Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: growth curve model with weights


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: growth curve model with weights
Date   Thu, 18 Jun 2009 13:15:04 -0500

On 6/18/09, P C <calicep@yahoo.com> wrote:
>  The sample I will be working on has a large sample size. It has over
> 15,000 cases. Will it take long for gllamm to estimate the growth curve model.

15000 students x 4 waves = 60000 entries... I've just been running a
two-level model (although a relatively more complicated) of the
similar total size, and it took about a day with Stata/MP4. This is
not a cup-of-coffee estimation exercise :((

>  When I specify student-level weight like gen wt2 = [student level weight],
> suppose I have three waves (w1weight; w2weight; w3weight), when I list them
> in a hierarachical order in the data file, the [student level weight] is no more
> a constant but changes within each respondent, right?

Technically nobody knows if there's a proper way of handling these.
But as a matter of putting those weights as inputs to -gllamm-, I
don't think it cares that much about weights not being constant
(although Sophia can correct me if she reads this :))

>  The other thing is that when I specify school-level weight,
> gen wt3 = [school level weight], I should use the method suggested
> in Sophia's paper to construct the school-level weight, right?

Frankly I don't remember what she was suggesting. You mean, the scaling issue?

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index