Friday, May 11, 2012

Women and children first? No

The infamous quadruplicate papers of Frey, Savage and Torgler have caused a lot of grief for their multiplicity, yet they yielded a somewhat interesting, yet old result: women and children get priority on maritime disasters, crew are last, and there are some subtle differences among passengers from different nationalities. This result, however, was obtained using a sample of two: the Titanic and the Lusitania. And one also argue that there was some selection bias for the Titanic, as this was a much celebrated inaugural voyage with, let's say, an unusual set of passengers. It could therefore not hurt to increase the sample size.

Mikael Elinder and Oscar Erixson jack up the sample from 2 to 18. And the results are completely reversed. Women are at a distinct disadvantage, crew fare much better than the rest. This is the outcome you would expect from a free-for-all situation where weaker women get pushed aside on the run for the lifeboats. The lesson from this: do not trust a sample of two, even if amplified by four publications.


LemmusLemmus said...

And on the Titanic itself, see Andreas Diekmann's paper (in German, with an English abstract):

Carlos said...

"...even if amplified by four publications."

hahaha that was bad

Kansan said...

So Frey said the study was worth four publications because it was so important. And now it turns out to be a fluke. Oh the irony.