It's somewhat surprising, but then, men can still be significantly more prone than women to consider themselves expendable, and still outsurvive women in shipwrecks if both genders tend to be non-self-sacrificing enough for the situations to devolve to "every man for himself." For purely physical reasons, men are more likely to make it out of a panicked crowd alive. I'm a bit surprised that the Titanic scenario was as exceptional as it was, but I would not necessarily have predicted that relative rates of self sacrifice would dominate survival rates.

If a reliable study were to find that women are as or more likely to risk or sacrifice their lives to save non-progeny compared to men, it would certainly be sufficient to change my mind.

I'd say the shipwreck data reinforces it: in the circumstances where heroism is least observable and where death is most likely (reducing the potential reward and increasing the incurred risk), we see less peacocking. If the relationship ran the inverse direction - the more the reward and the less the risk, the less risk-taking - that'd be pretty strange and hard to reconcile with the Baumeister paradigm.