Recently we published a study in the Proceedings of the National Academy of Sciences reporting a drop in testosterone levels among newly partnered new fathers. We were gratified to see those findings discussed widely, from the front page of the New York Times to spirited conversations among groups of parents. As the study percolated through the media and across blogs, inevitably some of its nuances, interesting implications, and important limitations were obscured. In addition, news of our results was also used by some to reinforce unfortunate stereotypes about masculinity. Critiques of the interpretation of the study and especially its portrayal in the media have been published (for example, see critiques here and here). Clearly, these findings struck a chord!

We would like to contribute to the discussion by providing more scientific detail and context for our findings. We welcome this opportunity to discuss, for example, the previous research that motivated our question, exactly what we found in our study, and why we interpret our findings as we do. We will also discuss the gaps in current understanding that leave open important questions. As others have done in response to this study, we would also like to reflect upon what we think these findings imply about the nature of human males and human fatherhood.

Brief introduction to the problem

The idea that testosterone might drop in new fathers first appeared in studies of other species in which mothers and fathers cooperate to raise their young. For instance, fathers are especially involved in caregiving in many bird species, because eggs need to be incubated and protected, and the chicks also need to be fed after they hatch. In many of these species, it was long ago shown that testosterone levels are often high during periods of mating, but then decline once males began caring for their young. Although less common, males among the rare species of mammal in which fathers care for offspring in some cases also show a decrease in testosterone. Researchers hypothesized that this might make good sense: given that testosterone motivates behaviors related to finding and competing for mates, perhaps this social drive was no longer needed, and might even be a distraction, as a male’s duties shifted to caring for dependent young.

Starting more than a decade ago, researchers began reporting evidence suggesting that human fathers might also share this pattern of testosterone response. In 2000, Anne Storey and colleagues published a study showing that expectant fathers had higher testosterone late in their partners’ pregnancies compared to new fathers shortly after their infants were born. Anthropologists working in societies across the globe found evidence that fathers have lower testosterone than non-fathers, and in some instances, also that married or partnered men have lower levels than single men (see these fascinating studies in Boston, China, Tanzania and Senegal; see an in-depth interview with Peter Gray, the author of several of these studies and a pioneer of fatherhood research, at this blog post by Kate Clancy as well as his recently co-authored book). In 2009, we published our first article on testosterone and fatherhood in the Cebu Study that was the focus of the recent PNAS article. Consistent with findings in these other societies, we found that fathers in committed relationships had lower testosterone than single non-fathers. We also showed that fathers who reported being a primary caregiver to their children had low testosterone, especially in the evening, compared to fathers who did not consider themselves one of their child’s primary childcare providers.

Thus, many biologists and anthropologists, including our research team, have assumed for some time that fatherhood likely lowers a man’s testosterone levels, as many of the past studies suggested that men likely conform to the pattern seen in other species with paternal care. However, this evidence, although fascinating, did not allow definitive conclusions to be reached about whether fatherhood reduced a man’s testosterone levels. This was because these studies were “cross-sectional”, meaning that they only measured fatherhood status and hormones as a single snapshot during one visit and interview. Had the act of becoming fathers reduced these men’s testosterone? Or alternatively, were the men who had low testosterone to begin with simply more likely to have become fathers? Both scenarios could explain these findings equally well, and so the matter remained unsettled.

What we found in our new study

As noted above, when we first measured testosterone among the men in our study in 2005 we found that fathers had lower testosterone levels than the men who were single at that time. What makes our new study unique is that we measured these same men a second time about 5 years later. Importantly, because the men were quite young, many of them became fathers for the first time during this 5 year period, which allowed us to test whether new fathers experienced a greater drop in their hormone levels compared to men who remained single and childless.

We found that the single men who ended up being partnered fathers 5 years later had higher testosterone when the study began. This is important, because it allowed us to rule out the other possible interpretation of our data – that men with low testosterone were simply more likely to become fathers. That was clearly not the case in our sample.

We also found that, once these men became fathers, their testosterone levels went down substantially, and more dramatically than other groups of men. In fact, most of the men in the sample experienced some decline in testosterone during this 5 year period. Testosterone levels tend to peak in early adulthood and then gradually decline with age, so this was expected. But the new fathers experienced much greater declines than would be expected based upon age alone.

Because researchers had previously speculated that the act of providing direct care of offspring might cause a decline in testosterone (see for instance here and here), we made sure to ask men in 2009 about the amount time they spent in childcare each day. We found that men who reported spending more time in childcare also had the lowest testosterone levels in 2009, which provided additional support for this hypothesis.

There were other nuances to our findings, some of which we were surprised to see. For instance, although all fathers as a group had comparatively greater drops in their testosterone across the 5 years of the study, the hormonal drop was much larger among men who had newborn babies. In these men, testosterone levels were about half their former pre-fatherhood levels measured in 2005. When we looked at men whose youngest child was older than 1 month, we found that their testosterone drop was less than in this group of men with newborns. Thus, our study also suggested that men experienced a very substantial drop in testosterone immediately after the birth of their child, and that their hormone levels rebounded somewhat (but still quite a bit below their pre-fatherhood levels) and remained relatively stable for a year or two afterwards.

These findings generated a great deal of media and scientific attention. They are the first demonstration in humans that fatherhood seems to drive down a man’s testosterone, rather than the other way around. Our data suggest that the drop in testosterone may be especially dramatic soon after birth but then rebounds partially afterwards. They also suggest that the extent to which a father’s testosterone drops or stays relatively low is somehow linked with his commitment to caring for offspring. While these findings helped answer certain questions, they also led to new ones. What is the cause of the decline in testosterone? What are the likely effects on a man’s behavior or health? As is often the case with new findings such as ours, we do not have definitive answers to many of these questions. However, we will do our best to address a few of the implications of these findings below.

Why should fathers have lower testosterone? Does this “make sense”?

There has traditionally been a great deal of debate about the effects that testosterone has on human behavior. In general, we can think of the hormone as providing a bit of a competitive and confidence boost. Classically, the hormone has been seen as motivating behaviors that help males compete for and attract mates. In our study, we found that the men who entered relationships and became fathers during the 5 years of the study had higher testosterone to begin with. Although we did not investigate what behaviors or other characteristics might explain the tendency of higher testosterone males to find mates, this finding is broadly consistent with our basic understanding of the role of the hormone in many species.

Some of the effects that are encouraged by testosterone focus one’s attentions and energies on the social realm. Just as these behaviors could improve the likelihood of a young man finding a mate and becoming a father, they might then become a detriment when duties shift to caring for dependents. For instance, it is known that men with higher testosterone are moderately more likely to express aggressive behaviors . Higher testosterone men are also more likely to be “sensation seekers” and to engage in risky behaviors, such as drug or alcohol abuse. In addition, prior research shows that men with higher testosterone are more likely to be promiscuous, and indeed, have been shown to have a greater number of lifetime sexual partners. Personality differences between men, and cultural variation in social norms, can modify the relationships between hormones and behaviors like these. But the general assumption among many anthropologists who study the role of hormones in human fatherhood is that these behaviors represent time investments that would likely detract from a man’s ability to be an effective and attentive father and caregiver.

Although there are very few available studies that have evaluated the effects of testosterone on fathering behaviors, there is preliminary evidence that lower testosterone can sensitize men to being more involved as caregivers. As we noted above, in our study we found that fathers who reported being involved in childcare had the lowest levels of evening testosterone. Perhaps more compelling evidence is reported by a study which showed that fathers with lower testosterone expressed greater sympathy towards, and a need to respond to, infant cries. Similarly, a recent study, released just this week, found that men’s testosterone decreased when they heard recorded infant cries, but only if they felt a need to nurture the upset infant. In a separate study examining hormonal changes during father-toddler interaction, Dr. Anne Storey and colleagues concluded that “…low baseline testosterone levels or greater testosterone decreases are associated with men being more responsive or attentive to their toddlers.”

Much more research will be needed to clarify the possible effects of lower testosterone on men’s abilities to perform as effective caregivers. However, as we see above, there is at least preliminary evidence that having high testosterone can focus a male’s attentions and energies on social activities, while low testosterone appears to improve the attention and care that men provide their own young. Both point to the potential benefits of reducing levels of the hormone as social duties shift with the onset of fatherhood.

What is the cause of the decline in testosterone among fathers?

In our study we found that fathers who reported spending the most time (3 or more hours each day) with their offspring had lower testosterone levels compared to other fathers who reported spending no time caring for their young. Here once again it is difficult to know what is “cause” and what is “effect”: does childcare reduce a father’s testosterone, or were fathers with lower testosterone simply more inclined to be involved caregivers? The short answer is that we cannot say for sure, but we do have some hints. Our best guess is that not only did fatherhood lead to a decline in testosterone, but interacting with a child, or some correlate of time spent with a child, influenced how far the levels dipped once men became fathers. Let us explain how we reached this tentative conclusion.

Researchers had previously speculated that providing direct care might lead testosterone levels to drop in new fathers. For instance, in a study in Senegal, fathers who showed greater investment in their children, including hands-on childcare, had lower testosterone than fathers who were reported to be less invested. Similarly, another fascinating study looked at testosterone levels in fathers in two cultural groups in Tanzania. Testosterone was only lower among the fathers in the cultural group in which fathers are expected to be involved in offspring care – in the second group, among whom fathers typically are not involved caregivers, fathers did not have lower testosterone than non-fathers. These findings suggested that it is not simply being a father, but a committed father who spends time with his offspring, that seems to influence testosterone levels. Paralleling these findings, we published an article in 2009 that considered a cross-sectional (“snap shot”) study of relationships between fatherhood, childcare, and testosterone in the young men from Cebu. We found that fathers who identified themselves as primary caregivers to their children had lower testosterone compared to fathers who did not identify themselves as caregivers. Taken together, these studies set the stage for us to consider in greater detail in our new study in PNAS whether the amount of time that a man spends in childcare relates to his testosterone levels.

In our new study, we found that men who reported spending more time in childcare each day had the lowest testosterone levels in 2009. As with examples above, there are several possible interpretations of this finding that we had to consider. If the men who had lower testosterone to begin with were “predisposed” by their lower testosterone to be more involved parents compared to other fathers, then we might expect that men who had lower testosterone before becoming a father, when we collected the first hormone measurement at 21 years of age, would end up spending more time with their children at age 26. In fact, as we reported in the study, we did not find this – the amount of time a man spent in childcare in 2009 was unrelated to his baseline testosterone levels. Based upon the finding, we think it unlikely that low testosterone before men were fathers led to more childcare. Instead our best guess is that childcare, or some behavior or activity that is associated with childcare, somehow lowered the testosterone levels of highly-involved fathers in our sample.

It is important to note that several recent experimental studies (here and here), including our own, have failed to find that testosterone drops immediately in fathers after they interact with their children. For instance, in a recent study of 42 fathers in our study, we looked at hormone levels before and 30 and 60 minutes after they interacted with their offspring. Testosterone levels were stable across this period. This could indicate that interacting with a child does not lead to a drop in testosterone. However it could also indicate that we did not measure men over a long enough period of interaction to see a biological response. For instance, a father’s body might only “reset” hormone levels in response to more extended periods of interaction, such as over weeks or months. In addition, the fathers in this smaller study likely had already experienced large declines in their hormone levels because they had been fathers for 4 years, on average, so perhaps we would have needed to perform this experiment when their offspring were younger to see the decline.

To complicate the picture further, there is another subtly distinct possibility that we were not able to test with our data: fathers who experienced the largest drop in testosterone as a result of becoming a father might then be more inclined to be more involved in childcare as a result of their lower testosterone. That is, was the intensity and number of hours spent in childcare influenced by the drop in testosterone experienced by new fathers, rather than vice versa? Addressing these various possibilities will require new studies with more fine-grained data on testosterone and paternal care as fathers engage with their newborn babies in the weeks and months after birth. We look forward to seeing what future studies reveal!

What are the possible effects of the testosterone drop on a man’s behavior and health?

Fathers who learned of our findings expressed a range of emotions, from excitement to fascination to concern. If testosterone declines with the onset of fatherhood, what effects on behavior or health might this have? Here we briefly review the evidence linking testosterone with several of the traits that men seemed to express most concern about, in hopes of dispelling some of the concerns that were raised.

First, as we noted above, the largest drops in testosterone were seen in fathers whose partners had only recently given birth to newborns. Although fathers’ testosterone declined more than non-fathers regardless of child age, evening testosterone declined a median of 49% in fathers of newborns compared to 34% in other new fathers. Thus, the first point to keep in mind about our study’s findings is that the biggest drop in testosterone appeared to be temporary, and to rebound somewhat about a month after birth.

Second, although testosterone is an important male reproductive hormone, many reproductive functions are maintained at quite low levels of the hormone. Common sense and every day experience tell us that fathers must maintain the ability to have additional children, even if their testosterone decreases, because many men have large families with multiple offspring. Indeed, of the 162 men in our study who transitioned to fatherhood between ages 21 and 26, roughly 27% had two or more children during that period. Keep in mind that there are also good evolutionary reasons for basic reproductive functions to be maintained even at low levels of testosterone: natural selection would not favor a biological response in which fatherhood caused testosterone to drop to sub-fertile levels, because this would prevent men from producing more than one child, which would minimize their contribution to the gene pool.

The magnitude of the hormonal change that we documented could have some subtle effects on some behaviors and traits, depending on how long they persist. For instance, higher testosterone has been associated with libido in some studies. It is widely known that couple’s sex lives often change after offspring enter the picture! How much of this might trace to the hormonal changes that we document? The simple answer is that we are not sure, but surely other factors are equally if not more important. Perhaps most obviously, the intense care required by infants and toddlers often leaves parents fatigued and stressed and can also reduce opportunities to engage in sexual intercourse. It would surely be a simplification to attribute all post-baby changes in sexual behavior to the modest changes in testosterone that we document.

Our study was not able to determine whether the changes in testosterone among new fathers were stable over longer periods of time, or whether they might rebound after a year or two. If more sustained, the decline in testosterone might be expected to have some effects on body composition, because testosterone promotes muscle mass and helps the body burn off body fat. This might be particularly true if men’s diet and physical activity levels change after they transition to fatherhood, as research suggests is often the case.

Other health effects will require additional research to clarify. On the one hand, high levels of sex steroids have long been known to predispose women to reproductive cancers, and there has similarly been some evidence for a link between high testosterone levels and prostate cancer in men. However, several recent studies found that fathers had higher prostate cancer risk, although their risk went down as they had more offspring. Similarly, there are debates among the medical community about a range of both good and bad effects that testosterone might have on a man’s risk for cardiovascular diseases and related risk factors, such as blood pressure or cholesterol profiles. As we pointed out in our study, we hope that our findings encourage researchers to design studies to evaluate the possible effects of a father’s decline in testosterone on health differences among fathers. As of now, we do not think there are any reasons to be concerned about the health effects of being a father. In fact, there is growing evidence that fathers tend to be healthier than non-fathers, such as was shown by this new large study published less than a week after our PNAS paper was reported.

Why did we conduct this study in the Philippines?

The men in this study live in urban and rural neighborhoods in the metropolitan area surrounding Cebu City, a city of almost 3 million people, and the second largest city in the Philippines (itself a nation of nearly 100 million). This is a fascinating time to study fatherhood in the Philippines, because more and more fathers are actively participating in childcare and there have been few studies of fatherhood in the country. Our research and that of others has shown that within this generation (men currently in their 20s and 30s) men are increasingly helping take care of their children, which is a role fewer fathers embraced in the past in Cebu and elsewhere in the Philippines. The caregiving duties that fathers in our study reported most commonly will likely sound familiar to many fathers who read this blog: watching TV, feeding and bathing, playing games, and taking their children on walks. Thus, we predict that the findings that we report for this population will be broadly-applicable to other global populations with similar models of paternal involvement.

In addition, there were good practical reasons to conduct the study in this population. Our study required that almost 1000 men were interviewed and had samples collected for hormone analysis at two times that were years apart – before and after many were fathers. As you might imagine, this was no small undertaking! We were able to conduct such a large and detailed study because these men were already being followed as part of a longer-term study called the Cebu Longitudinal Health and Nutrition Survey (what we refer to above as the “Cebu Study” for short), on which all of the authors on this paper collaborate. The study has been run “on the ground” since its inception by the many scholars and researchers at the University of San Carlos Office of Population Studies Foundation, in Cebu City, where one of us serves as Director (AF). The Cebu Study is known as a "birth cohort" because the participants were first enrolled when their mothers were pregnant with them in 1983-84, and it has since followed each of these offspring from birth to the present as they have grown into young adults. It was this ongoing study, which was already following a large group of young adults, that allowed us to test how becoming a new father might influence a man’s testosterone levels.

A few final observations

It has recently become clear to anthropologists that humans reproduce using a strategy of “cooperative breeding,” meaning that during human evolution it was not just mothers alone, but a larger, flexible social group that helped mothers raise their offspring. As anthropologists have developed this model, there has been much discussion and debate about who would have been involved as caregivers. Some have argued that grandmothers and other female kin were likely key providers of care who assisted mothers with childrearing duties, with others noting that older siblings are often important as caregivers of younger siblings. Although there is little question that parenting roles vary a great deal across cultures, fathers were likely in a good position to cooperate with mothers to raise their young under many circumstances during human evolution. Our study lends support to this idea, because it confirms that human male biology has an ability to respond as childcare duties emerge with the transition to fatherhood. We suggest that this built-in responsivity in male biology likely evolved, in part, because the survival and well-being of children were improved if fathers were highly invested in their partners and offspring.

Fathers have been shown to have lower levels of testosterone in many societies, and our study simply confirms that this is likely a result of fatherhood leading to a decline in the hormone. Because paternal care is not present in the living primates most closely-related to humans (chimpanzees and gorillas), it seems likely that this biological capacity appeared sometime more recently during human evolution in conjunction with the unique way that we raise our offspring. These biological changes have likely been experienced by countless generations of human fathers, for hundreds of thousands of years if not more. We hope that fathers will share our enthusiasm in knowing that men, in ways that may be unique to our species, are capable of responding biologically to parenthood, suggesting that their male ancestors have long played an important role in raising children.

Photos by Fe Largado