles too:) On 10/14/05, Steve Stillman <stillman@motu.org.nz> wrote: > Along these lines, a correlated random effects (CRE) model specifies a (linear) relationship between the random effects and the Xs. The citations below discuss these models in more detail. This type of model can be estimated in stata using sureg on panel data in wide form (separate Ys and Xs for each panel dimension) with proper constraints. However, a completely unrestricted CRE model cannot be estimated using this approach because some of the constraints are non-linear (which stata doesn't allow). These models can also be estimates using a minimum distance approach which I think would be do-able in stata. > > Chamberlain, Gary (1984), "Panel Data", Handbook of Econometrics, Chapter 22 in Vol. 2, 1247-1318, Elsevier Science B.V. > Mundlak, Yair (1978), "On the Pooling of Time Series and Cross-section Data", Econometrica, 46, 69-85. > > -----Original Message----- > From: owner-statalist@hsphsun2.harvard.edu > [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of Sean > Sent: Saturday, October 15, 2005 12:05 PM > To: statalist@hsphsun2.harvard.edu > Subject: Re: st: RE: random effects, fixed effects and mixed effects > models > > > Then what relationship do you have in mind in regards to c(i) to x(i)? > if you have any idea, you can write it down explicitly and estimate > the structual model. > > > On 10/13/05, Shuaizhang Feng <fengsz@yahoo.com> wrote: > > Thanks. I guess what I am really looking for is an > > approach that could use both within and between > > variations to estimate the model but without assuming > > that the "random effect" is orthogonal to explanatory > > variables. Is this possible with STATA? > > > > Thanks a lot. > > > > --- Steve Stillman <stillman@motu.org.nz> wrote: > > > > > GLLAMM and XTMIXED are capable of estimating a > > > wide-range of models, but in general, these models > > > are in the random effects family and assume that c > > > is orthogonal to x. I am not sure what you mean by > > > "I have too few observations". There is nothing in > > > particular about fixed effects models that require > > > more observations than OLS or random effects > > > (assuming that you have at least two time periods of > > > data on many/most individuals/firms). > > > > > > Steve > > > > > > -----Original Message----- > > > From: owner-statalist@hsphsun2.harvard.edu > > > [mailto:owner-statalist@hsphsun2.harvard.edu]On > > > Behalf Of Shuaizhang > > > Feng > > > Sent: Friday, October 14, 2005 8:17 AM > > > To: statalist@hsphsun2.harvard.edu > > > Subject: st: random effects, fixed effects and mixed > > > effects models > > > > > > > > > Dear all: > > > > > > I have a general question on what to choose with a > > > panel data model like this: > > > > > > y(it)=x(it)b+c(i)+v(it) > > > > > > If I use xtreg to estimate this, when using random > > > effects models (re), then I am assuming that c is > > > orthogonal to x, right? But that is some assumption > > > I > > > am reluctant to make. I do not want to use fixed > > > effects estimates too because I have too few > > > observations. > > > > > > So could I use GLLAMM or XTMIXED. I suppose in those > > > c > > > is not assumed to be orthogonal to x. Is that true? > > > also, if both GLLAMM and XTMIXED work, which one > > > should I choose? I may have to use a two level mixed > > > effects model like this one. > > > > > > y(ijt)=x(ijt)b+c1(ij)+c2(i)+v(ijt). > > > > > > Thanks a lot. > > > > > > SZ > > > > > > > > > > > > > > > __________________________________ > > > Yahoo! Mail - PC Magazine Editors' Choice 2005 > > > http://mail.yahoo.com > > > * > > > * For searches and help try: > > > * > > > http://www.stata.com/support/faqs/res/findit.html > > > * http://www.stata.com/support/statalist/faq > > > * http://www.ats.ucla.edu/stat/stata/ > > > > > > > > > * > > > * For searches and help try: > > > * > > > http://www.stata.com/support/faqs/res/findit.html > > > * http://www.stata.com/support/statalist/faq > > > * http://www.ats.ucla.edu/stat/stata/ > > > > > > > > > > > > > > > __________________________________ > > Yahoo! Mail - PC Magazine Editors' Choice 2005 > > http://mail.yahoo.com > > * > > * For searches and help try: > > * http://www.stata.com/support/faqs/res/findit.html > > * http://www.stata.com/support/statalist/faq > > * http://www.ats.ucla.edu/stat/stata/ > > > > > > > -- > Best Regards, > > Sean > > * > * For searches and help try: > * http://www.stata.com/support/faqs/res/findit.html > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > > > * > * For searches and help try: > * http://www.stata.com/support/faqs/res/findit.html > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > > -- Best Regards, Sean * * For searches and help try: * http://www.stata.com/support/faqs/res/findit.html * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

**References**:**RE: st: RE: random effects, fixed effects and mixed effects models***From:*"Steve Stillman" <stillman@motu.org.nz>

- Prev by Date:
**st: from scientific to real numbers** - Next by Date:
**st: A vvvvvery naive question about "controll for"** - Previous by thread:
**RE: st: RE: random effects, fixed effects and mixed effects models** - Next by thread:
**st: RE: Re: Marginal effect after xtlogit, fe** - Index(es):

© Copyright 1996–2017 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |