STAT 8630 — Advanced Statistical Applications II
Homework 5 – Due Thursday, April 23
Homework Guidelines:
•
Homework is due by 4:30 on the due date speciﬁed above. You may turn it in at
the beginning of class or place it in my mailbox in the Statistics Building.
No
late homeworks will be accepted without permission granted prior to
the due date.
•
Use only standard (8
.
5
×
11 inch) paper and use only one side of each sheet.
•
Homework should show enough detail so that the reader can clearly understand
the procedures of the solutions.
•
Problems should appear in the order that they were assigned.
Assignment:
Read chapters 12–13 from FLW (our text), ch.14 from Molenberghs and
Verbeke’s book (handed out in class), and ch.14 from McCulloch et al.’s book (also handed
out in class). Do exercise 12.2 (all parts) from FLW and exercises 14.3 and 14.12 from
McCulloch et al. In addition, do the following problem:
4. Recall the hospital visit data from problem 9.12 in Davis’ book. We analyzed these
data in problem 3 of homework 4. In that analysis, a fourway interaction between
age, sex, maternal smoking status and quarter (the repeated measures factor) was
found to be signiﬁcant using GEE methodology to ﬁt a marginal Poissonlike GLM.
Here, we simplify the analysis somewhat by considering only the subjects from that
dataset for which mother’s smoking status is “Yes”. Reﬁtting a marginal Poisson
like GLM to this subset of the data, it can be found that some interactions are
no longer signiﬁcant and a parsimonious model that excludes all nonsigniﬁcant
interactions and main eﬀects (except those required by the principle of model
hierarchy) is one that includes main eﬀects of quarter, sex and age (where age is
treated as continuous), and which contains an age*sex interaction (that is where
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '08
 Staff
 Statistics, Probability theory, Binomial distribution, Negative binomial distribution, Hij, withinsubject correlation, hospital visit data

Click to edit the document details