"Hormones and behaviour are linked in dynamic and complex ways, more complex than we often think," LiveScience quoted van Anders as saying.
"Hormones can change depending on the context, and our behaviours and perceptions can even affect these endocrine responses. So even the same situation can elicit different patterns of hormonal responses depending on how people behave or perceive the situation," she said.
In other research, she found that surging testosterone doesn't necessarily mean an increase in stereotypically masculine behaviours. In young men, for example, she found that individuals with higher testosterone levels have stronger acceptance of safe sex, despite stereotypes that such men would be sexual risk-takers.
Both human and animal research, however, has linked parenting to lowered testosterone levels in men; a few outlier studies have found that the sound of babies crying increases, rather than decreases, male testosterone.
"This completely contradicts theory and the larger body of evidence," she said. The discrepancy got van Anders and her colleagues thinking. After all, van Anders said, "parenting" is a broad behavioural umbrella, as, one can cuddle their children, or discipline them, or protect them from harm.
Perhaps different contexts and behaviours would yield different hormonal effects. So the researchers set up an experiment using interactive lifelike baby-dolls of the sort often used to teach high-school students about the responsibilities of parenthood. The dolls can make a variety of noises, including loud crying. The only way to stop the crying is to swipe a sensor bracelet on the doll and then to comfort it as you would a real baby.
Fifty-five men, mostly college-age, came to the lab one-by-one to try their hand at calming down these animatronic infants. Before they began, the men provided saliva samples for testosterone measurements, and also answered questions about their mood.
Next, they were assigned to one of four groups. Some of the men simply sat quietly, leafing through a book of photography, before giving a second saliva sample and heading home. These men were the control group.
The other three groups all experienced a short stint with an upset baby-doll, programmed to cry with increasing intensity for about eight minutes. Some of the men were given the sensor and told to comfort the baby. Others were told to comfort the baby, but were not given the sensor, so their efforts were doomed to fail. Men in the third group only heard the baby cries via recording, and had no chance to stop the wailing.
The dolls "are so realistic, they can be almost kind of creepy when you first realize they're not real, but it's really quite amazing how invested people do get", van Anders said.
"They really, really want to help the baby calm down. It's sort of hard to believe if you've never seen it,' she said.
After their attempts to calm the baby-dolls, the men provided a second saliva sample so that researchers could measure changes in their testosterone over the experiment.
The results confirmed van Anders' suspicion that different situations would have different hormonal effects. The men who comforted the babies unsuccessfully saw no testosterone changes. But men who got to soothe the crying doll saw their testosterone levels drop by 10 percent.
Those who only heard the cries and couldn't respond, on the other hand, experienced 20 percent increases in their testosterone levels. According to van Anders, the opposing hormonal changes could be linked to different parenting behaviours.
"Hearing an increasingly upset baby, with shrieking cries that are rising, without being able to provide a nurturant response might cue a danger or emergency physiological response for infant protection," she said.
That could spur a flood of testosterone, as theories have linked higher testosterone to just this type of "I will protect you" behaviour. Meanwhile, the nurturing men were busy providing close and warm care, potentially stimulating a drop in testosterone.
The study will be published in an upcoming issue of the journal Hormones and Behavior.