Re: Re: st: repeated time values within panel

[Nick Cox <njcoxstata@gmail.com>]

This adds some substantive detail which may help others suggest
_particular_ techniques of panel data analysis, if that is what you
are seeking.

Otherwise the only Stata issue you raise is one that has been raised
twice before and answered by me twice before.

Nick

On Wed, Sep 28, 2011 at 10:27 AM, S.H. Former <S.H.Former@uvt.nl> wrote:
> Thanks for your reaction and sorry for the vague problem description! Maybe I should explain the study I am performing so it is more clear what I want. I am investigating what the effect is of knowledge characteristics on internal knowledge transfer (internal knowledge transfer is the dependent variable). I have certain characteristics that could be of influence (value of knowledge, complexity of knowledge and tacitness of knowledge). These variables I am operationalizing using patent-usage in the period 1976-2006. But certain companies transfer knowledge more often than other which leads to more experience and probably a more efficient knowledge transfer. So this experience I want to include using panel data analysis. Now I come to my problem; a lot companies assigned more than one patent per year and Stata gives the command repeated time values within panel. I understand that this is a problem when panel data analysis is trying to be performed but do you know whether the!
 re!
>  is a solution for this in Stata or that I should adjust the dataset / use another method of analysis?
>
> Thank again for the time and effort!
>
> best regards,
>
> Sytze
>
> -----Original Message-----
> From: Nick Cox <njcoxstata@gmail.com>
> To: statalist@hsphsun2.harvard.edu
> Date: Tue, 27 Sep 2011 15:19:34 +0100
> Subject: Re: st: repeated time values within panel
>
> I stand by my earlier answer, ignored here, at
>
> http://www.stata.com/statalist/archive/2011-09/msg01198.html
>
> The fact of multiple patents per company-year is not a problem except
> in so far as you might have a fixed idea that you must apply time
> series methods which require balanced panels. Of course, you could
> reduce your data to counts of patents for each company and year and
> that could be handled using such methods, but that may or may not be
> helpful in your project. That aside, what you describe seems a natural
> and unsurprising feature, but as in your previous you don't give any
> details on what you intend to do, so useful comment is difficult.
>
> By the way, if knowledge transfer is neither dependent or independent,
> how is it to be taken accounf of?
>
> Nick
>
> On Tue, Sep 27, 2011 at 2:37 PM, S.H. Former <S.H.Former@uvt.nl> wrote:
>
>> I want to perform a panel data analysis but I have some problem with my data. For each entity (companies in my case) I am looking at the patents they have assigned. Only some companies have assigned multiple patents in certain years. So for example, company X has assigned 4 unique patents in 1999 while Company Y assigned 143 and Company Z only 1. Stata gives the command that there are repeated time values within the panel.
>>
>> So to clarify it, when a company assigned multiple patents I still want to incorporate them all because it tells me something about the experience in knowledge transfer. The experience of knowledge transfer is taken into account by performing a panel data analysis (it is not a dependent or independent variable). Has anyone
>>
>> Has anyone got a solution for me so that I still can use the database and incorporate all the patents so that experience is considered as well?

*
*   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/