Exercise #2, all parts, ELMR Chapter 2, pages 52-53, Due: Thursday, Nov. 8 by end of day.
Due: please submit your completed assignment by e-mail by end of the day, Tuesday, Nov. 20
This datafile from Jeffery Simonoff's textbook, Analyzing of Categorical Data (also linked on the course web-page) describes the yearly production of lugs (boxes) of grapes for each of 52 rows of grapes in an Ohio vineyard over the period of 1983 to 1991. Note that the while the response variable Lugs contains decimal values, it is essentially discrete in that the fractional components denote half-lugs. Note also that since the rows are numbered contiguously (i.e. side by side) it is reasonable to consider a continuous (but not necessarily linear) pattern over rows. The vineyard owners were particularly interested in the potential for a fertility variations over field in which the rows were planted and whether these fertility patterns were the same from year to year.
Fit generalized models using (a) gaussian, (b) poisson and (c) negative binomial families and provide the conventional summaries of the fitted models supplemented by graphic representations of the patterns as described above. Briefly compare the interpretation of each of these models with regard to the objective described above and discuss theoretical and/or practical advantages and disadvantages of the three approaches. (Note: you may omit all the preliminary and intermediate, e.g. diagnostic, steps in your model building process in your submitted assignment)
Assignment 3 was due by the end of the day, Thursday Nov. 29th
Assignment 4 was due by the end of the day, Thursday Dec. 6th