On 24 May 2003 Gayathri <[EMAIL PROTECTED]> wrote (edited):
> I am working on a paper for school.
> I have data with 2 factors and 1 block.
> Factor A (variety of bread) is random with 7 levels,
> Factor B (temp) is fixed with 4 levels,
> and there are 5 levels of block (days).
> Could someone pls help me with the model for this?
> How do I write the model for this?
>
> Gayathri
What kind of language are you supposed to use? If this were an
assignment in a course of mine, I'd expect an answer something like
R(VxTxB) with T fixed, V and B random, r=1, v=7, t=4, b=5
where V = variety of bread, T = temperature, B = blocks, and
R = "the ubiquitous nested factor" as someone has called it.
I would also expect a complete layout of the ANOVA summary table,
including the <expected mean squares> and the <error mean squares>
(aka the <denominator mean squares>).
But your instructor may want something like
y = \mu + \alpha_i + \beta_j + \gamma_k + (\alpha\beta)_ij + ...
... + a collection of random error terms,
equivalent to R(VxTxB) etc. but written in different language.
(Identifying the error terms, and the effects to which they apply,
may be the point of the assignment.)
Or you may be expected to describe how you would specify the model
to a statistical package like SAS or SPSS or MINITAB. Again, this
would entail equivalent information, but in different language.
In any case, to write the model at all you need to specify the
relationships between the factors you have so far identified:
Is A crossed with B, or is one of them nested in the other?
Are <blocks> crossed with both A and B; or with one of them,
and then is it nested within the other, or vice versa?
Is the factor <blocks> fixed or random?
All of this information will help you identify the proper
error terms for the several sources of variation.
I refrain here from commenting on the problems likely to be met in
actually trying to carry out an experiment of this kind, and the
uncontrolled sources of variation almost certainly entailed in
doing so: mostly because this looks to me like a more or less
standard homework assignment, of the kind where the designer of the
question devotes no attention whatever to the practicalities of the
experiment.
-----------------------------------------------------------------------
Donald F. Burrill [EMAIL PROTECTED]
56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
