In article <rrg.1463.000A2C0F at aber.ac.uk> rrg at aber.ac.uk (Roy Goodacre) writes:
>It is worth pointing out that Genstat which I use to analyse pyrolysis mass
>spectral data is *not* very user friendly since it is an old program that has
>been moved from mainframe to the PC with a lot of problems.
I agree that Genstat is not easy to learn, though neither is SAS.
However once you have learned to use it, it is astonishingly powerful
for GLM's. It is considerably less opaque than GLIM, and has a lot
of facilities which are not available elsewhere. In particular as both
GLIM and GENSTAT use the same formalisms for GLM's as McCullagh and
Nelder's book, unlike SAS, it is fairly easy to come to a real
understanding of what is going on.
I stick by my original statement, if you do a lot of GLM's, and
I do, nothing else comes close. Note that, of course, you can fit
GLM's in all sorts of packages, including SAS. What I like is the
ease and flexibility of GENSTAT. I use SAS every day, indeed the bulk
of my data lives in SAS datasets, but I haven't fitted a GLM in SAS in
earnest for five or six years. (i *have* tried the new features as far
as ver 6.08).
PS - Make sure that you get version 5.3, it's better than 5.1 or 5.2,
and vastly better than versions 4.anything.
E-Mail - A.Staines at leeds.ac.uk Dial +44 (0)113 Tel - 2443517 Fax - 2426065
S-Mail - LRF Epidemiology Unit, 17 Springfield Mt, Leeds, LS2 9NG, UK.