Dear all
i'm really disappointed just having one response to my question on
standard error for dummy. coef.
So being more precise., I try again:
I made a lognormal regression (using function lm) with a dataframe
containing quantitative and qualitative variables.
Log Y= AX+E
Then I want to make a new regression using each coefficient of one
factor in the previous model as observations. In order
to have an unbiased estimator of this coefficient using Laurent or
Golberger correction, I need its standard error.
exp(a_i-(sderr(a_i)/2)) = BX+E
In the summary(lm) I have the standard error of contrast coefficients.
If a used dummy.coef(lm), I have the real coefficient
of the predicting model, but no standard error. And I can't use
model.tables(lm) asking for standard error
because my datframe is unbalanced!!!
Does anyone know of such an S-plus function, or such a way to do
this?
Thanks in advance.
(Splus5.0 with unix)
Stephanie Mahevas
Ifremer Nantes
smahevas@ifremer.fr
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|