2 ajob Winter 2001, Volume 1, Number 1

� 2001 by The MIT Press

Pre co ncep tio n Ge nd er S ele ctio n

Preconception Gender Selection1 John A. Robertson, School of Law, University of Texas at Austin

Safe and effective methods of preconception gender selection through �ow cytometric separation of X- and Y-bearing sperm could greatly increase the use of gender selection by couples contem-

plating reproduction. Such a development raises ethical, legal, and social issues about the impact

of such practices on offspring, on sex ratio imbalances, and on sexism and the status of women.

This paper analyzes the competing interests in preconception gender selection, and concludes that

its use to increase gender variety in a family, and possibly for selecting the gender of �rstborn, might in many instances be ethically acceptable.

Advances in genetics and reproductive technology present prospective parents with an increasing number of choices about the genetic makeup of their children. Those choices now involve the use of carrier and prenatal screening techniques to avoid the birth of children with serious genetic dis- ease, but techniques to choose nonmedical charac- teristics will eventually be available. One nonmedical characteristic that may soon be within reach is the selection of offspring gender by pre- conception gender selection (PGS).

Gender selection through prenatal diagnosis and abortion has existed since the 1970s. More re- cently, preimplantation sexing of embryos for transfer has been developed (Tarin and Handyside 1993; The Ethics Committee of the American So- ciety of Reproductive Medicine 1999). Yet prena- tal or preimplantation methods of gender selection are unattractive because they require abortion or a costly, intrusive cycle of in vitro fertilization (IVF) and embryo discard. Attempts to separate X- and Y-bearing sperm for preconception gender selec- tion by sperm swim-up or swim-through tech- niques have not shown consistent X- and Y-sperm cell separation or success in producing offspring of the desired gender.

The use of �ow cytometry to separate X- and Y- bearing sperm may turn out to be a much more re- liable method of enriching sperm populations for insemination. Laser beams passed across a �owing array of specially dyed sperm can separate most of the 2.8% heavier X- from Y-bearing sperm, thus producing an X-enriched sperm sample for insemi- nation.2 Flow cytometry has been used successfully in over 400 sex selections in rabbit, swine, ovine, and bovine species, including successive genera- tions in swine and rabbit (Fugger et al. 1998). A human pregnancy was reported in 1995 (Levinson, Keyvanfar, and Wu 1995).

The United States Department of Agriculture (USDA), which holds a patent on the �ow cy- tometry separation process, has licensed the Genet- ics and IVF Institute in Fairfax, Virginia, to study the safety and ef�cacy of the technique for medical and “family balancing” reasons in an institutional review board–approved clinical trial.3 In 1998 re- searchers at the Institute reported a 92.9% success rate for selection of females in 27 patients, with most fertilizations occurring after intrauterine in- semination (Fugger et al. 1998). A lower success rate (72%) was reported for male selection. 4

At this early stage of development much more research is needed to establish the high degree of safety and ef�cacy of �ow cytometry methods of PGS that would justify widespread use. With only one published study of outcomes to date, it is too soon to say whether the 92% success rate in deter- mining female gender will hold for other patients, much less that male selection will reach that level of ef�cacy. Animal safety data have shown no ad- verse effect of the dye or laser used in the technique on offspring, but that is no substitute for more ex- tensive human studies (Vidal et al. 1999). In addi- tion, if �ow cytometry instruments are to be used for sperm separation purposes, they may be class- i�ed as medical devices that require U.S. Food and Drug Administration (FDA) approval. Finally, the holder of the process patent—the USDA—will have to agree to license the process for human uses.

If further research establishes that �ow cy- tometry is a safe and effective technique for both male and female PGS, and regulatory and licensing barriers are overcome, then a couple wishing to choose the gender of their child would need only provide a sperm sample and undergo one or more cycles of intrauterine insemination with separated sperm.5 A clinic or physician that offers assisted re- productive technologies (ART) and invests in the

Target Article


gender discrimination


procreative liberty

sex selection

Open Peer Commentary

Norman Daniels, p. 10

Rebecca Dresser, p. 11

Carson Strong, p. 12

Scot D. Yoder, p. 14

Mary B. Mahowald, p. 15

Julian Savulescu, p. 16

Owen D. Jones, p. 19

David B. Resnik, p. 21

Judith Daar, p. 23

Sandra Anderson Garcia, p. 24

Diane Paul, p. 26

Carl H. Coleman, p. 27

Mark V. Sauer, p. 28

Nancy King Reame, p. 29

Rosamond Rhodes, p. 31

Gregory Stock, p. 33

William Ruddick, p. 34

Dorothy C. Wertz, p. 36

Jenny Dai, p. 37

John Oberdiek, p. 38

�ow cytometry equipment could run the separa- tion and prepare the X- or Y-enriched sperm for in- semination, or it could have the sperm processed by a clinic or �rm that has made that investment. Flow cytometry separation would not be as cheap and easy as determining gender by taking a pill be- fore intercourse, but it would be within reach of most couples who have gender preferences in off- spring.6

Demand for Preconception Gender Selection

Unkown at present is the number of people who have offspring gender preferences robust enough to incur the costs and inconvenience of PGS. Al- though polls have often shown a preference for �rstborn males, they have not shown that a large number of couples would be willing to forego coital conception in order to select the gender of their children. If PGS proves to be safe and effec- tive, however, it may be sought by two groups of persons with gender preferences.

One group would seek PGS in order to have a child of a gender different from that of a previous child or children. A preference for gender variety in offspring would be strongest in families that have already had several children of one gender. They may want an additional child only if they can be sure that it will be of the gender opposite to their existing children.7 Couples who wish to have only two children might use PGS for the second child to ensure that they have one child of each gender. If social preferences for two-child families remain strong, some families may use PGS to choose the gender of the second child.

A second group of PGS users would be those persons who have strong preferences for the gender of the �rst child. The most likely candidates here are persons with strong religious or cultural beliefs about the role or importance of children with a particular gender. Some Asian cultures have belief systems that strongly prefer that the �rstborn child be a male. In some cases the preference re�ects reli- gious beliefs or traditions that require a �rstborn son to perform funeral rituals to assure his parents’ entrance into heaven (for a discussion of son prefer- ences in India and China, see Macklin 1999, 148– 151). In others it simply re�ects a deeply embed- ded social preference for males over females. The �rst-child preference will be all the stronger if a one-child-per-family policy is in effect, as occurred for a while in China (Greenlagh and Li 1995, 627). While the demand for PGS for �rstborn children is likely to be strongest in those countries, there has

been a sizable migration of those groups to the United States, Canada, and Europe.8 Until they are more fully assimilated, immigrant groups in West- ern countries may retain the same gender prefer- ences that they would have held in their home- lands.

Other persons with strong gender preferences for �rstborn children would be those who prize the different rearing or relational experiences they think they would have with children of a particular gender. They may place special value on having their �rstborn be male or female because of per- sonal experiences or beliefs. Numerous scenarios are likely here, from the father who very much wants a son because of a desire to provide his child with what he lacked growing up, to the woman who wants a girl because of the special closeness that she thinks she will have with a daughter (Belkin 1999).

The Ethical Dilemma of Preconception Gender Selection

The prospect of preconception gender selection ap- pears to pose the con�ict—long present in other bioethical issues—between individual desires and the larger common good. Acceding to individual desires about the makeup of children seems to be required by individual autonomy. Yet doing so leads to the risk that children will be treated as ve- hicles of parental satisfaction rather than as ends in themselves, and could accelerate the trend toward negative and even positive selection of offspring characteristics. The dilemma of reconciling procre- ative liberty with the welfare of offspring and fami- lies will only intensify as genetic technology is fur- ther integrated with assisted reproduction and couples seek greater control over the genes of off- spring.

Arguments for Preconception Gender Selection

The strongest argument for preconception gender selection is that it serves the needs of couples who have strong preferences about the gender of their offspring and would not reproduce unless they could realize those preferences. Because of the im- portance of reproduction in an individual ’s life, the freedom to make reproductive decisions has long been recognized as a fundamental moral and legal right that should not be denied to a person unless exercise of that right would cause signi�cant harm to others (Robertson 1994, 22–42). A corollary of this right, which is now re�ected in carrier and prenatal screening practices to prevent the birth of

Winter 2001, Volume 1, Number 1 ajob 3

Preconception Gender Selection

See http://ajobonline:

More online commentary and author response s

Relevant news

AJOBlive discussion

Write your own response to the contributors

children with genetic disease, is that prospective parents have the right to obtain preconception or prenatal information about the genetic characteris- tics of offspring, so that they may decide in a par- ticular case whether or not to reproduce (Robertson 1996, 424–435).9

Although offspring gender is not a genetic dis- ease, a couple’s willingness to reproduce might well depend on the gender of expected offspring. Some couples with one or more children of a par- ticular gender might refuse to reproduce if they cannot use PGS to provide gender variety in their offspring or to have additional children of the same gender (E. F. Fugger, personal communication to author). In other cases they might have such strong rearing preferences for their �rstborn child that they might choose not to reproduce at all if they cannot choose that child’s gender. Few persons con- templating reproduction may fall into either group; but for persons who strongly hold those preferences, the ability to choose gender may de- termine whether they reproduce.

In cases where the gender of offspring is essen- tial to a couple’s decision to reproduce, the freedom to choose offspring gender would arguably be part of their procreative liberty (Robertson 1996, 434). Since respect for a right is not dependen t on the number of persons asserting the right, they should be free to use a technique essential to their repro- ductive decision unless the technique would cause the serious harm to others that overcomes the strong presumption that exists against government interference in reproductive choice. Until there is a substantial basis for thinking that a particular use of PGS would cause such harms, couples should be free to use the technique in constituting their fam- ilies. The right they claim is a right against gov- ernment restriction or prohibition of PGS. It is not a claim that society or insurers are obligated to fund PGS or that particular physicians must pro- vide it.

Arguments Against Preconception Gender Selection

There are several arguments against preconception gender selection. Although such methods do not harm embryos and fetuses or intrude on a woman’s body as prenatal gender selection does, they do raise other important issues. One concern is the poten- tial of such techniques to increase or reinforce sex- ism, either by allowing more males to be produced as �rst or later children, or by paying greater atten- tion to gender itself. A second concern is the wel- fare of children born as a result of PGS whose par-

ents may expect them to act in certain gender spe- ci�c ways when the technique succeeds, but who may be disappointed if the technique fails. A third concern is societal. Widely practiced, PGS could lead to sex-ratio imbalances, as have occurred in some parts of India and China due to female infan- ticide, gender-driven abortions, and a one-child- per-family policy (Sen 1990). Finally, the spread of PGS would be another incremental step in the growing technologization of reproduction and ge- netic control of offspring. While each step alone may appear to be justi�ed, together they could constitute a threat to the values of care and concern that have traditionally informed norms of parent- ing and the rearing of children.

Evaluation of Ethical and Social Issues

Concerns about sex-ratio imbalances, welfare of offspring, and technologizing reproduction may be less central to debates over PGS than whether such practices would be sexist or contribute to sexism. If the number of persons choosing PGS is small, or the technique is used solely for offspring gender variety, sex-ratio imbalances should not be a prob- lem. If use patterns did produce drastic changes in sex ratios, self-correcting or regulatory mecha- nisms might come into play. For example, an over- abundanc e of males would mean fewer females to marry, which would make being male less desir- able, and provide incentives to increase the number of female births. Alternatively, laws or policies that required providers of PGS to select for males and females in equal numbers would prevent such im- balances.10 A serious threat of a sex-ratio imbalance would surely constitute the compelling harm nec- essary to justify limits on reproductive choice.

It may also be dif�cult to show that children born after PGS were harmed by use of the tech- nique. Parents who use PGS may indeed have spe- ci�c gender role expectations of their children, but so will parents who have a child of a preferred gen- der through coitus. Children born with the desired gender after PGS will presumably be wanted and loved by the parents who sought this technique. Parents who choose PGS should be informed of the risk that the technique will not succeed, and coun- seled about what steps they will take if a child of the undesired gender is born.11 If they commit themselves in advance to the well-being of the child, whatever its gender, the risk to children should be slight. However, it is possible that some couples will abort if the fetus is of the undesired

4 ajob Winter 2001, Volume 1, Number 1

The American Journal of Bioethics

gender. PGS might thus inadvertently increase the number of gender-selection abortions.

Finally, technological assistance in reproduc- tion is now so prevalent and entrenched that a ban on PGS would probably have little effect on the use of genetic and reproductive technologies in other situations. With some form of prenatal screening of fetuses occurring in over 80% of United States pregnancies, genetic selection by negative exclu- sion is already well-installed in contemporary re- productive practice. Although there are valid con- cerns about whether positive forms of selection, including nonmedical genetic alteration of off- spring genes, should also occur, drawing the line at all uses of PGS will not stop the larger social and technological forces that lead parents to use ge- netic knowledge to have healthy, wanted offspring. If a particular technique can be justi�ed on its own terms, it should not be barred because of specula- tion of a slippery slope toward genetic engineering of offspring traits (for an analysis of the slippery- slope problem with genetic selection, see Robert- son 1994, 162–165).

Is Gender Selection Inherently Sexist?

A central ethical concern with PGS is the effect of such practices on women, who in most societies have been subject to disadvantage and discrimina- tion because of their gender. Some ethicists have argued that any attention to gender, male or fe- male, is per se sexist, and should be discouraged, regardless of whether one can show actual harmful consequences for women (see Grubb and Walsh 1994; and Wertz and Fletcher 1989). Others have argued that there are real differences between male and female children that affect parental rearing ex- periences and thus legitimate nonsexist reasons for some couples to prefer to rear a girl rather than a boy or vice versa, either as a single child or after they have had a child of the opposite gender.

To assess whether PGS is sexist we must �rst be clear about what we mean by sexism. The Compact OED (1991, 1727) de�nes sexism as “the assump- tion that one sex is superior to the other and the re- sultant discrimination practised against members of the supposed inferior sex, especially by men against women.” By this de�nition, sexism is wrong because it denies the essential moral, legal, and political equality between men and women. Under this de�nition, if a practice is not motivated by judgments or evaluations that one gender is su- perior to the other, or does not lead to discrimina- tion against one gender, it is not sexist.

Professor Mary Mahowald, an American bio-

ethicist writing from an egalitarian feminist perspective, makes the same point with a conse- quentialist twist:

Selection of either males or females is justi�able on medical grounds and morally defensible in other situa- tions [emphasis added] so long as the intention and the consequences of the practice are not sexist. Sexist intentions are those based on the notion that one sex is inferior to the other; sexist consequences are those that disadvantage or advantage one sex vis-à-vis the other. (2000, 121)

In my view, the OED de�nition, modi�ed by Mahowald’s attention to consequences, is a persua- sive account of the concept of sexism. If that ac- count is correct, then not all attention to the bio- logic, social, cultural, or psychological differences between the sexes would necessarily be sexist or disadvantage females. That is, one could recognize that males and females have different experiences and identities because of their gender, and have a preference for rearing a child of one gender over an- other, without disadvantaging the dispreferred gender or denying it the equal rights, opportuni- ties, or value as a person that constitutes sexism.

If this conjecture is correct, it would follow that some uses of PGS would clearly be sexist, while others would clearly not be. It would be sex- ist to use PGS to produce males because of a paren- tal belief that males are superior to females. It would be nonsexist to use PGS to produce a girl because of a parental recognition that the experi- ence of having and rearing a girl will be different than having a boy. In the latter case, PGS would not rest on a notion of the greater superiority of one gender over another, nor, if it occurred in coun- tries that legally recognize the equal rights of women, would it likely contribute to sexism or further disadvantage women. As Christine Overall, a British feminist bioethicist, has put it, “sexual similarity or sexual complementarity are morally acceptable reasons for wanting a child of a certain sex” (1987, 27; quoted in Mahowald 2000, 117).

Psychological research seems to support this position. It has long been established that there are differences between boys and girls in a variety of domains, such as (but not limited to) aggression, activity, toy preference, psychopathology, and spa- tial ability (Maccoby and Jacklin 1974; Gilligan 1980; Kimura and Hampson 1994; Feingold 1994; Collaer and Hines 1995; and Halpern 1997). Whether these differences are primarily in- born or learned, they are facts that might rationally lead people to prefer rearing a child of one gender

Winter 2001, Volume 1, Number 1 ajob 5

Preconception Gender Selection

6 ajob Winter 2001, Volume 1, Number 1

The American Journal of Bioethics

rather than another, particularly if one has already had one or more children of a particular gender. In- deed, Supreme Court Justice Ruth Bader Gins- burg, a noted activist for women’s rights before her appointment to the Supreme Court, in her opinion striking down a male-only admissions policy at the Virginia Military Institute (United States v. Vir- ginia, 116 S. Ct. 2264 [1996]), noted that:

Physical differences between men and women . . . are enduring: “[T]he two sexes are not fungible; a community made up exclusively of one [sex] is dif- ferent from a community composed of both.” . . . “Inherent differences” between men and women, we have come to appreciate, remain cause for celebra- tion.

Some persons will strongly disagree with this account of sexism and argue that any attention to gender difference is inherently sexist because per- ceptions of gender difference are themselves rooted in sexist stereotypes. They would argue that any offspring gender preference is necessarily sexist be- cause it values gender difference and thus rein- forces sexism by accepting the gendered stereo- types that have systematically harmed women (Grubb and Walsh 1994; and Wertz and Fletcher 1989, 2112). According to them, a couple with three boys who use PGS to have a girl are likely to be acting on the basis of deeply engrained stereo- types that harm women. Similarly, a couple’s wish to have only a girl might contribute to unjusti- �ed gender discrimination against both men and women, even if the couple especially valued females and would insist that their daughter re- ceive every bene�t and opportunity accorded males.

Resolution of this controversy depends ulti- mately on one’s view of what constitutes sexism and what actions are likely to harm women. Al- though any recognition of gender difference must be treated cautiously, I submit that recognizing and preferring one type of childrearing experience over the other can occur without disadvantaging women generally or denying them equal rights and respect. On this view, sexism arises not from the recognition or acceptance of difference, but from unjusti�ed reactions to it. Given the biological and psychological differences between male and female children, parents with a child of one gender might without being sexist prefer that their next child be of the opposite gender. Similarly, some parents might also prefer that their �rstborn or only child be of a particular gender because they desire a spe- ci�c rearing and companionship experience.

If it is correct that using PGS for offspring di- versity is sexist, then those who deny that biologi- cal gender differences exist, or who assume that any recognition of them always reinforces sexism or disadvantages women, will not have carried the burden of showing that a couple’s use of PGS for offspring gender variety or other nonintentionally sexist uses is so harmful to women that it justi�es restricting procreative choice. Until a clearer ethi- cal argument emerges, or there is stronger empiri- cal evidence that most choices to select the gender of offspring would be harmful, policies to prohibit or condemn as unethical all uses of nonmedically indicated PGS would not be justi�ed.

The matter is further complicated by the need to respect a woman’s autonomy in determining whether a practice is sexist. If a woman is freely choosing to engage in gender selection, even gen- der-selection abortion, she is exercising procreative autonomy. One might argue in response that the woman choosing PGS or abortion for gender selec- tion is not freely choosing if her actions are in�uenced by strong cultural mores that prefer males over females. Others, however, would argue that the straighter path to equal rights is to respect female reproductive autonomy whenever it is exer- cised, even if particular exercises of autonomy are strongly in�uenced by the sexist norms of her com- munity (Mahowald 2000, 188).

Public Policy and Preconception Gender Selection

Because of the newness of PGS and uncertainties about its effects, the best societal approach would, of course, be to proceed slowly, �rst requiring ex- tensive studies of safety and ef�cacy, and then at �rst only permitting PGS for increasing the gender variety of offspring in particular families.13 Only after the demographic and other effects of PGS for gender variety have been found acceptable should PGS be available for �rstborn children.

However, given the close connection between parental gender preferences for offspring and re- productive choice, public policies that bar all nonmedical uses of PGS or that restrict it to choos- ing gender variety in offspring alone could be found unconstitutional or illegal. If there are phys- ical, social, and cultural differences between girls and boys that affect the rearing or relational experi- ences of parents, individuals and couples would have the right to implement those preferences as part of their fundamental procreative liberty. The risk that exercising rights of procreative liberty would hurt offspring or women—or contribute to

sexism generally—is too speculative and uncertain to justify infringement of those rights.

The claim of a right to choose offspring gender is clearest in the case of PGS for gender variety. If �ow cytometry or other methods of PGS are found to be safe and effective, there would be no compel- ling reason to ban or restrict their nonmedical use by persons seeking gender variety in the children they rear. Couples with one child or several chil- dren of a particular gender might, without being sexist or disadvantaging a particular gender, prefer to have an additional child of the opposite gender. ART clinics should be free to proceed with PGS for offspring variety in cases where couples are aware of the risk of failure, and have undergone counsel- ing that indicates that they will accept and love children of the dispreferred gender if PGS fails. Clinics providing PGS should also ask couples to participate in research to track and assess the ef- fects of PGS on children and families.

The use of PGS to determine the gender of �rstborn children is a more complicated question. The choice to have one’s �rst or only child be fe- male has the least risk of being sexist, because it is privileging or giving �rst place to females, who have traditionally been disfavored.14 The use of PGS to select �rstborn males is more problematic because of the greater risk that this choice re�ects sexist notions that males are more highly valued. It is also more likely to entrench male dominance. The danger of sexism is probably highest in those ethnic communities that place a high premium on male offspring, but it could exist independentl y of those settings.

Yet restricting PGS to offspring gender variety and �rstborn females may be dif�cult to justify. Given that individuals could prefer to have a boy rather than a girl because of the relational and rear- ing experiences he will provide, just as they might prefer a girl for those reasons, it might be dif�cult to show that all preferences for �rstborn males are sexist. Nor could one easily distinguish �rstborn male preferences when the couple demanding them is of a particular ethnic origin. Although the risk that �rstborn male preferences would be sexist is greatest if the PGS occurred in a country in which those beliefs prevailed, the chance that PGS would contribute to societal sexism lessens greatly if the child is reared in a country that legally pro- tects the equal status of women and men.

If prohibitions on some or all nonmedical uses of PGS could not be justi�ed and might even be unconstitutional, regulation would have to take different forms. One form would be to deny public

or private insurance funding of PGS procedures, which would mean that only those willing to pay out-of-pocket would utilize them. Another form would be for the physicians who control access to PGS techniques to take steps to assure that it is used wisely. If they comply with laws banning dis- crimination, physician organizations or ART clin- ics could set guidelines concerning access to PGS. They might, for example, limit its use to offspring gender variety or �rstborn female preferences only. As a condition of providing services, they might also require that any couple or individual seeking PGS receive counseling about the risks of failure and …