Testosterone and the Ultimatum Game

The common wisdom on testosterone is that it contributes to risky and aggressive behavior, but new research reveals a different pattern. In a study, 121 women were dosed with testosterone or a placebo and then played the ultimatum bargaining game (see Chapter 3 of SuperFreakonomics for more than you ever wanted to know about Ultimatum). The results were counterintuitive: women given testosterone actually made higher initial offers in the bargaining game, resulting in less conflict and more efficient social interactions. The researchers attributed the results to “a desire of the testosterone group to maintain their images — by avoiding rejection — aligning with the so-called social status hypothesis.” [%comments]

Eli Baker

So...If it comes out as suspected it confirms the old theory...if it doesn't we have a reasonable explanation.

Therefore, the theory holds!!!!

Unbelievable that this stuff should even get into print.


I wish all those MBA grads who were taught the Ultimatum Game in class and went on to Wall St or leadership positions in the Republican Party would APPLY IT when thinking about executive compensation and tax policy. If they don't start making a reasonable offer on both and rein in wealth inequality, we're going to end up in some ugly political disaster instead of a moderate course correction.

Chris M.

What about the really interesting result of the study? If they thought that they'd been given testosterone, they behaved in a manner exactly opposite of the statement you've made there, because of the cultural perceptions behind it. Being able to test four categories (received testosterone/did not receive testosterone combined with thought did/thought didn't) resulted in strong evidence of social perception on our biological processes.

In both categories, those who thought they had been given testosterone and those who were uncertain, those given testosterone made the larger initial offer. However, those who thought they were given testosterone in both comparisons behaved less fairly.

The strong, opposite perceived effect of testosterone, tested directly against its actual effect, is at least as interesting.

As far as Eli's comment... the researchers eventually came up with a reasonable explanation for a result that ran counter to prevailing dogma, and I'm having trouble seeing a problem with that in the scientific sense. It's a wonderfully simple and powerful result, for both perception and biology affecting our behavior.


Timmy B

What kind of women accept testosterone for an economics experiment? This isn't cancer research. I would raise questions as to whether this group represents a random sample of society. Also, as the book brings up, would they behave similarly if they were not being watched?


The results seem pretty straightforward: testosterone makes women more fair and reasonable. Apparently, lack of testosterone is the answer to the musical question:

Why can't a woman be more like a man?
Men are so honest, so thoroughly square;
Eternally noble, historically fair.
Who, when you win, will always give your back a pat.
Why can't a woman be like that?


Giving more money = better odds of being perceived as generous = indicator of higher income and status = better chance of getting laid.

Welcome to manhood, ladies.

*throws wad of 100's up in the air*


The most parsimonious interpretation of the data:

1. Women who received testosterone made offers that were more fair and reasonable.

2. Women who did not receive testosterone continued to make offers that were unfair.

3. Women who received testosterone were no more likely to accept offers from other women because, as newly reasonable people, they rejected the unfair offers that other women continued to make..

4. Women who thought they had received testosterone made offers that were even less fair than usual, because women (being unreasonable) incorrectly perceive men as being unfair; hence, they adjusted their behavior to conform to their incorrect perception.

I suppose what we need to know, in order to complete the circle, is whether women who received testosterone were more likely to accept offers made by women who received testosterone than by women who had not.

A different D

Which may explain why it takes a group of men about 10 seconds to figure out where to grab a drink, whereas with women.....something in the range of two hours.

yet another D

Real men don't do stupid research studies. Testosterone drives us to say, "this is nonsense" and go out for the aforementioned beer.


In studying the actual vs. perceived effects of alcohol, researchers divided the subjects into four groups along two dimensions: alcohol vs.non-alcohol, and having been told the drinks were alcoholic vs. non-alcoholic. Thus some people who had been told they were consuming alcohol, in fact were not, and some who had been told they were consuming orange juice were actually drinking screwdrivers. One result? People who drunk o.j. believing it to be alcoholic acted more "drunk" than the people who had actually consumed alcohol, but believed that they hadn't.


VERY interesting research that explores parasites' role in promoting risky behavior in mice:


the viking

Wow, what brilliant research! Lets give them all hysterectomies and testosterone patches.
Where can I get a job doing research like this?

chris hauser

jeez, if you want it, ante up.


Timothy B. is correct. If a woman takes testosterone for other than a health issue, she is not of the general population of women.

Francis Galton

There is a simple alternative explanation that would be consistent with testosterone having no effect (or even mildly cause aggressiveness): The women were asked about whether they believed they had received testosterone or placebo after the experiment. Their guesses were correlated with the treatment, but only weakly so. Suppose women who were greedy tended to justify their greed by ascribing it to testosterone. Given that there are also a few women who actually recognize that they have received a non-placebo drug and do not behave as aggressively, it will necessarily but falsely seem as if testosterone makes people gentle.

Frank Mirer

The results of the ultimatum game mimic those you would expect from a multiple encounter game. Players who deny the offer seem to act as if the counter player will learn from the result and make better offers in the future after being punished. In multiple encounter prisoner's dilemma, and in evolution, a propensity to strike the first blow reduces community welfare.

How did this get by their IRB? Testosterone injections can't be free of side effect risks.

a bit delayed

Perhaps for most men, testosterone prompts them to behave expediently and due to years of engagement in the business of civilization in ways that women tend not, partly as a result of finding ourselves forced into the background and because the interests of family and sexual life come first. My first cousin just died yesterday as the result of behavior that apparently was risky and self destructive. I just wish that I had seen it coming. I should have.


As a fan of the popular poker game Texas Hold 'Em, (yes, I'm making this relevant....) I've often wondered if there is a difference in the betting strategies of men and women if played online, where other social/cultural differences could be isolated. It seems from my experience that men are much more comfortable in making appropriately large initial bets, as well as folding when the time comes; but women are better at trap plays.

Tying into the testosterone study, late poker bets are often riskier, while early raises are perceived as risky and aggressive. Furthermore, get a table of people trapped into a big bet late in the game and you'll hear a lot more sighs and grumbling when people lose.

I've seen similar in my highly male-dominated industry, where men tend to value an early decision more over flirtation around a deal (unless it's actual flirtation from the fairer gender, which is why I would isolate that out of any experiment). Am I just sexist, or doesn't this generally confirm what we know about testosterone and gender roles in our society? BTW, are we allowed to safely discuss typical gender roles in society, or is that automatically sexist?


Original D

a bit delayed - I am sorry to hear this news.

My condolences.

Bobby G

Who's to say being risky and aggressive is inefficient and causes conflict? I commute an hour to and from work every day on a freeway, and I hate drivers who take things nice and easy in the slow lanes, slowing down all the traffic behind them as a consequence. Based on this article and what I've written so far, care to guess which gender is typically the culprit I described above?