Rich Iverson
Brett Kramer
Tina Liu
Lab 3
1. There is a strong linear relationship between all five physical characteristics. In
particular, we note the following:
 In general, the data exhibit a strong linear relationship (correlation) between all
variables.
 Whenever Rear Width is involved, we see a split in the plot at higher values of
both variables. We speculate that this separation at higher levels (as the crabs
mature) is due to species and/or sex differences between crabs.
 In regard to rear width, variability increases with levels. We would want to
explore this variability for indications of heteroscedasticity.
 There is an unusually strong linear relationship between carapace width and
carapace length.
2. As carapace length increases, the rate of change in females’ rear width is greater,
relative to that of males.
3. We noticed the carapace length and rear width linear relationship diverge for the
two sexes. We notice that for the same value of carapce length, the rear width for
the females is greater than that of males.
The color separation by sex in the scatter plot makes this relationship more
discernable to the eye, especially when compared to the coplot. By choice, we
would use colors to enhance our graphics.
4. CW, FL, and BD. Using colors as a graphical separator between the two species
in a scatter plot matrix, we can clearly see these characteristics differ across
species. For example, comparing CW and FL, their slopes are similar, the values
of one characteristic are higher at each of the corresponding characteristics.
5. The shape of the overall distribution is fairly elliptical, perhaps more variation at
the upper levels. The species distribution demonstrates a shift in relationship to
each other; however, in regard to sex, the shape of the distributions appear to be
different. They have different slopes.
However, for lower values of the variables, these four groups may prove difficult
to differentiate.