Refer to model (4.11) with width and color effects for the horseshoe crab data.

Using the data at www.stat.ufl.edu/∼aa/intro-cda/appendix.html:

a. Fit the model, treating color as nominal-scale but with weight instead of width as x. Interpret the parameter estimates.

b. Controlling for weight, conduct a likelihood-ratio test of the hypothesis that having a satellite is independent of color. Interpret.

c. Using models that treat color in a quantitative manner with scores

{1, 2, 3, 4}, repeat the analyses in

(a) and (b).