|
Model
|
Assumptions
|
Relative likelihood of model (from DIC)
|
Notes
|
|---|
|
Humans
|
Animals
|
|---|
|
Random participation
|
Non-participation
|
|---|
|
Low-risk
|
High-risk
|
Low-risk
|
High-risk
|
|---|
|
1
| ✓a
| | | | |
< 10− 100
|
Least good fit
|
|
2
| ✓ | ✓ | | | |
≈ 10− 5
|
Implausible parameterisation
|
|
3
| ✓ | | ✓ | | |
< 10− 100
|
Poor fit
|
|
4
| ✓ | | | ✓ | |
0.955
|
Good fit
|
|
5
| ✓ | ✓ | ✓ | ✓ | |
0.026
|
No improvement over Model 4 but more parameters
|
|
6
| ✓a
| | | | ✓ |
≈ 10− 9
| |
|
7
| ✓ | | | ✓ | ✓ |
1
|
Best fit
|
-
aModels 1 and 6 have homogeneous risk for all humans, which is equivalent to all humans being “low-risk” ✓ = included as a distinct host category in the model
