Panel data - heteroskedasticity test

[Stefania Curion]

Hello,

I have a panel and I would like to test for heteroskedasticity. I'm trying to understand this post http://www.stata.com/support/faqs/st...tocorrelation/ but I really don't understand (I'm new in using Stata...). What the dots "..." after xtgls stand for? What is e(N-g)? N and g are numbers or what else? ... I'm really in troubles because I don't understand the answers I find online (many things are supposed to be "known" but they are not trivial or obvious to me... ). I would like to know if there is a test for heteroskedasticity with panels and how does it work.. My aim is to decide whether to use a Linear model (FE or RE) or a PPML model, and I need to know if there is heteroskedasticity.

Thanks in advance

S.C.

[Carlo Lazzaro]

Stefania may find the following thread useful: http://www.stata.com/statalist/archi.../msg01258.html

[Mark Schaffer]

Stefania - Stata's online help is very good, and links you directly to the manual entry. If you type "help xtgls", you will find a basic description of the command, including (at the bottom) what the macros e(N), e(N_g) etc. mean. In the upper right-hand corner of the help window you will see a drop-down menu for "Also See". The first item is the Stata manual entry for xtgls in PDF form. You'll find lots of detail there.

[Pouya Tabesh]

Hi. I have a related question (I think it is related) so I bring it up as a part of this older post:
I am dealing with panel data (84 panels, unbalanced, average panel size: 4, 10 IVs some categorical and some continuous).
I am testing heteroskedasticity by using the LR test mentioned in the famous FAQ page (http://www.stata.com/support/faqs/st...tocorrelation/). However, the never-ending iterations just keep going (with a repeated pattern of numbers).

How can I deal with this?
Thanks
Pooya