Re: Normalizing a non-normal distribution

[Thom Baguley]

Brian MacDonald wrote:
 
 I am doing a series of analyses using discriminant analysis to predict group
 membership.  Several of the variables I am using show distributions that are
 not normal.  My question is can these (and for that matter shold they) be
 somehow transformed so that the resulting distribution looks and presumably
 acts in the analyses) like a normal distribution.

It depends. For some distributions it is easy to do the
transformations (e.g., log is often appropriate for +ve skew). An
alternative approach might be to consider logistic regression which
has several advantages over discriminant analysis and doesn't
require normality.

Thom


=
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
  http://jse.stat.ncsu.edu/
=

Re: Double mediation

[Duncan Smith]

Sylvia J. Hysong, Ph.D. [EMAIL PROTECTED] wrote in message
[EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
 Hello,

 I'm hoping someone can help me with this.  I have looked at a
 multitude of resources including the David Kenny page, this and other
 newsgroups, Pedhazur (1982), Cohen  Cohen (1983), and Darlington
 (1990?), to no avail.  I am hoping someone can direct me to the right
 resource.  I am trying to conduct a test of double mediation.  In
 other words, I am trying to test the hypothesis that x--z1--z2--y.
 Is there a way to do this (and if so, what is it?), or must I result
 to a path analysis or a structural equation model?

 Thanks in advance for any help.

If I understand the question correctly, this implies a number of conditional
independence relationships which can be tested.  i.e. x cond. ind. of z2 and
y given z1; x and z1 cond. ind. of y given z2; x cond. ind. of y given z1
and z2.  If these, and only these, independence relationships hold then you
have either x--z1--z2--y or x--z1--z2--y.  To decide which, you need
some background knowledge or to conduct an experiment.  You might want to
check out links at http://www.cs.berkeley.edu/~murphyk/Bayes/bnsoft.html




=
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
  http://jse.stat.ncsu.edu/
=