From Ergo Log
The more testosterone there is circulating in a man's body, the more honest he's likely to be. Even when he knows that no one is likely to find out that he is lying, he's more likely to tell the truth than a man with less testosterone in his body. Researchers at the University of Bonn in Germany write about this in PLoS One.
Until recently researchers have tended to emphasise the negative psychological effects of testosterone, such as a reduction in empathy and increased aggressive tendencies. [Biol Psychiatry. 1999 Feb 1;45(3):254-60] But it seems that testosterone also has positive psychological effects. Ten years ago, for example, psychologists at Georgia State University published a study in which a group of firemen with high testosterone levels were more likely to do their work well. [Journal of Research in Personality Volume 37, Issue 2, April 2003, Pages 107–115.]
In 1997 the same psychologists published a study on the relationship between behaviour and the testosterone level of female prisoners. [Psychosom Med. 1997 Sep-Oct;59(5):477-80.] While women with relatively high levels of testosterone in their blood were often more violent, the prison guards surprisingly often described women with low testosterone levels as being 'sly' and 'unreliable'.
The Germans decided to study this aspect of testosterone in more depth. Testosterone is linked to pride and status seeking, they reasoned. Proud people lie less often. So subjects should become more honest if you give them testosterone – for example by rubbing the contents of a sachet of Testogel into their skin, which would supply them with 50 mg testosterone.
A score of 6 earned nothing. The researchers didn't check whether the men were cheating.
A similar-sized control group were given a placebo, and had to go through the dice routine too.
The men that had been given testosterone had an average level of 7.78 nanograms per millilitre during the experiment. The level in the placebo-group men was 6.79 nanograms per millilitre.
The men in the testosterone group won 3.33 euros per throw, the men in the placebo group won 4.18 euros per throw.
The figure above shows that the testosterone men said more often that they had won nothing or only a few euros, and less often that they had scored the maximum 5 euros.
"Our main finding is a lower incidence of self-serving lies in the testosterone group", the Germans write. "We observe this result in a setup where subjects cannot be caught lying. To the best of our knowledge this is the first piece of evidence on a causal relationship between testosterone administration and prosocial behavior when actions are not observable to others."
PLoS One. 2012;7(10):e46774.