st: RE: Pearson Goodness of Fit

 From "Nick Cox" To Subject st: RE: Pearson Goodness of Fit Date Sun, 16 Jul 2006 17:46:39 +0100

```You can put your sequences in string variables.

bysort id (t) : gen y_seq = string(y[1])
by id : replace y_seq = y_seq[_n-1] + string(y) if _n > 1
by id : replace y_seq = y_seq[_N]

and similarly for the other variable.

Then you select just observation for each panel

egen tag = tag(id)
tab y_seq yhat_seq if tag, chi2

Nick
n.j.cox@durham.ac.uk

Nikolaos Kanellopoulos

> I want to estimate the Pearson Goodness of fit after a Random Effects
> Probit. To do that I need to count the frequencies of each
> sequence that my
> actual variable (y) and its prediction (yhat)have.
>
> For instance in the following example y takes 3 sequences by
> id and t 0001,
> 0000, 0011 with frequencies 2 ,1, 1 respectively and yhat
> takes 0001, 0000,
> 1000, 0011 each with frequency 1. My problem is that I can't count the
> different sequences that y and yhat follow.
>
> Any help is more than welcome.
>
> 	+-------------------+
> 	id   t   y   yhat
> 	-------------------
> 1.	1   1   0      0
> 2.	1   2   0      0
> 3.	1   3   0      0
> 4.	1   4   1      1
> 5.	2   1   0      0
> 	-------------------
> 6.	2   2   0      0
> 7.	2   3   0      0
> 8.	2   4   1      0
> 9.	3   1   0      1
> 10.	3   2   0      0
> 	-------------------
> 11.	3   3   0      0
> 12.	3   4   0      0
> 13.	4   1   0      0
> 14.	4   2   0      0
> 15.	4   3   1      1
> 	-------------------
> 16.	4   4   1      1
> 	+-------------------+

*
*   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/
```