[R] plm ? tests of poolability ? error: insufficient number

[Millo Giovanni]

Hello Cecilia, 

nice hearing from you again. I must restate a couple of my old hints,
though ;^)
1) please always put the authors c/c, as we are not guaranteed to browse
through the r-help every day
2) please provide reproducible examples.

As example(pooltest) keeps working fine, as do some other cases I tried
(Grunfeld data etc.), I don't know what the problem is but evidently
your data are peculiar in exposing it. 

Sure the data numerosity must be enough, but maybe you have something
which doesn't vary along some dimension, and gets dropped? or some
groups which are all-NA? Do the "random" and "pooling" options crash as
well?

You might use traceback() to better identify the problem, and/or send a
complete example to the list or even to me separately if it's too heavy.

Cheers,
Giovanni

Giovanni Millo
Research Dept.,
Assicurazioni Generali SpA
Via Machiavelli 4, 
34132 Trieste (Italy)
tel. +39 040 671184 
fax  +39 040 671160 

### original message ###
------------------------------

Message: 106
Date: Thu, 10 Dec 2009 10:53:01 +0000
From: "Cecilia Carmo" <cecilia.carmo at ua.pt>
Subject: [R] plm ? tests of poolability ? error: insufficient number
	of observations
To: r-help at r-project.org
Message-ID: <web-92270892 at controller.cgpmail.ua.pt>
Content-Type: text/plain;charset=iso-8859-1;format="flowed"

Hi everyone!

I?m running the pooltest in plm package, like this
pooltest(cstfin12~lmaccdiscrz+lcobjur+lliq+lcollateral+ldimensao, 
data = dados3,
model = "within")

But I got the following error:
Error in FUN(X[[1L]], ...) : insufficient number of 
observations

My data is an unbalanced panel with 20907 observations 
(6971 individuals and years 2001 to 2007).   This is not 
enough? Could anyone help me?

Cec?lia Carmo (Universidade de Aveiro ? Portugal)

### end original message ###

Ai sensi del D.Lgs. 196/2003 si precisa che le informazi...{{dropped:13}}