Re: help with ANOVA design

klange wrote:
Hi all,

I'm currently constructing some ANOVA tables for complex experimental
designs, and would like to see if I'm on the right track.

6 plots are chosen (PLOT = random factor, 6 levels)
Each plot is split into 2 sub-plots with a burning treatment (natural
or experimental) applied to sub-plots. Each plot has one natural burn
sub-plot and one experimental burn sub-plot. (BURN = fixed factor, 2
levels)
Within each sub-plot 2 transect measurements are taken (TRANSECT =
random factor, nested within sub-plot)

The outcomes are various environmental measures that give us one value
per transect, and the transects are visited in two sequential years
(YEAR = fixed factor, 2 levels).

The way I have it at the moment is with YEAR, PLOT and BURN as three
fully crossed factors (ie, all three main effects, three two-way
interactions and the three-way interaction), TRANSECT is nested within
LOC*BURN, and then the error term is TRANSECT*YEAR.

How does that look? I wasn't sure about being able to separate the two
transect terms...?

There's an extension to this design that I'll have to put together
next, but first I wanted to get this one up and running.

Thanks all.

Kylie.

Fixed: Burn(B,2), Year(Y,2)
Random: Plot(P,6), Transect(T|BP,2)

Sources of Variance, and their df:

Y(1)
B(1) BY(1)
P(5) PY(5)
BP(5) BPY(5)
T|BP(12) TY|BP(12)

Components of Expected Mean Squares, and the error terms they imply:

B P BP T|BP Y BY PY BPY TY|BP e Error Term

B x . x x . . . . . x BP
P . x . x . . . . . x T|BP
BP . . x x . . . . . x T|BP
T|BP . . . x . . . . . x -
Y . . . . x . x . x x PY
BY . . . . . x . x x x BPY
PY . . . . . . x . x x TY|BP
BPY . . . . . . . x x x TY|BP
TY|BP . . . . . . . . x x -

Row labels are sources of variance.
Column labels are theoretical components of variance.
An "x" indicates that the column component contributes to the row EMS.
(I have omited the multipliers, to keep the table simple.)

.

Relevant Pages

  • help with ANOVA design
    ... Each plot is split into 2 sub-plots with a burning treatment (natural ... sub-plot and one experimental burn sub-plot. ... Within each sub-plot 2 transect measurements are taken (TRANSECT = ...
    (sci.stat.consult)
  • Re: help with ANOVA design
    ... Within each transect there are 11 samples taken. ... sub-plot and one experimental burn sub-plot. ... and then the error term is TRANSECT*YEAR. ...
    (sci.stat.consult)
  • Re: What is a sub-plot?
    ... In that case I'm afraid I read it for the sub-plot (Sorry Charlie, it's just the way I'm put together. ... To me a sub plot is something that, if it were to be removed, the basic shape of the books would still be the same. ... If the romance was absent from Singularity Sky would Rachel have ever rescued Martin from the New Rebublic's flagship? ... Imagine a character with a profound hatred of something. ...
    (rec.arts.sf.composition)
  • Re: flight cancellations today
    ... Paul Heslop wrote: ... out in a big plot involving planes going to the USA but they're concerned that there may be a sub-plot in case the big one failed. ...
    (uk.local.geordie)