Re: ANOVA with empty cells

[Mike Miller]

On Mon, 11 Apr 2005, Isaac Neuhaus wrote:

> Is there a function to calculate ANOVA for a design with empty cells? 
> Please see the data below (f1 and f2 are the independent variables and 
> response is the dependent variable).  The model that I would like to fit 
> is:
>
> response ~ f1 + f2 + f1*f2

If you mean to fit a model with 4 levels of f1 (3 dummy coded predictors) 
and 2 levels of f2 (one dummy coded predictor), and the interaction of f1 
with f2 (another 3 products of the dummy coded variables), you must either 
have data in all 8 cells or you must have some restrictions on the 
parameter space.  Otherwise, Octave can't help you, but neither can any 
other program.

Your data....

>     f1   f2   response
> S1    1    1    7
> S2    2    1    1
> S3    3    1    4
> S4    4    2    5
> S5    3    2    5
> S6    1    2    9
> S7    3    2    5
> S8    1    2    10

....will not allow you to fit the model with interaction, but you can fit 
the "response ~ f1 + f2" model.

Mike

--
Michael B. Miller, Ph.D.
Assistant Professor
Division of Epidemiology and Community Health
and Institute of Human Genetics
University of Minnesota
http://taxa.epi.umn.edu/~mbmiller/



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------