Dear all,
I am not a statistician, but as a forestry researcher I am involved
with the analysis of basic trial designs used in forestry research.
Generally I only use appropriate designs that I know are valid and that
will allow me to interpret results. Recently, I was asked to help with
the analysis of the following trial by an associate, and because it is
not properly replicated, I am not sure what to do.
The treatments consisted of an irrigated block and an un-irrigated
block. These were not replicated. Within each of these blocks were 8
plots that were either fertilized (4 plots) or not fertilized (4 plots).
When analyzed using ANOVA:
Source of variation DF
Irrigation 1
Fertilization 1
Irrigation x Fertilization 1
Residual 12
Total 15
The irrigated treatment has no error against which it can be compared
as it was not replicated. In other words the mean of irrigated plots
versus un-irrigated plots can be observed, but no statistical
significance can be attached to this figure. Differences between
fertilized plots and this interaction with Irrigation can be estimated
and the SE calculated.
The problem arises in that my associate would still like to test for
statistical significance between the irrigated and un-irrigated plots.
Could one carry out a t-test by comparing the 8 irrigated and 8
un-irrigated plots, or are there other tests that would be more
appropriate?
Your assistance/suggestions would be very much appreciated
Regards
Keith
********************************************************
Keith M. Little (PhD)
Programme Manager: Re-establishment Research
Institute for Commercial Forestry Research
PO Box 100281
SCOTTSVILLE
SOUTH AFRICA
3209
Tel: +27 33 386-2314
Fax: +27 33 386-8905
Cell: 082 373 1854
**********************************************************
------------------------------------------------------------------------
DISCLAIMER
This e-mail message is privileged and confidential. If you are not the
intended recipient please delete the message and notify the sender.
Any views or opinions presented are solely those of the author.
------------------------------------------------------------------------
<<<<GWIASIG 0.07>>>>
|