Sounds, as far as Stata is concerned, like a case for -rolling-.
The constraint that -flag == 1- might better be applied
retrospectively.
Nick
[email protected]
Siyam, Amani
> Below is a segment I have of a child's record that includes
> 22 visits:
>
> +---------------------------------------------------------------+
> | visit sex age tage weight tweight thresh flag |
> |---------------------------------------------------------------|
> | 19 2 549 548 12.1 12.1 7 0 |
> | 20 2 583 609 12.35 . 7 1 |
> | 21 2 670 670 13.85 13.85 7 0 |
> +---------------------------------------------------------------+
>
> I need your help to structure a STATA code that would perform
> a linear fit within a child's record as follows:
>
> By child:
> take the observed weight when flag==1, the
> weight[_n-1] and weight[_n+1];
> regress the 3 points weights on the 3 points age
> (age[_n], age[_n-1], age[_n+1])
> and return the beta-coefficient of that model
> or
> predict tweight for tage(=609) if flag==1 using that model
*
* 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/
