Re: Effect of cattle stocking rate on soil P (fwd)

---------- Forwarded message ----------
Date: Thu, 27 Feb 1997 13:01:42 -0600
From: Jenny Kendall <zenomind@earthlink.net>
To: sustag@beta.tricity.wsu.edu
Subject: Re: Effect of cattle stocking rate on soil P (fwd)

Tom Hodges (moderated newsgroup) wrote:

>   2. (My Burning Question!): Given the variability of these fields due to
> (1) natural soil variability, and (2)     variability of manure deposition
> by grazing animals, what sampling strategy should we follow?
>   I appreciate if any body can share his/her ideas and experiences with me.
> Please excuse cross posting.
>   Thank You
>  M. Mozaffari

One way to handle experimental effects such as 'natural variability' is
to consider using a 'mixed model' approach to analyzing the data.  Treat
these effects as random components in your model, rather than fixed.  To
analyze this type of data, you would use something like PROC MIXED in
SAS.  Naturally, you would want to talk w/your local statistician about
the design..  What you gain by using this approach is that you treat
these 'design aspects' of your overall design as they actually are.  

A good reference book/short course on these methodologies is 'The
Analysis of Messy Data" by Milliken and Johnson.  I'm sorry but I have
an old edition, and I believe there is a newer edition out.  

Happy sampling,
Jenny Kendall