One problem in replying to your email is that I don't know how
much you already know about multivariate statistics. If you know a
lot then my reply will sound patronising!
Are you committed to using a particular multivariate package? If
you are, then you are sure to find courses specifically geared to
the multivariate workings of that package.
However there is a problem with most courses in multivariate
analysis. They disassemble the algebra of a particular method
(such a principal components analysis) but offer little in the way
of diversity of worked examples and explanation of results. From
the consumer's point of view, it's a bit like wanting to fly from
London to Paris, but before taking off the airline insists you
understand the physics and mechanics of a jet engine.
If you are not committed to any particular software, then you
could try Oyvind Hammer's superb package PAST. It is free of
charge, but has a somewhat unconventional user interface. You are,
nevertheless, able to paste in copies of multivariate tables
already assembled in Excel. Go to:
PAST is certainly suitable for somebody with an archaeological or
biological bent. The author has a palaeontological background. It
will do any type of multivariate analyses that people in those
fields are ever likely to need - the most common needs being for
clustering, seriation, principal components, correspondence
analysis, and linear discriminant analysis.
PAST is, however, not a course. So you could couple use of PAST
with such lucid texts as Bryan Manly's "Multivariate Statistical
Methods: a primer". I have the 2nd edition of 1994, but I see now
that there is a 3rd edition. There is lots of advice on what
substantive problems require which multivariate method, and lots
of raw data with worked examples.
In brief, what I am suggesting is a way to go. But it is clearly
not a course as such.
I will be interested to learn whether members of this list have
had constructive experiences with online multivariate courses.
Richard
On 29/06/2011 18:37, fiona beglane wrote:
[log in to unmask]"
type="cite">
Hello
Does anyone know of an online course or a short course (e.g. a week) in
multivariate statistics, preferably with an archaeological or biological bent?
Thanks
Fiona