This might be a matter of language.  Planned contrasts are the same thing
as planned comparisons, if that's any help.

On 9 Oct 2002, wuzzy wrote:

> Has anyone come across an intro on planned contrasts.
> I suppose a background in algebra is required

Fluency, at least, since it is cumbersome to try to describe (let alone
specify) contrasts (or comparisons) in any other language.

> to teach vector-space contrasts.

Not sure what you intend by the qualifier:  as distinguished from what
other kind(s) of spaces, or of contrasts

> I've also seen multiple regression taught in terms of vector-space and
> matrix algebra.  Are contrasts considered important to teach for
> experimental design, any recommended texts?  eg., the argument could
> be made that omnibus tests are sufficient.

Except that they aren't, in the presence of formal factors having more
than two levels or categories.  The omnibus F test only permits one to
conclude (if H0 is rejected) that among these 3 (or 5, or 18) categories,
not all of them have the same mean (or expected value).  Which immediately
begs the question, "OK, which categories are different from which other
ones?" -- and this question can only be answered by considering contrasts
(or, if you prefer, comparisons).

(When there are only two levels, there is only one difference between
levels, and if the difference is found to be non-zero one needs only
observe the sign of the difference to determine which level (or treatment)
has the superior results.  But with more levels, the number of imaginable
contrasts increases startlingly:  even with only 3 levels, there are
already 6 obvious possible contrasts:  three pairwise comparisons (A vs B,
A vs C, B vs C) and three comparisons of the average of two levels vs the
remaining level.

> Usually I don't see the term "planned contrast" in non-statistical
> literature, maybe I just haven't been aware of it.

'Twould nearly always follow an ANOVA (with its omnibus tests).

 -----------------------------------------------------------------------
 Donald F. Burrill                                            [EMAIL PROTECTED]
 56 Sebbins Pond Drive, Bedford, NH 03110                 (603) 626-0816
 [Old address:  184 Nashua Road, Bedford, NH 03110       (603) 471-7128]

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

Reply via email to