Quote:

---

you can model in both directions

---

There is an order of magnitude (or two orders of magnutude) more data from external evidence for SOC mortality. This may argue towards using more data, i.e. option 1.