The most common method is with some form of economical game where in an experimental setting you either reward either aggression or cooperation, and then inject half the participants with testosterone and the other with a placebo. The finding of that is just as I mentioned above, ie that regardless if cooperation or aggression is rewarded, what correlate with testosterone is how much effort is spent on defending a position. If it is cooperation, then cooperation goes up after injecting testosterone. The real life example that a professor in neuroendocrinologist gave is that if you inject a buddhist monk with testosterone the expected results, considering their culture, would be increased zen-like behavior.
The reason testosterone get associated with aggression is that zen-like behavior is not the common cultural appropriate response to being challenged. Lets imagine the case of a man disrespecting an other man (a different word for challenging someone socially) by cutting a waiting line in front of an other man. Testosterone would in that case amplify what ever the cultural appropriate behavior is for defending their social status, which in many places of the world would be a violent response. From the outside that would look very much like aggression.
The common difference between animal studies and human is not that animals and humans has fundamental different biology or behavior in regard to impulses, but rather that humans tend to use it more novel and complex way. Loosing your job can easily be converted to a feeling of loosing social status, especially if someone or a group can be made to be the perceived as taking that social status.
In a rather fascinating study researcher looked at sport fans and found that their biology mimicked the social status effect of their teams. Fans testosterone went up if their team won, or down when it lost. A similar effect has been found in primary elections, where men on the winning side gain testosterone after the election results was announced.
If somebody's chance of reproducing is say 2%, they may start to think, "Hey, if me and a bunch of other losers invade the next town, my chance of "getting some" goes up to 5%." The math of evolution would favor such behavior.
You see it in sports also: if your team is are clearly losing, you gamble more with "tricky" plays and perhaps cheating, which may result in interceptions (losing possession of the ball) and/or penalties.
I should point out this is probably a factor with any inequality, not just marriage. Pitchforks come out when enough feel left behind.
No, it doesn’t work like that. Not getting something which you never had is not a loss in status. And only loss in status is affected by testosterone, not any possible gains in status.
I.e. only people who feel they are losing something which they feel they currently have are affected by testosterone.