The Biology of Sex Differences in Vocational Choice

There are striking disparities between the vocational choices of men and women in the United States and other developed countries. Generally, occupations relating to the manipulation of objects and things tend to be dominated by men while occupations relating to interactions with people tend to be dominated by women. For instance, in the United States roughly 90% of nurses and therapists are female as well as roughly 80% of elementary school teachers and social workers [1]. In contrast, occupations that the Bureau of Labor Statistics classifies as architecture and engineering are around 85% male and many computer related occupations such as software development are over 80% male. These may be the extremes, but the pattern is fairly consistent across the full spectrum of occupations and is also apparent in the educational choices of men and women. In 2012 75-85% of bachelor’s degrees in health professions, public administration, education, and psychology went to women while over 80% of bachelor’s degrees in engineering and computer science went to men [2]. Indeed these differences are consistent with the findings from the RIASEC model that uses a questionnaire to determine suitability for careers based on ones interests. It has been found men and women differ most in occupation related interests along the “realistic” and “social” dimension [3].


Percent of Bachelor’s degrees conferred to women in the United States by major between 1970 to 2012

While these differences are striking they come after a period of large increases in women’s representation in the workforce and in post-secondary education. For instance, in 1950 less than half of one percent of all bachelor’s degrees in engineering went to women compared to 17% today [2]. Similar, but less drastic, rises have been seen across other fields in higher education that were traditionally dominated by men. Although the increase has halted recently, and in some cases reversed, this is still a powerful demonstration of the effect that social and economic changes can have on the career choices that women make.

The topic of this post will be the biological factors that are likely influencing this persistent sex gap in regard to people-oriented and object-oriented vocations. Teasing apart biological from social influences can be hard especially considering the powerful effect that the environment can have on sex differences. In light of this I will be drawing on evidence that avoids entanglement with the effects of socialization such as studies of non-human primates, human neonates, and studies on the effects of prenatal exposure to sex hormones.

One way to disentangle biological and social influences on behavior is to measure the behavior of infants who have had little or no exposure to social conditioning. There are four studies that looked at the viewing preferences of male and female infants for social and mechanical objects. The ages of the participants in these studies range from 12 months old to 24 hours old. These studies typically have infants presented with stimuli made to represent social and mechanical interest such as faces and crib mounted mobiles and then track the amount of time spent looking at them. Three of the four studies found that male infants showed a greater preference for mechanic objects than did females [4][5][6]. The study that failed to replicate these findings differed in one crucial respect in that it used pictures of faces and toys rather than the actual objects or video of the objects [7]. These sex differences are consistent with toy preference later in childhood that I discuss below and demonstrate that the influence of biology on interests is present almost immediately after birth and so suggests that other sex differences in interests observed later on in life also have a significant biological component.

Another indicator that sex differences in interests are in part due to biological differences comes from studies that examine the toy preferences of non-human primates. What these studies have found is that vervet and rhesus monkeys have similar sex differences in toy preferences as do humans [8][9]. That is to say, female monkeys and female humans both tend to prefer dolls and stuffed animals while male monkeys and male humans both tend to prefer toy such as balls and cars. Because these monkeys and humans have much greater biological similarity than cultural similarity (if any at all) this suggests that these sex differences in toy preferences are influenced by our shared biology.

vervet with toys

Female and male vervet monkeys contacting toys

While studies of infants and non-human primates provide a compelling case for innate human sex differences in interests a large corpus of relevant data is available from the study of the effects of prenatal exposure to sex hormones. Two ways that scientists can explore the contributions that prenatal hormones have on human sex differences in behavior is to study the effects of extreme and abnormal variations in hormone activity caused by disease or medical intervention as well as the normal variation in hormone levels. Before exploring the finer details of how prenatal hormones effect behavior and interests it’s important to grasp the range of influence that they have on human sexual dimorphism. Congenital adrenal hyperplasia (CAH) is a genetic disorder that causes a deficiency in an enzyme needed to produce cortisol and results in elevated levels of androgen during prenatal development [10]. In addition to being associated with a wide range of increased male-typical in females behavior that will be discussed later, girls with CAH are also born with varying degree of genital virilization such as fusion of the labia and enlargement of the clitoris to the point of resembling a penis. On the other end of the spectrum, males (or rather individuals with and X and a Y chromosome) who suffer from androgen insensitivity syndrome (AIS) have defective androgen receptors resulting in the inability of cells to respond to androgens such as testosterone. Individuals with this disorder develop female external genital, incur increased feminization during puberty, and do not differ psychologically from females with two X chromosomes [11]. In fact, individuals afflicted with AIS are not normally identified until it becomes apparent that they do not menstruate. These disorders make it clear that sex hormones alone play an important role in sex differences.


XY individuals with AIS and related DSD conditions

As we can see, biological influence on sex differences in interests is supported by evidence from studies of the viewing preferences of newborns and infants, the toy preferences of non-human primates such as vervet and rhesus monkeys, and the effects of prenatal hormones play styles and toy preferences. Another valuable source of insight into the biological factors in sex differences in occupational interest and choice comes from studies that explore their relationship prenatal exposure to testosterone. A well established proxy for prenatal exposure to testosterone is the ratio of the 2nd to 4th digit length (2D:4D) where increased exposures is associated with a lower ratio [16]. The advantages of this measurement is that it uses persistent features of the body and so unlike amniocentesis can be administered after birth. It is also such a relatively easy measurement to make that researchers often have participants carry out the measurements themselves at home making online surveys possible. Unfortunately, one downside is that it has been hypothesized that the 2D:4D digit ratio is a very “noisy” indicator of prenatal exposure to testosterone and so has limited explanatory power. This limited explanatory power may account for the inconsistencies and outright contradictions found in the study of prenatal testosterone exposure and its effects.

digit ratio

A smaller ratio between the second and four digit indicating male-typical prenatal testosterone exposure

In regard to occupational interests and prenatal testosterone exposure there are two relevant studies. In one study published in 2007 in the journal Personality and Individual Differences researchers investigated the relationship between participants’ 2D:4D and their scores on the RIASEC model of career interest mentioned earlier [17]. What this study found was that for both males and females lower 2D:4D (indicating higher prenatal exposure to testosterone) was associated with a higher score in the “Enterprising” dimension of career interests and among males lower 2D:4D was associated with a higher score on the “Realistic” dimension. Unfortunately, a major drawback of this study is its low sample size of forty-seven which could account for the lack of a statistically significant association between 2D:4D and other dimensions like “Social” and “Conventional” that are known to have large sex differences. The other study that explores this topic also comes from the journal Personality and Individual Differences and was published in 2011 [18]. Again, it looked at the association between 2D:4D and scores on the RIASEC model of career interests. Unlike the previous study this study had a sample size of over 8000 participants who submitted their self measured digit ratios online. The results were that there was a very weak negative correlation between 2D:4D and “Realistic” interests among males, but no other statistically significant associations among females or other dimensions of career interests. These inconsistent results could be due to differences in method and sample composition between the studies. For example, the first study mentioned had a much smaller and more homogeneous sample size and a more reliable method of finger measurements.

Despite these inconsistencies studies that look at the association between 2D:4D and occupational choice rather than occupational interests have found more consistent results. Two studies have explored this relationship directly. One 2010 study published in the journal Personality and Individual Differences found that women in occupations with a lower proportion of women such as manufacturing and engineering had more male-typical 2D:4D. It also found that the proportion of women in an occupation was moderately correlated with overall 2D:4D although this is unsurprising since males have lower 2D:4D [19]. Another study published recently in the same journal found that women in Moscow working in “Enterprising” occupations had the most male-typical 2D:4D [20]. The results of these studies strongly support the theory that exposure to prenatal testosterone is one of the causes of the disproportionate sex composition of certain occupations, specifically those relating to sex-typical interests such as mechanics and object for men and social interaction in the case of women.

In addition of occupations, across multiple studies there seems to be a persistent relationship between 2D:4D and cognitive ability despite there being some inconsistencies. A 2011 article published in the journal Perspectives on Psychological Science  provides and overview of cognition-related 2D:4D studies[21]. Some of the findings of the studies that were reviewed are as follows:

  1. Among females 2D:4D negatively correlated math-intensiveness of college major, but no effect among men [22].
  2. Among females those with below-median 2D:4D outperformed those with above-median 2D:4D on numerical and spatial tests, but no effect among males [23].
  3. Among males 2D:4D negatively correlated with numerical ability, but no effect among females [24].
  4. 2D:4D negatively correlated with overall profitability in high frequency stock traders [25].
  5. Male 2D:4D negatively correlated with math ability. Female 2D:4D positively correlated with verbal ability [26].

Although there remain some contradictions as other studies found that 2D:4D did not predict mental rotational ability [27][28] and that faculty from math intensive disciplines had higher 2D:4D than social science faculty [29]. Despite these inconsistencies the relationship between prenatal exposure to testosterone may yield insights into the under representation of women in mathematical and quantitative fields of study and occupations since prenatal testosterone exposure has a large influence on overall sexual dimorphism. On the other hand it is clear that further research into this area is necessary.

An obvious test of whether or not sex differences in prenatal testosterone can account for sex differences in vocational choice would be to compare population with different level sexual dimorphism in regard to 2D:4D to see if there is a correlation with the vocational choices of men and women. Thankfully, a 2014 study published in the journal Evolutionary Psychology does just that [30]. It looked at whites in a sample of 29 predominantly white nation (since 2D:4D is substantially influenced by ethnicity) and found a moderate correlation between political and economic gender equality and the degree similarity between male and female 2D:4D That is, in countries where the sexes have more similar prenatal exposure to testosterone also they have more gender equality in parliamentary and labor force participation.

When taken as a whole it is clear that there is a concordance of evidence to suggest that biology plays a role in sex differences in vocational choice. Studies of infants indicate that these differences are apparent before any socialization can occur, the toy preference of non-human primates suggest that the sex differences exist in our common biology, and although it can be inconsistent at times, the relationship between sex-typical behavior and preferences and prenatal hormones makes a strong case that the different choice men and women make in regard to their vocations is not solely a function of socialization.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s