Economic theory mostly assumes that individuals are self-interested, yet there is plenty of evidence that they also value fairness and incorporate the utility of others, including people they are not related to. So they may not be that self-interested or, put it in another way, their utility function has other arguments that pertain to other individuals or to aggregate measures. Sociologists and psychologists would argue here that at least part of this interest in others comes from education or societal pressure. Many studies have highlighted that some societies are more caring, and others more greedy. Much of this is based on surveys.
Economists do generally not like surveys because they do not reflect actual decisions. But controlled experiments or good data are very hard to come by. Bruno Frey, David Savage and Benno Torgler found an interesting data set: the passengers of the Titanic. There is a well established societal norm that women and children have priority in situation of life and death where a future reciprocity is not expected. It this norm really applied?
Take as an example the contrast between British and American customs: In Britain, people queue for everything and apply strictly the norm of :first come first served". In the US, the price is much more used as a selection mechanism, thus wealth matters more. In the context of the Titanic, this would mean: among Brits, a larger proportion of women and children than men should have survived; among Americans, the survivors should be more frequent in first class than third class. According to this analysis by John Henderson, this hypothesis would be correct: the survival rate in third class for British women is 47%, British children 35% and British men 12%. Americans in first class survived at a rate of 67%, 47% in second class and 28% in third class. Interestingly, 58% of Americans survived against of 33% Brits. Did Americans push their way to the lifeboats?
But the analysis is not that simple: there can be composition effects: American may have been proportionally more numerous is higher classes or with more females. Few staff survived, and they may have been disproportionately British. Thus, simple averages are not enough, some proper regression analysis is required using a variety of controls. This is what Frey, Savage and Torgler do and they confirm that women and children had higher survival rates, thus societal norms seem to have largely prevailed. Passengers from higher classes also had a higher survival probability, but this result is likely tainted by the fact that 1) they were better informed as early on few believed the Titanic would sink, 2) most boats were on the first class deck.
And even with all those controls, the British are more likely to die and the Americans are those most likely to survive. But this distinction is not visible among women. Ah, those selfish American men.