-xtmixed- is likely to be a better choice for a linear model with
continuous response, as it is notably faster. If I were fitting this,
I would put a bunch of B-splines in age (I think they became official
with Stata 10) into my model formulations, and then figure out the
derivatives. I am pretty sure there are better methods aimed at
derivatives directly though, but it's been a long a while since I read
proper nonparametric regression books. Otherwise, your discipline may
indeed have a crowned model or method that everybody uses, but that's
likely outside of most statalisters immediate span of attention.
On 8/10/07, Nick Cox <[email protected]> wrote:
> This fills in some of the background science.
>
> It seems to leave unclear:
>
> 1. Quite how demographic data are to be related to
> individuals. More broadly, whether you intend fitting
> individuals separately or together. My guess is
> the latter, but it helps to spell things out.
>
> 2. Whether you have specific discipline-approved
> growth curve functional forms in mind or are going
> to use more general methods.
>
> Even if you answer these questions, I will not be
> able to add very much more, if anything, but others
> might start tuning in and give you specific advice.
>
> Nick
> [email protected]
>
> Joseph Wagner
>
> > The research involves craniofacial skeletal growth measurements. The
> > data consist of several angle measurements at junctions between bones
> > and length measurements of specific areas of bone. We have
> > demographic
> > data as well. We get measurements (using x-rays) from individuals
> > annually over their lifetime. We want to know at which point the
> > greatest rate of growth occurs and the peak growth velocity for
> > individuals and the mean for our entire sample. We also want to know
> > the rate of bone recession in older persons.
> >
> > Nick Cox wrote:
> > > It isn't obvious from this information alone that -gllamm-
> > > is the only or even the best way to do it. That said,
> > > there will be ways to get key properties of fitted
> > > growth curves, even if only numerically. Quite how
> > > will depend on what you do precisely, so better answers
> > > might depend on further information.
> > >
> > > Nick
> > > [email protected]
> > >
> > > Joseph Wagner
> > >
> > >
> > >> I have data composed of yearly jaw xray measurements over a 20 year
> > >> period. I wish to do some growth curve analyses of these
> > >> data so I plan
> > >> to use -gllamm- but I was wondering, is there a way to get
> > >> the mean peak
> > >> velocity and the mean peak height of the growth curve?
>
> *
> * 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/
>
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: Please do not reply to my Gmail address as I don't check
it regularly.
*
* 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/
