Re: ANOVA QUESTION - Terminology


jp wrote:
I have looked at an intro to biostatistics book.

I know two groups is t-test and more than two is ANOVA. I'm sorry this
sounds so basic, but I didn't find any example of this in the book.

Then perhaps you should get one written by a statistician, and not
someone calls himself/herself biostatistician. I am just INFERRING
from your brief discription that the author(s) are MISSING the most
basic (and simple) ingredient in their introdution of the topic of
ANOVA.


Most examples of ANOVA are 3 groups of race (Chinese, White, Black)
with one dependent variable (IQ). I can see how this is a one-way as
there is one independent and one dependent.

Actually the used of the terms "dependent" and "independent" variables
are NOT necessary. That is often introduced as a One-Way
classification
of ONE variable, IQ. You're merely comparing the three means based
on the classification or RACE (which can be nominal, ordinal, or
real-valued). Now some biostatisticians even get THOSE concepts
of DATA all mixed up, as in a recent example in sci.stat.math where
Richard Ulrich was making the same mistake another phoney
statistician from the MedStat group was making.

The introduction of the term DEPENDENT and INDEPENDENT
variables comes into play only as a ALTERNATIVE way of formulating
the SAME problem, using Multiple Regression analysis technique
for analyzing many different groups by the introduction of INDICATOR
variables (sometimes called DUMMY variables) for the categories
in the classification.

It's a different FORMULATION of the problem that delivers the
IDENTICAL statistical tests and results!

But that is usually done in LINEAR MODELS after the introduction
and study of Multiple Regression Techiniques.

That is one reason Richard Ulrich, for all his claim to be a
biostatistician,
NEVER got that far, because after teaching others his Quackery in
these groups for nine years, he still didn't have the slightest idea
what
LINEAR MODELS are -- this FACT has been throughly documented
in the archives of sci.stat.math.

That's what I mean when I say you need to learn something property
from creditable statisticians!

There is just the right book for YOU -- the one I've taught from for
decades, and recommended to others then, and now.

That's the book titled Applied Linear Models, by Neter, Wasserman,
Kutner, and Nachtsheim.

Is Richard Ulrich had the EDUCATION of what is covered that book,
he would NOT have made the DOZENS of blunders he had made
and left the trail in the sci.stat.math archives since 2005!

That book is suitable for an UNDERGRADUATE class in a second
(and third) course in Applied Statistics. It has excellent coverage
on the analysis of ALL kinds of one-way, two-way, multi-way,
as well as designs of experiments, Latin Squares, ete., etc., ALL
fit nicely as SPECIAL cases of the Linear Model!

Instead of answering your questions below point by point, I think
referring you to that book which has ALL the answers (and 10 times
more the answers to questions you don't even know how to ASK
yet).

They are ALL there.

BTW, I've taught the entirety of that book in TWO semesters as
an undergrad sequence, after a first course in statistics.

That book can be used as a beginning graduate course textbook,
in many topics, such as regresssion, analysis of variance and
covriance, experimental design, and in the latest editions, a
bit about multivariate analysis.

I've used that as a supplement textbook in my graduate course in
Data Analysis, that presumes almost all of the regression material
as the pre-requisite.

That is BY FAR, the most appropriate, and the most efficient way,
not only for you, but for EVERYONE, to learn the substance of the
topics within it.

I would put it on record here, that I STRONGLY recommend the
book specifically to Richard Ulrich and Anon Bob O'Hara, because
they are both frequent pariticpants in these groups, and they are
ABSOLUTELY and POSITIVELY deficient in the topics of
Regression, Linear Models, that are pre-requite to the ANOVA
models. I noted in particular that RIchard Ulrich even gave
you some bad advice out of his own deficiency on the topics
he never learned.

-- Reef Fish Bob.


When repeated measures is
described in the book, the ANOVA now changes to something like
monitoring heart rate at several times during biking. A two-way is
described as having two independent variables, say gender and race in
which IQ is again, the dependent variable.

The question I have is what would the analysis be for for what I was
suggesting above. To me you could treat it as a two-way in that Group
(A,B,C) and Test# (Pre Post) could serve as two independent variables,
but to me I thought it sounded like a repeated measures because the
Test Score was the dependent variable measured at two time points.
There is no example like this in my book, and perhaps I not understand
it. I would imagine this newsgroup is for individuals of all learning
levels?

To me, this sounds like a two-way design, but I am confused because it
is more like a one-way repeated measures since the dependent variable
is measured twice and there is the independent variable Group. I would
really like to know if this distinction.


Reef Fish wrote:
jp wrote:
If you perform a t-test on the change (difference from pre and post)
don't you have to look at the initial scores to see if they were
different?

I was really more concerned with the terminology.

That's why I suggested you get an elementary text book and read up
on the comparison of means.

Two groups; generally a T-test. More than two, the comparison of
MEANS is called, strangely, the ANOVA (analysis of VARIANCE).
Then you go on to other forms of ANOVA and designs.

The newsgroup is not where you can, or want to, learn the basics from
scratch. JMHO.

-- Reef Fish Bob.



Say I wanted to
compare the same dependent variable (pre and post scores), and I have
one independent variable (Group- A, B, C). I can see if I use the
differences between the pre and post this would be a simple one-way
with repeated measures. Assuming I don't use the differences what
would this be called, a two-way? That is, treating Group and Score as
independent variables?

Thanks for the responses!

Reef Fish wrote:
Richard Ulrich wrote:
On 9 Sep 2006 12:53:19 -0700, "jp" <jpopovich@xxxxxxxxx> wrote:

I have the following set up:

One group (males, females)
One dependent variable (test scores), but tested pre and post tutoring.

Would this be a repeated measures design?

It can be analyzed that way, with the two periods.

The two periods are pre-tutor score and post-tutor scores and
it has only ONE variable, of the "difference" (improvement).


What would this be called (i.e., One-way ANOVA with repeated measures,
Two-way, etc.)?

For Pre-Post, the testing usually assumes that the group means do
not differ at Pre;

Why? Where did you get that?

and the efficient test is usually the one-way
ANCOVA, using the Pre as covariate.

That's just a simple T-test for two independent groups on the
DIFFERENCE of the pre-and-post scores. Freshman/sophomore
stuff, Richard.

jp, get you an elementary testbook on testing means.

That may save you much time in unlearning some bad advice you
get from these newsgroups.

-- Reef Fish Bob.



For the repeated measures, it is "one-way" ANOVA with repeated
measures.

In general, a repeated-measures analysis with several
periods might have a few 'missings' and you could do
the closest thing to the same analysis by doing a two-way
ANOVA with Period and Subject (or ID) as the factors.


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

.

Relevant Pages

  • Re: ANOVA QUESTION - Terminology
    ... I know two groups is t-test and more than two is ANOVA. ... Most examples of ANOVA are 3 groups of race ... and Test# could serve as two independent variables, ... a repeated-measures analysis with several ...
    (sci.stat.consult)
  • Re: ANOVA QUESTION - Terminology
    ... MEANS is called, strangely, the ANOVA. ... -- Reef Fish Bob. ... a repeated-measures analysis with several ...
    (sci.stat.consult)
  • Re: ANOVA QUESTION - Terminology
    ... Most examples of ANOVA are 3 groups of race ... and Test# could serve as two independent variables, ... measures design, probably because the procedure they use in their stats package is found under "GLM->Repeated Measures", or something similar. ... The interaction F-test is equivalent to an independent groups t-test on the change scores. ...
    (sci.stat.consult)
  • Re: Is a 2-way ANOVA appropriate?
    ... independent variables. ... I was planning to use a 2-way ANOVA for my statistical analysis and I was wondering if it is okay to have an independent variable that is discontinuous (Soil texture), and an independent variable that is continuous and still use a 2-way ANOVA. ... If your ANOVA software allows for covariates and for interactions between factors and covariates, ...
    (sci.stat.math)
  • Re: Repeated measures ANOVA with one measurement
    ... suggestion that I do a repeated-measures ANOVA to remove some of the ... ANCOVA the power of post-hoc tests are dramatically reduced. ... Are the power of post-hoc tests reduced in RM ANOVA like an ANCOVA? ... Newman-Keuls, Dunnett's test, etc) are not available for repeated measures factors in SPSS, because those procedures were designed for between-Ss ANOVA. ...
    (sci.stat.math)