Description Usage Format Source References Examples
Scientists in horticulture have been interested in how plants adapt
to cold climates such as we find in Wisconsin. They have noticed that
plants which have been previously ‘acclimatized’ suffer less damage.
Palta and Weiss (1993) conducted a
controlled laboratory experiment to examine two species of
potato
es (1, 2) subjecting plants to one of two acclimatization
regime
s (1 kept in cold room; 0 kept at room temperature) for
several days. Plants were later subjected to one of two cold
temp
erature treatments (1 just below freezing, -4 C; 2 way below
freezing, -8 C). Two responses were measured, both being damage
scores for photo
synthesis and ion leak
age.
Initially there were 80 plants, but some were lost during the experiment.
Your analysis should take care to address the imbalance due to unequal
sample sizes for the different factor combinations.
1 |
Hardy data frame with 75 observations on 6 variables.
[,1] | potato | factor | potato identifier |
[,2] | regime | factor | hardiness regime |
[,3] | temp | factor | temperature |
[,4] | code | factor | plot code |
[,5] | photo | numeric | photosynthesis score |
[,6] | leak | numeric | ion leakage score |
Jiwan P Palta (mailto:palta@calshp.cals.wisc.edu) \& Laurie S Weiss, U WI Horticulture, (http://www.hort.wisc.edu)
Palta JP and Weiss LS (1993) 'Ice formation and freezing injury: an overview on the survival mechanisms and molecular aspects of injury and cold acclimation in herbaceous plants', in Advances in Plant Cold Hardiness, ed. by PH Li and L Christersson. CRC Press, Boca Raton, LA.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 | data( Hardy )
Hardy$potato <- factor(Hardy$potato)
Hardy$regime <- factor(Hardy$regime)
Hardy$temp <- factor(Hardy$temp)
# full model: you may want to consider various reduced models */
Hardy.fit <- aov(leak ~ potato * regime * temp, Hardy)
# Type I sums of squares
summary( Hardy.fit )
# Type III sums of squares
drop1( Hardy.fit, formula( Hardy.fit ), test = "F" )
# Least Squares means
Hardy.lsm <- lsmean(Hardy.fit)
# default plots for lm object
plot(Hardy.fit)
# diagnostic plot of resid vs predict with 0 line and +/- 1 SD
Hardy$code = factor( Hardy$code )
xyplot( resid(Hardy.fit) ~ jitter(predict(Hardy.fit)),
groups = Hardy$code, pch = levels( Hardy$code ), cex = 1.5,
panel = function(x,y,sd=std.dev(Hardy.fit),...){
panel.xyplot(x,y,...)
panel.abline(0,0,lty=2)
panel.abline(sd,0,lty=3)
panel.abline(-sd,0,lty=3)
} )
# interaction plot with LSD bars (but play with this!)
lsd.plot(Hardy.fit,Hardy,c("potato","regime"),
more = TRUE, split = c(1,1,2,1) )
# the following give a slightly different plot -- why?
attach(Hardy)
lsd.plot(potato,leak,regime, split = c(2,1,2,1))
detach()
# side-by-side interaction plots for the subsets of t=1,2
lsd.plot(Hardy.fit,Hardy,c("potato","regime"),
Hardy.lsm[Hardy.lsm$temp==1,],
more = TRUE, split = c(1,1,2,1) )
lsd.plot(Hardy.fit,Hardy,c("potato","regime"),
Hardy.lsm[Hardy.lsm$temp==2,],
split = c(2,1,2,1) )
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.