Well, I happen to think that this is true, but I recently read it in a different context, a medical one:
In a somewhat unexpected finding, societal male dominance over women -- patriarchy -- may help explain why men have a lower life expectancy than women worldwide.
British researchers analyzed rates of female murders and male death rates from all causes in 51 countries in Europe, Asia, Australasia, and North and South America. The prevalence of violence against women was used to indicate the extent of patriarchal control in each of the countries. Socioeconomic factors were also taken into consideration.
The study found that women lived longer than men in all 51 countries. The study also found that those countries with higher rates of female murders (indicating higher levels of patriarchy) also had higher rates for male death and shorter male life expectancies, compared to countries with lower female murder rates, the researchers said.
In fact, statistical analysis showed that variations between countries in rates of violence against women accounted for close to half (49 percent) of the variation in male death rates, the researchers noted.
Lots of examples here about the difficulties of studying something that doesn't have an easy measurable equivalent (patriarchy) and of the use of data outside laboratory conditions. Most social science studies use such data, of course, and it is almost always possible to argue that a study may not have taken into account all possible explanatory causes or controlled for them adequately. For example, the study I'm discussing here seems to have taken into account poverty rates and such, but did they also check to see if the male mortality rates correlated with male murder rates? And what about female mortality rates in general? Did they show the same pattern as the male rates?
Perhaps they did all these things. Which means that I should dig up the original study and look at it. Sigh. Maybe I will if I feel especially good.