Agriculture Reference
In-Depth Information
--------------------------------------------------------
Variety:1-Nir |
vana, |
2-Pegasus, |
3-PS 7092, | Sowing Date: 1-10/5/01, 2-10/15/01,
4-Sweet | 3-10/29/01
Vidalia | 5 Oct. 2001 15 Oct. 2001 29 Oct. 2001
--------------+-----------------------------------------
Nirvana | 59.8965 53.801 27.74
Pegasus | 71.7955 61.466 44.16
PS 7092 | 72.343 58.546 19.385
Sweet Vidalia | 78.6575 71.175 9.785
--------------------------------------------------------
In the first table, the results appear similar over the three sowing
dates. The differences appear to be largely the magnitude of the yield
with the third sowing date, 29 Oct. 2001, having much lower yields
overall compared to the 5 or 15 Oct. 2001 sowing date. An examina-
tion of the second table of variety by sowing date indicates a difference
in variety ranking for each of the separate sowing dates. A further
examination of the separate analyses of variance (data not shown) for
each of these sowing dates indicated that only on the 29 Oct. 2001
sowing date was there significant differences in the variety means.
These results also can be seen with the following commands:
margins fertility # sowingdate, marginsplot and margins
variety # sowingdate, marginsplot . The order of the fertility and
sowingdate or variety and sowingdate will affect the marginsplot
graph. FigureĀ 6.7 shows the margins variety # sowingdate followed
by the marginsplot command in the first graph. The second graph
had margins sowingdate # variety entered followed by the mar-
ginsplot command.
Split-Split Plot Design
The split-split plot design is an example of a three-factor experiment
where the layout of the experiment is such that the factors occur as a
main-plot effect with a subplot effect and finally a sub-subplot effect
(Table 6.4). Generally the precision with which the factors can be
analyzed increases from the main-plot effect to the sub-subplot effect.
For this reason, if possible, the factor of most importance should be
assigned to the sub-subplot, which has the greatest precision. This can
Search WWH ::




Custom Search