Re: Rolling up variance

On 8 Dec 2006 08:14:38 -0800, "Reef Fish"
<large_nassua_grouper@xxxxxxxxx> wrote:


Richard Ulrich wrote:
[snip]
2) Bob has an idiosyncratic meaning for a word, and
defends it, and attacks the mainstream usage. Multi-
collinear. Multivariate.

This is just typical of Richard's NOISE and MISREPRESENTATION.

There is no need for me explain other than his deliberately
misleading use of WORDS. It's all in the archives.

Search for:

"multivariate regression" -- NOT "multivariate". Ulrich does not
know the definition of a "multivariate REGRESSION"
and makes the error of equating it with "multiple
regression"


Reef Fish Bob has a problem of badly misreading many of
my posts. That is how he started making the above ERROR
in attribution to me. For no apparent reason, in several threads
in October, he interjected the word 'regression' into his version
of my discussion ... which I finally described as his 'hallucination,'
as mentioned below.

Here are the lines where I corrected his ERROR most completely.
http://groups.google.com/group/sci.stat.math/msg/a1869b413a306a04

====== Included in my post of Oct 16.

We agree, and Bob has no evidence *from me* (I cannot
be responsible for his hallucinations) to think otherwise -
"Multivariate regression" in a textbook would denote multiple
dependent variables, since "multiple regression" is the
accepted name for the ordinary procedure.


Bob does not mention or dispute -
The same phrase in a sci.stat.math question will be a
mis-reference to multiple regression, almost all the time.


Bob does not mention or dispute -
Most people use "multivariate" in other places to refer to
multiple variables.
======= end of cite.
Bob's photographic memory....




"multicollinearity" or "multicollinearity condition" -- NOT
"multicollinear".

There are DOZENS of threads on those two keywords above, and
you'll find Richard Ulrich's ERRORS everywhere.

If anyone could stand to read them, they will find that
I do agree with Bob that multicollinearity can be dealt
with fairly summarily for "prediction" equations.

But the problems are much more acute when the intention is
inference. Inference is a place where Bob's nihilism cuts in.
His participation, then, falls under the errors of "Topics where
Bob knows little about the question and refuses to politely
step aside".


--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.

Relevant Pages

  • Re: Univariate Analysis
    ... So this is just the 'regular' kind of statistics that you may be taught ... 'multivariate' analysis is not the topic of interest. ... confused by the use of "multiple regression" and "multivariate ... "Multivariate" occasionally means "multiple dependent ...
    (sci.stat.math)
  • Re: Univariate Analysis
    ... So this is just the 'regular' kind of statistics that you may be taught ... 'multivariate' analysis is not the topic of interest. ... confused by the use of "multiple regression" and "multivariate ... "Multivariate" occasionally means "multiple dependent ...
    (sci.stat.math)
  • Re: Univariate Analysis
    ... Why don't you come up with a reference from a statistical MULTIVARIATE ... regression with TWO OR MORE Dependent variables. ... A multiple regression, is by definition and by exclusion, a regression ...
    (sci.stat.math)
  • Re: Richard Ulrich used THREE posts to distort one post in one thread
    ... That was when Richard Ulrich took it as his opening to push his ... Quackery in regression analysis. ... Those are patterns with Bob that I've noted several times. ... Linear regression might be the least ...
    (sci.stat.math)
  • Multivariate regressions with paneling, constraints, and analysis
    ... Anybody got a good reference on multivariate ... regression with an eye towards the various ways different explanatory ... and response variables can be connected, ... regressions w/ panelling from the GMM point of view and its reduction ...
    (sci.stat.math)
Quantcast