Dear all,
I think this is a very relevant discussion, so let me share some thoughts.
I agree with Jack that setting up a general discussion of econometrics could end in a "do-my-homework-for-myself" thing.
I'm not an econometrician, just an amateur user of econometrics. As such, hundred of times I was in desperate need of quick answers for econometric issues of mixed origins. My problem is I do not understand Gretl or Econometrics? My problem is that I have not read the Gretl manual enough?
I have tried to avoid specific econometric discussions in Gretl list, because I think everyone has been polite and patient, I should not gone too far using others people time.
I think Google could be a much better research find place than the proposed list, although some people have generously share their knowledge on a lot of issues.
I think there are two roads that maybe should be explored, and I am trying to sum up some of the emails that others send upon this topic. Please, do not feel the following reflections as critics, nor demanding more work to those that gently spend their time to simplify our work. I expect them to be path for improving Gretl. My reflections will be targeted upon the use of the Gretl console or the construction of inp files.
1) The Gretl documentation some times is obscure on the specific way that one should proceed in order to do things work. This came in two ways. First, sometimes the commands itself have sometimes a narrow explanation. Some have examples and others not, and as in Murphy's Laws the day you lack examples when you need some specific command. As an example, I found that Gretl documentation is scarce on how commands work in panels. But not how to do econometric, but how to manipulate and how work specific commands in panels. Second, in the Gretl Guide, which I think is also a good reference econometric text, some inp examples are good for the topic at hand (and work!), but some lines need more explanation. I regret not having a good example right now, and this weaken my argument. This also happen when they are references to C++ commands.
2) The Gretl User guide abounds in econometric explanations (how to decide between random and fixed effects) which is excelent! But it lack explanation on the specific way it is calculated. As an example, in the command meantest --unequal-vars I was asked on which is the specific formula used. I have no answer but "in the usual way, I think". In panels I have another answer from a colleague on the specific way Gretl estimate the parameters (which I do not remember now, sorry) and I could not find the answer.
The first point could be easily handled, and those who do not program could make some suggestions. The second one I think could require more work in order to translate the code to a readable text.
I think this two lines of work could improve Gretl and also spread its use. When I talk to people which use STATA in order to convince them that Gretl is not less than STATA, I found a barrier in the details of the specific calculations (among some spread fear that associate free with not god as payed).
I hope this email contribute to the discussion,
Best regards,
Leandro2011/5/4 Talha Yalta <talhayalta@gmail.com>
Although I have brought up the issue in the first place, I must
emphasize that I don't have a strong opinion on this issue. Various
people have expressed their opinions on the subject and we see there a
pros and cons. I think the pros may outweigh the cons, is slightly.
Then again, I don't have much experience in running such lists.
For the sake of completeness of the discussion, I would like to
mention that such a list, if moderated, could be very useful for
improving gretl's documentation. There are experienced econometricians
in the list and their answers for my initial coint2 related question
provide a very thorough view on the subject, missing in the
documentation. (By the way, I am grateful for all those responses.)
An often cited problem regarding open source software is insufficient
documentation. For example, gretl's user manual is dwarfed by Stata's
3 volume Base Reference Manual.
So I think it may be worth considering to set up a list for which
subscription is limited to a balanced mix of experienced
econometricians and maybe less experienced econometricians who know
gretl well and actively involve in its development (such as yours
truly). It is expected that this mailing list would have a very low
message frequency and the posts will be edited and used for improving
a topic or a discussion in gretl's documentation. Of course, the
participants to the list would also be acknowledged in the user's
manual. This would provide additional incentive to participate in
addition to knowing one is doing something helpful for the developers,
mainly Allin and Jack.
If successfull, this could even be a case study in open source
development process.
Best regards,
Talha
--
On Wed, May 4, 2011 at 3:34 AM, MICHAEL BOLDIN <mboldin@temple.edu> wrote:
> I was part of asking about making new mailing list but after reading
> the replies and recalling my experiences running such a list at one
> time, it is probably not worthwhile. If it becomes popular because
> experienced GRETL users do respond we will surely encourage a lot of:
> 'do my econometrics homework for me' (as someone else points out) or
> worse: 'how do I apply a technique for which I have no idea about how
> it works but want to use it for my dissertation'. Many of the
> remainder are likely to be naive and show the person asking the
> questions did not bother to first look in their econometrics textbook
>
> Most important: Anyone motivated to make it work (inside or outside
> of the GRETL community) could easily start such a list independently
> of GRETL. A Google
> group is easy to set up. And if Allin, Sven, Jack and others involved
> in development and enhancements to GRETL were to encourage something
> linked to GRETL, the value added is unlikely to exceed the
> distraction.
>
> So I change my vote to letting things stand as they are-- everyone
> seeing a few econometrics questions related to GREL every once in a
> while.
>
> However, I think an item in the GRETL Help menu for 'Check and use
> the GRETL developers listserve' would be a worthwhile addition.
> Instead of making it a direct 'Send question to listserve' I suggest
> encouraging new users to search the listserve first by simply pointing
> to http://lists.wfu.edu/mailman/listinfo/gretl-devel and adding the
> ground rules to this page.
>
> --my 2 cents
> _______________________________________________
> Gretl-users mailing list
> Gretl-users@lists.wfu.edu
> http://lists.wfu.edu/mailman/listinfo/gretl-users
>
“An expert is a person who has made all the mistakes that can be made
in a very narrow field.” - Niels Bohr (1885-1962)
--
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users