Home Overview Press Room Blog Publications For Students about us

Girls R Us: Sex Selection, Sound Bites and Weak Data

by Gina Maranto, Biopolitical Times guest contributor
November 21st, 2012

I finally got around to reading Hanna Rosin’s article, “The End of Men,” which ran in the Atlantic Monthly in July/August this year. I’d heard some of the noise about the piece back then and figured that like other members of the genre (e.g., Francis Fukuyama’s The End of History and the Last Man, or John Horgan’s The End of Science) the title served more as a provocation than a believable prognostication – Fukuyama published in 1992 and Horgan in 1996, and last I checked, history and science are still ginning along.

I was actually kind of sorry to find that I was right. I’m not going to take on Rosin’s overarching argument about the ascendancy of women on a post-industrial planet, mainly because where to start? Despite a few nods to trends in Korea, India, and China, her optic is white, 1-per-centish, and American. But I am going to take apart her claim that there is a global shift afoot that is “eroding the historical preference for male children,” which is based mostly on, it seems, wishful thinking, suspect data, and faulty logic.

Rosin fills most of her seven-paragraph lede with a lot of fluff and color about the cowboy predilections of reproductive physiologist Ronald Ericsson. Given all the details about his ranch and Marlboro commercials and his Old West “swagger,” it comes as a surprise that the real reason for his being there is so that he can deliver a pronouncement about an epochal shift. Rosin quotes him as saying, “Did male dominance exist? Of course it existed. But it seems to be gone now. And the era of the firstborn son is totally gone.”

Why choose Ericsson to deliver this news? Well, because in 1973, Ericsson and two colleagues had published a letter in Nature describing a technique for sorting Y-chromosome-bearing sperm – that is, those that would produce male offspring – from those carrying X chromosomes. The method could reliably produce samples that contained 85% Y-bearing sperm.

When I interviewed Ericsson in 1984 for a Discover piece I wrote on sex selection, the cowboy schtick was thankfully not in evidence. The concern of ethicists, feminists, and others I spoke with then was that such techniques would be used overwhelmingly to have sons. Ericsson was unconcerned, categorically dismissing any objections to sperm sorting.

He pointed me to a colleague, Nan Chico, a sociologist who had, as I wrote then,

analyzed more than 2,000 letters received by Gametrics, the Sausalito biotechnology company that Ericsson founded. About 51 per cent of the correspondents already had daughters and wanted a boy, or boys to round out their families. Only twelve people, or six-tenths of a per cent, specifically wanted a first-born son.
Chico’s conclusion was that sex selection was being sought primarily for family balancing, and a number of clinicians I spoke with said that was also their experience. But even then, what Chico saw in the letters to Gametrics – obviously, a biased sample – ran counter to scores of surveys and studies of American preferences, which consistently showed that boys were preferred, and often, first sons. For example, in a 1992 review article published in Women’s Studies International Forum, “Sex Selection and Reproductive Freedom,” political scientist Alison Dundes Renteln noted that the worldwide preference for sons “has been extensively documented and there is no evidence that this preference is declining” –and provided plenty of citations to prove it.

Yet Rosin reports that,
In the ’90s, when Ericsson looked into the numbers for the two dozen or so clinics that use his process, he discovered, to his surprise, that couples were requesting more girls than boys, a gap that has persisted, even though Ericsson advertises the method as more effective for producing boys. In some clinics, Ericsson has said, the ratio is now as high as 2 to 1. Polling data on American sex preference is sparse, and does not show a clear preference for girls. But the picture from the doctor’s office unambiguously does. A newer method for sperm selection, called MicroSort, is currently completing Food and Drug Administration clinical trials. The girl requests for that method run at about 75 percent.
Rosin’s “unambiguous” picture is wholly inadequate as evidence of anything other than the power of small samples to mislead. Ericsson’s claim about what goes on at clinics using his method cannot possibly be generalized. And Rosin doesn’t say whether he published these findings (I surveyed the available abstracts of 52 of his publications from 1962 to 2012 via Web of Science and saw no article that appears to present that data, but I could have missed it).

It’s also unclear from Rosin’s phrasing who the source for the percentage of “girl requests” at clinics using MicroSort is, and whether data was systematically collected or the 75% figure is merely anecdotal. Call me skeptical, but I conclude only that Ericsson says that there’s a 2 to 1 preference for girls at clinics using his method, and maybe there is and maybe there isn’t.

The fact is, we know extremely little, through direct measures, about the degree to which people are engaging in gender selection in the doctor’s office in the U.S. Take just the data from IVF clinics for starters. The voluntary Assisted Reproductive Technology (ART) Reports filed by many IVF clinics annually (443 in the 2010 edition, the latest available from the Centers for Disease Control and Prevention), tally types of procedures, pregnancy success rates from types of cycle (like, how many live births resulted from fresh embryos from nondonor eggs, or how many from frozen embryos from nondonor eggs), and asks clinics what services they provide. Clinics are not asked if they provide gender selection. Moreover, the gender of infants is not broken out in the nationally reported data on live births. We would have to go into data kept by the clinics themselves to get those figures, and as far as I know, no one has done that on a wide scale. So in looking for evidence of gender selection we hit a dead end via that route.

We could look at a number of other pieces of data in the ART Reports, such as the number of cycles in which pre-implantation genetic diagnosis (PGD) was performed. PGD is typically used to detect genes linked with some heritable diseases, but it can also determine the presence of X or Y chromosomes. But the PGD numbers don’t tell us much, because, again, we don’t know how many of those cycles actually involved a choice of the embryo’s sex. Even if we knew, we would have no clue what the motivation was (a desire to have a child of a certain gender vs. a desire to screen out embryos for sex-linked disease genes).

So that brings us to sex ratios. Demographically, skewed sex ratios are the best single indicator that a species’ natural proportion of male-to-female births has been interfered with somehow. Among humans, the natural number of males to females at birth is 105 to 100, or a ratio of 1.05. When we look not at clinics but at total births in the U.S. what we see is that something is skewing the ratio. Take, for example, the results from a 2008 PNAS article by economists Douglas Almond from Columbia University and Lena Edlund of the National Bureau of Economic Research, titled “Son-biased sex ratios in the 2000 United States Census.”

Almond and Edlund looked at total births among “White” Americans, and found that when running the numbers for first, second, and third children, the sex ratio was normal. But when they examined births among “U.S.-born children of Chinese, Korean, and Asian Indian parents,” the ratio for the second child shifted to “1.17 if the first child was a girl. At third parity, boys outnumbered girls by 1.51:1 if the two previous children were girls.” In other words, as families grew in size, the sex ratio skewed, meaning that the birth outcomes were not random. Some sort of selection was going on, whether due to abortion after prenatal testing or sonography, or to embryo selection in IVF clinics, or to some other unidentified factor. (Multiple studies have shown that the sex ratio is not altered by previous birth outcomes, so that could not be a cause of the shift.)

Almond and Edlund note that such an effect among Chinese, Korean, and Asian Indian immigrants parallels well established trends in their home countries, where sex ratios have been significantly skewed for decades due to preference for sons. The trend in America, they add also “appears to be recent. In the 1990 U.S. Census, the tendency for males to follow females among Indians, Chinese, and Koreans is substantially muted.”

Similar findings concerning the same populations were reported in Prenatal Diagnosis in 2011 by a team from the University of Connecticut Health Center that examined live births from 1975 to 2002 using National Center for Health Statistics birth certificates in 4-year intervals. They concluded that, “The male to female livebirth sex ratio in the United States exceeded expected biological variation for third+ births to Chinese, Asian Indians and Koreans strongly suggesting prenatal sex selection.”

Which brings us to the second big claim in Rosin’s lede, which is that preference for male offspring is waning globally. Using South Korea as her case in point, she writes,
As recently as 1985, about half of all women in a national survey said they “must have a son.” That percentage fell slowly until 1991 and then plummeted to just over 15 percent by 2003. Male preference in South Korea “is over,” says Monica Das Gupta, a demographer and Asia expert at the World Bank. “It happened so fast. It’s hard to believe it, but it is.”

But surveys are not evidence of actual sex ratios, just of attitudes toward hypothetical scenarios, and the extent to which attitudes line up with actual choices isn’t known. As Sue Hall, Erin Reid, and Theresa M. Marteau wrote in a 2006 review article in Prenatal Diagnosis among other things, when answering a survey, “Many people may not consider the context in which actual decisions would be made, such as whether their partner desires a child of a particular sex.”

Such surveys have to be compared with actual sex ratios. When we look at that data, the South Korea situation is more complex than Rosin indicates. In 2011, Therese Hesketh, of the UCL Centre for International Health and Development, reported in an article in Early Human Development that the sex ratio at birth (SRB) did seem to be declining in South Korea, but from an extraordinarily high level: “South Korea was the first country to report a very high SRB, since the widespread access to sex-selective technology preceded other Asian countries. The SRB started to rise here in the mid-1980s, and by 1992 the SRB was reported to be as high as 125 in some cities.”

As for Rosin’s statement that “The same shift [as in South Korea] is now beginning in other rapidly industrializing countries such as India and China,” it remains to be seen. Although China’s SRB went down slightly in 2010 to 119, Hesketh writes, its “huge population means these ratios translate into very large numbers of excess males. In 2005 it was estimated that 1.1 million excess males were born, and that the number of males under the age of 20 exceeded females by at least 25 million.” Too, the national figures conceal pronounced regional variations; in some provinces, the SRB is currently as high as 140.  At an upcoming conference at Stanford’s Center for East Asian Studies, postdoctoral fellow Shuang Zhang will deliver a paper on “Gender Imbalance in China: A cautionary tale of land reform, income, and sex ratios,” which will present results showing that land reform and better education have not led to preferences for girls, rather to higher incidence of sex selection for boys. Hesketh says the SRB in India, particularly in urban areas, appears to have risen over the last few years.

All in all, there is little solid indication that we are seeing a major global shift in attitudes toward female offspring, much as one might wish it were true. Unfortunately, Rosin’s lede is symptomatic: to my mind, many claims in the rest of her piece are backed with equally problematic warrants that weaken or fall apart altogether under scrutiny.

Gina Maranto is Co-Director of Ecosystem Science and Policy and coordinator of the Environmental Science and Policy program at the University of Miami's Leonard and Jayne Abess Center. She is the author of Quest for Perfection: The Drive to Breed Better Human Beings (1996).

Previously on Biopolitical Times:


Add a Comment


home | overview | blog | publications| about us | donate | newsletter | press room | privacy policy

CGS • 1936 University Ave, Suite 350, Berkeley, CA 94704 • • (p) 1.510.665.7760 • (F) 1.510.665.8760