Confession: I yell at TV commercials. I mean I scream at them. The trigger may be poor edits, overwriting, bad enunciation or—the worst of the worst—bad statistics combined with minimal knowledge of basic biology.

These days the most serious offenders are advertisements for testosterone replacement. We can dispose quickly of the bad statistics implied by the slogan “it’s low T, not age.” Age and testosterone levels are correlated, so it is not an issue of one versus the other. But linking manliness with testosterone, as the ads do, is a meme of long standing and harder to dislodge; indeed, it is embedded in our language. Androgen, the hormone group that includes testosterone, means “to build a man,” while estrogen means “to provoke a frenzy.” This terminology compromises our ability to think clearly about the complex biological roles hormones play in our development and behavior.

Which leads me to the bad biology. Testosterone is not, as the same ad says, just a number. It is a mutable, subtle, modulated hormone. Its levels change with social context. It makes friends with and also antagonizes other hormones. And as biologists come to understand it better, we find that if we want to keep it as an emblem of masculinity, our understanding of what counts as masculine and feminine needs to change. The results of a decade and a half of studies on fatherhood, specifically the relationship between hormones and nurturing behaviors, are instructive.

Fathers? We are used to thinking of new mothers as hormonal and emotional basket cases. Muddled thoughts, flowing tears, glowing faces. But what about men? Even expectant fathers experience morning sickness and other symptoms of pregnancy, a phenomenon known as couvade. Studies have found couvade in as little as 11 percent of subjects and as much as 97 percent. We don’t know why these numbers vary so wildly, but there have been lots of theories to explain why dads get morning sickness. Psychoanalytic ones focus on pregnancy envy or imagining the fetus as a rival. Psychosocial theories examine the marginalization of men during pregnancy and birth. But other ideas are more grounded in biology, especially the possibility that couvade symptoms evolved to prepare young men for fatherhood by synchronizing them with their partners and focusing their attention on the prospective birth.

Testosterone is thought of as a molecule that creates social structure. This false belief leads to troubling consequences.

Couvade appears to be mediated through hormonal changes that resemble those of pregnancy. One study sampled levels of four hormones: cortisol (related to stress physiology) and prolactin (noted for its role in stimulating milk production before and after birth) in men and women, testosterone in men, and estrogen in women. Prolactin levels rose in both the prospective fathers and mothers early and late in pregnancy and even more after birth. Among all of the fathers, testosterone levels fell right after birth, and men who showed greater concern for and responsiveness to baby stimuli or had couvade symptoms had larger drops in testosterone and larger increases in cortisol and prolactin. Men’s hormone levels correlated with their partners’; that is, in a man, testosterone and fellow travelers respond to the biosocial context—in this case the partner’s hormonal state.

Several scientists have offered evidence that becoming a father leads to lower testosterone levels. In particular, the anthropologist Lee Gettler tells a nuanced tale using well-designed longitudinal studies. There is much still to be learned, but his basic finding is that partnered fathers average much lower testosterone levels—as much as 30 percent lower, depending on the time of day—compared to their pre-fatherhood selves. Gettler’s data support the hypothesis that child interaction suppresses testosterone: fathers who spend three or more hours a day on childcare have lower testosterone compared to uninvolved fathers. The data also show that nurturing an infant causes the drop in testosterone, not vice versa.

Now, I can hear the gears grinding. Is that low testosterone going to affect a new dad’s sex drive? This is just the kind of fear the hormone replacement ads play on. There is no doubt that having children changes a couple’s sex life. But within the normal range of testosterone, which includes the lower levels common among new fathers, the evidence linking libido and testosterone levels is pretty weak. So, not to worry. Becoming a nurturing dad does not turn you into a eunuch.

In fact lower testosterone levels might actually be good for one’s health. Lower testosterone may improve immunity, including resistance to a variety of infections. It may reduce propensity toward unsafe behavior, improve heart function, and decrease the risk of type II diabetes and prostate cancer. Fathers who nurture their children just might live a little longer than those who don’t.

There is, of course, a bigger picture when it comes to thinking about hormones. Many testosterone enthusiasts not only correlate masculinity with high levels of the hormone but also believe that testosterone causes aggression, high sex drive, and a desire to mate. This same framework associates low testosterone with femininity and parenting.

Does this mean that “real” men can’t be caregivers? Psychologist Sari van Anders doesn’t think so. She argues that testosterone levels are linked to, but not necessarily causes of, behaviors that can be associated with both men and women. For example, high testosterone in either sex correlates with competitive behaviors, while nurturing behaviors encourage lower testosterone levels. Testosterone and other hormones do not define masculinity or femininity—“real” men can be competitive and nurturing and “real” women can be caring but also competitive and have a high sex drive.

Testosterone is thought of as a molecule that creates social structure. That belief can lead to troubling consequences. Consider a new publication by the legal scholar Kingsley R. Browne. He explains that women earn less than men because they have lower testosterone levels. Their differing hormonal profiles guide women and men toward different choices in education, the workplace, and the family, and it is these choices that lead women to take jobs offering lower wages. “Despite the frequent assertion that the gaps that favor men (although not those that favor women) are results of invidious social forces, the truth seems to be somewhat more basic,” Browne writes. “If the various workplace and non-workplace gaps could be distilled down to a single word, that word would not be ‘discrimination’ but ‘testosterone.’” Using a hierarchical model with testosterone at the beginning of a causal chain, Browne links a long list of traits—including dominance seeking, aggression, spatial-reasoning ability, occupational preferences, entrepreneurial spirit, nurturance, and empathy—to higher or lower levels of testosterone. Because he believes gender gaps in earning and employment reflect natural differences between men and women, he also thinks they will not go away.

Browne agrees with the ad that testosterone is just a number, one that we should accept when it comes to low wages, if not when it comes to restorative pharmaceuticals. But if Gettler, van Anders, and others are right that circumstances and behaviors often change testosterone levels rather than vice versa, and if Browne’s hierarchical model cannot account for the evidence that testosterone levels change according to life history and social relations, then Browne is wrong. Social structure shapes that other number—seventy-nine cents on the dollar.

Hormone levels, behaviors, and social circumstances interact in complex ways that produce embodied gender. We—scientists and laypeople alike—try to tag bodies as masculine or feminine, but our attempts at prescribing gender fail because our biology is both more complicated and more interesting than a simple set of labels allows.