Re: multicollinearity in regression


Paul wrote:
Bob,

If you know so much about the topic, then how would you approach the
problem. Can you offer more constructive advice?

Yes I can. That's why I gave you the constructive advice:

RF> You should take a first course in multiple regression analysis
that is
RF> taught by a competent statistician

I taught many such courses -- each of which had one of more
prerequisite courses in statistics. This is neither the appropriate
forum nor is there sufficient time to go through even a small part
of such a course -- though I have posted dozens in sci.stat.math
that covered some small portions of such a course.

I don't mean to be facitious when I said, more than once, that you
should not expect to be able to perform a brain surgery without
killing the patient if all you learned about brain surgery are from
some newsgroups. It takes at least as much training as it takes
to be a competent brain surgeon as it takes to be a competent
statistician -- of the DATA ANALYSIS kind.

You can be a mathematical statistician very easily -- as easily as
it takes to get a degree in mathematics; but you most likely
wouldn't know how to analyze an data using statistical and data
analytic methods.

That's the sad REALITY of the Real World. IMO.

-- Bob.

.