Question - Table 10.9 summarizes a study with variables age of mother (A), length of gestation (G) in days, infant survival (I), and number of cigarettes smoked per day during the prenatal period (S). treat G and I as response variables and A and S as explanatory.
1) Explain why a loglinear model should include the λAS term.
2) Fit the models (AGIS), (AGI, AIS, AGS, GIS), (AG, AI, AS, GI, GS, IS), and (AS, G, I). Identify a subset of models nested between two of these that may fit well.
3) Use (i) forward selection and (ii) backward elimination to build a model between two of the models listed in (b). Compare the results of the strategies, and interpret the models chosen.
|
Table 10.9 Data on Gestation and Infant Survival for Exercise
|
|
Age
|
Smoking
|
Gestation
|
Infant Survival
|
|
No
|
Yes
|
|
< 30
|
< 5
|
≤ 260
|
50
|
315
|
|
> 260
|
24
|
4012
|
|
5 +
|
≤ 260
|
9
|
40
|
|
> 260
|
6
|
549
|
|
30 +
|
< 5
|
≤ 260
|
41
|
147
|
|
> 260
|
14
|
1594
|
|
5 +
|
≤ 260
|
4
|
11
|
|
> 260
|
1
|
124
|
|
Source: N. Wermuth, pp. 279-295 in Proc. 9th International Biometrics Conference, Vol. 1 (1976). Reprinted with permission from the Biometric Society
|