Logistic Regression: A Self-learning Text, Third Edition (Statistics in the Health Sciences)

SMK¼ 1

E notE

D 112

notD 022

4

SMK¼ 0

E notE

D 213

notD 213

6

25. What is the estimated MOR from these data?


26. What type of matched analysis is being used here,
    pooled or unpooled?


27. Which type of analysis should be preferred for these
    matched data (where smoking status is the only
    matched variable), pooled or unpooled?

The data below switches the nonsmoker control of stratum
2 with the nonsmoker control of stratum 4 from the data set
provided for Exercises 19–22. Let W¼no. of concordant
(E¼1,E¼1) pairs,X¼no.ofdiscordant(E¼1,E¼0)
pairs,Y¼no.ofdiscordant(E¼0,E¼1) pairs, andZ¼no.
of concordant (E¼0,E¼0) pairs for the “switched” data.

ID STRATUM CASE EXP SMK
11 110
21 010
32 100
42 000
53 111
63 001
74 110
84 010
95 101
10 5 0 0 1

28. What are the values forW, X, Y,andZ?


29. What are the values ofMOR (unpooled) andd MORd
    (pooled)?

Based on the above data and your answers to the above
Exercises:

30. Which of the following helps explain why the pooled
    MOR should be preferred to the unpooledd MOR? (Cir-d
    cle the best answer)
    a. The pooledMORs are equal, whereas the unpooledd
    MORs are different.d
    b. The unpooledMORs assume that exchangeabled
    matched pairs are not unique.
    c. The pooled MORs assume that exchangeabled
    matched pairs are unique.
    d. None of the choices a, b, and c above are correct.
    e. All of the choices a, b, and c above are correct.

Test 423