Re: st: xtserial error: "no observations"

[Rodrigo Badilla <rodrigo.badilla@gmail.com>]

Hi all,

i used numeric data  type , below is a description of my data with codebook:

fecha: numeric daily date (long) : this is my panel data (fixed effect)
 colocaciones: numeric (float)
tasa_min numeric (float)
tasa_max numeric (float)
imacec   numeric (float)
icv  numeric (float)
ipc  numeric (float)
cae   numeric (float)
ingreso,numeric (float).

When i write this command the output was:

 xtserial  colocaciones  tasa_min tasa_max imacec   icv  ipc  cae
ingreso, output
no observations
r(2000);

 alternatively i used xtpcse and xtgls command to get some idea about
autocorrelation in my data panel

Cross-sectional time-series FGLS regression

Coefficients:  generalized least squares
Panels:        homoskedastic
Correlation:   no autocorrelation

Estimated covariances      =         1          Number of obs      =       386
Estimated autocorrelations =         0          Number of groups   =         8
Estimated coefficients     =        15          Obs per group: min =        46
                                                               avg =     48.25
                                                               max =        49
                                                Wald chi2(14)      =   1575.78
Log likelihood             = -6270.753          Prob > chi2        =    0.0000

i know that my data number is small and probably the problem could be
in other side, I am working in that...

regards
Rodrigo

2013/4/8 Nick Cox <njcoxstata@gmail.com>:
> You need to tell us more about exactly what you typed and the
> incidence of missing values.
>
> In addition, note that -xtserial- is user-written, so that you are
> expected to explain that and say where it comes from.
>
> I doubt that having 386 observations is itself a problem.
>
> Nick
> njcoxstata@gmail.com
>
> On 8 April 2013 15:34, Rodrigo Badilla <rodrigo.badilla@gmail.com> wrote:
>
>> Thanks for you reply and information, i did checked my variables and
>> are not strings... probably the missing values could be the answer...
>>
>> anyway i have been read that autocorrelation test with xtserial should
>> be used with panel data with a large number of data, could be this in
>> relation with my problem ( i have only 386 samples)
>
> 2013/4/8 Nick Cox <njcoxstata@gmail.com>:
>
>>> This typically means that
>>>
>>> either some variable you think is numeric is really string
>>>
>>> or some variable contains missing values, prohibiting some calculation.
>>>
>>> We can't say more without knowing more about your data.
>
> On 8 April 2013 13:52, Rodrigo Badilla <rodrigo.badilla@gmail.com> wrote:
>
>>>> I am triying to run a autocorrelation tests with xtserial command,
>>>> lamentably only get this error output: "no observations" previously i
>>>> set my data with xtset command, my data set have 386 samples.
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/