Repros Therapeutics, Inc. (RPRX)

[exwannabe]

Program Survival Bias for Dummies:

JC has a contract with McD's to provide McChich nuggets. The contract calls for the company to weigh the McChichs on McD's provided scales (with some known margin of error), and only accept those over 10g. Furthermore, the same weighing will be used to provide the gross weight for billing (McD pays by the gram).

After a few months McD notices that the weight they measure upon random sampling of deliveries is tending to be less than that stamped on the box. What's up?

It's really simple. Some McChicks that truely are over 10g were not accepted, while some that were under 10g were mistakenly accepted. Both of these cause a BIAS of the final product weight.

This is all simple stats, hard to argue with it.

Clinical trials?

Exactly the same. There will definitely be a tendancy for P3s to have a HR less than than the P2s even if all the data is pristine.

One can argue to hell and back about how significant the effect is, but to deny it makes no sence.

[IMHO, seams kind of OT here as few doubt the efficiacy aspect of Proellex]