Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: xtmixed with nonrtolerance. What happens?


From   Maarten Buis <maartenlbuis@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: xtmixed with nonrtolerance. What happens?
Date   Fri, 24 Jun 2011 15:54:43 +0200

You seem to be under a time constraint, and getting this fixed is not
going to be easy as that variable looks extremely problematic. Can't
you change your dependent variable to a dummy, did that country export
products of animal x or not, and than model it with -xtmelogit-?

-- Maarten

On Fri, Jun 24, 2011 at 3:26 PM, "Lukas Bösch" <L.Boesch@gmx.de> wrote:
> sum quantity
>
>    Variable |       Obs        Mean    Std. Dev.       Min        Max
> -------------+--------------------------------------------------------
>    quantity |      6192    8545.829    116704.3          0    4019264
>
> and here transformed:
>  sum centquantity2
>
>    Variable |       Obs        Mean    Std. Dev.       Min        Max
>
> -------------+--------------------------------------------------------
>  centquanti~2 |      6192    2.17e-09           1  -.0732263   34.3665
>
>
> The dependent variable is however not a real count. It was a count in the original dataset, or i think so, but i had to transform the original data and i am not sure wheter it still is a count as it involves dezimals, and transformed it even has negative numbers. I tried the xtmepoisson but it did not work at all. I didn't had the time to work on it as i had to go to work, so i will try it later. But for 1 thing, I surely cant work with the z-scores in xtmepoisson as negative numbers aren't alloud. The untransformed dependent variable also seems not to fit to the xtmepoisson criterias and it doesnt even start to calculate.
>
> The sample population are 40 countries. I started with 130 but because of the complexity of the data and because of all the missings i am working with 40 countries now. I have been told that i need a completely balanced data, and as i still have a representativ sample of countries i thought it would be allright. Having no missings in the data cant be wrong, was my thought.
>
> Thanks
>
> Lukas
>
>
> -------- Original-Nachricht --------
>> Datum: Fri, 24 Jun 2011 08:54:45 -0400
>> Von: Anders Alexandersson <andersalex@gmail.com>
>> An: statalist@hsphsun2.harvard.edu
>> Betreff: Re: st: xtmixed with nonrtolerance. What happens?
>
>> I too would like to see the frequency distribution of your original
>> dependent variable (i.e., before standardization).
>> and, like Maarten, I would use xtmepoisson instead of xtmixed if the
>> outcome variable is a count.
>>
>> Lukas, what is your sample population, e.g, the 40 or the 130 countries?
>> Why do you need completely balanced data (no missing values)?
>>
>> Anders Alexandersson
>> andersalex@gmail.com
>>
>> Lukas wrote:
>> > 1) Concerning the missings, i took care to delete all data i don't have
>> the complete time series for.
>> > This means that i had to drop from 130 countries over a 20 years period
>> to 40 countries over a 15 years
>> > period. On the other hand, there are deffinitely no missing values.
>>
>> On Fri, Jun 24, 2011 at 6:30 AM, Maarten Buis <maartenlbuis@gmail.com>
>> wrote:
>> > --- On Fri, Jun 24, 2011 at 12:11 PM, "Lukas Bösch" wrote:
>> >> One more information:
>> >>
>> >> When i run the modell only with the complete export time series.
>> Without any zeroes in the dependent data, the output looks like this:
>> > <snip>
>> >>                 sd(R.genus) |   2.93e-06   .0221629
>>       0
>> >
>> > I would not trust that result. First, leaving out 0s means you select
>> > on your dependent variable, which is a very bad idea. Second, your
>> > results indicate that the genus variation is de facto 0 (2.93e-6 =
>> > 0.000000293), which given your context is very unlikely.
>>
>> *
>> *   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/
>
> --
> Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir
> belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de
> *
> *   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/
>



-- 
--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   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   |   Site index