# st: RE: xttrans2: matcell and prob option

 From "Nick Cox" <[email protected]> To <[email protected]> Subject st: RE: xttrans2: matcell and prob option Date Wed, 5 Jan 2005 13:49:22 -0000

```Falko Juessen

> I calculate transition probabilities (multiplied by 100) for
> a variable
> "per" using Nick Cox's -xttrans2- command (for information type -ssc
> describe xttrans2-).
> When specifying the -prob- option, the following error occurs:
>
> .         xttrans2 per, prob
> matcell() required

The -prob- option is for saving a matrix of probabilities. -xttrans2-
can't do that unless you also specify a matrix. I guess the help
should make that clearer.

> When I include the matcell option, the following error occurs:
>
>
> .         xttrans2 per, prob matcell(T)
>
>            |                          per
>        per |        .5        .75          1       1.25        1.5 |
> Total
> -----------+--------------------------------------------------
> -----+-------
> ---
>         .5 |      3.70      88.89       7.41       0.00       0.00 |
> 100.00
>        .75 |     46.55      24.14      29.31       0.00       0.00 |
> 100.00
>          1 |      0.00      13.39      72.32      14.29       0.00 |
> 100.00
>       1.25 |      0.00       0.00      29.31      63.79       6.90 |
> 100.00
>        1.5 |      0.00       0.00       0.00      31.25      68.75 |
> 100.00
> -----------+--------------------------------------------------
> -----+-------
> ---
>      Total |     10.33      19.56      43.17      21.40       5.54 |
> 100.00
>
> 1:  operator invalid
> r(198);

> The displayed numbers are still frequencies but not probabilities. The
> matrix T contains the same values (frequencies) as displayed in the
> output. Does someone have an idea what I am doing wrong when
> using the -
> xttrans2- command?

Some Stata matrix function is gagging on your implied row
and column name 1.25, so you are being thrown out of
-xttrans2- before the matrix of frequencies has been converted
to probabilities. If you multiply -per- by 100 and live with
the consequences of inaccurate labelling you can get a probability
matrix, I guess.

Nick
[email protected]

*
*   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/
```