Question - Table 7.1 Beetles Killed After Exposure to Carbon Disulfide
|
Log Dose
|
Number of Beetles
|
Number Killed
|
Fitted Values
|
|
Comp. Log-Log
|
Probit
|
Logit
|
|
1.6907
|
59
|
6
|
5.6
|
3.4
|
3.5
|
|
1.7242
|
60
|
13
|
11.3
|
10.7
|
9.8
|
|
1.7552
|
62
|
18
|
21.0
|
23.5
|
22.5
|
|
1.7842
|
56
|
28
|
30.4
|
33.8
|
33.9
|
|
1.8113
|
63
|
52
|
47.8
|
49.6
|
50.1
|
|
1.8369
|
59
|
53
|
54.1
|
53.3
|
53.3
|
|
1.8610
|
62
|
61
|
61.1
|
59.7
|
59.2
|
|
1.8839
|
60
|
60
|
59.9
|
59.2
|
58.7
|
For the flour beetle mortality data in table 7.1, fit models using the (a) logit, (b) probit, (c) complementary log-log, and (d) log-log links, with dosage entered in the model in standardized form. Report and interpret model parameter estimates. What aspect of the data pattern causes the model with log-log link to fit so poorly?