In its recent marriage decisions, the Supreme Court wisely declined to hold—as did the lower court in one of the two cases—that “sexual orientation” was a constitutionally “suspect” class, comparable to classes of race, color, and ethnicity. Any legal rule that classifies people on a “suspect” basis is usually presumed to be invidiously discriminatory and thus unconstitutional.

Not surprisingly, a new law introduced in both houses of Congress would go where the Supreme Court chose not to tread. The Employment Non-Discrimination Act (ENDA) would prohibit any adverse treatment in matters of employment based on an individual’s “actual or perceived sexual orientation or gender identity.”

Enacting ENDA would be a grave mistake, and a disservice to the cause of equality.

ENDA defines sexual orientation as “homosexuality, heterosexuality, or bisexuality,” and gender identity as “the gender-related identity, appearance, or mannerisms or other gender-related characteristics of an individual, with or without regard to the individual’s designated sex at birth.”

Start your day with Public Discourse

Sign up and get our daily essays sent straight to your inbox.

But social science research continues to show that sexual orientation, unlike race, color, and ethnicity, is neither a clearly defined concept nor an immutable characteristic of human beings. Basing federal employment law on a vaguely defined concept such as sexual orientation, especially when our courts have a wise precedent of limiting suspect classes to groups that have a clearly-defined shared characteristic, would undoubtedly cause problems for many well-meaning employers.

The Courts’ Precedents on Suspect Classes

Ten federal circuits have rejected the claim that sexual orientation defines a “suspect class” entitled to “heightened scrutiny” under the Fourteenth Amendment’s Equal Protection Clause. Only the Second Circuit, in its decision against the Defense of Marriage Act, held that “homosexuals compose a class that is subject to heightened scrutiny.” But the Supreme Court, reviewing the Second Circuit’s decision, instead based its ruling on the prior granting of marital status to same-sex couples in some states, and argued that it was unconstitutional for federal law to withhold recognition of that status for federal purposes.

The Court applies heightened scrutiny to protect “discrete and insular” groups, which are said to need “extraordinary protection from the majoritarian political process.” But the Court has repeatedly declined to apply heightened scrutiny where discreteness—the capability of being clearly defined—is lacking.

When the Court heard cases involving claims of discrimination against old age or poverty, for instance, it considered but rejected the argument that the elderly or the poor are clearly defined groups that need special legal protection. Perhaps the Court declined to create a new suspect class for sexual orientation in the Windsor case because sexual orientation is an even less discrete characteristic than age or poverty.

Sexual Orientation Is Not Biologically Determined

What does the science say about sexual orientation? While some people may display signs of same-sex attraction at an early age, researchers still don’t agree that sexual orientation is determined solely or even primarily at birth, as is one’s race or ethnicity. Some researchers have even concluded that biological and genetic factors play little to no role in sexual orientation.

Their conclusions are supported, for example, by studies of identical twins. Professors at Columbia University also reported in a 2002 study that “among male [opposite-sex] twins, the proportion reporting a same-sex romantic attraction is twice as high among those without older brothers (18.7%) than among those with older brothers (8.8%).”

Other studies have found strong correlations between sexual orientation and external factors, such as family setting, environment, and social conditions, which are difficult if not impossible to explain under exclusively biological theories.

In 2003, researchers in Australia discovered “a major cohort effect in same-gender sexual behavior” and noted that this had “implications for purely biological theories of sexual orientation, because there must be historical changes in environmental factors that account for such an effect.”

The well-known “Chicago Sex Survey” found that men were twice as likely and women nine times as likely to identify as gay, lesbian, or bisexual if they had completed college.

Researchers in New Zealand noted a correlation between sexual orientation and certain social conditions, reporting that:

The overall higher rate of same-sex attraction and contact for women in New Zealand in relation to other comparable countries, almost certainly represents a recent increase in prevalence. As such it argues strongly against a purely genetic explanation and suggests the environment can have a significant influence. It might be related to social changes which have happened with particular intensity and rapidity in this country.

In brief, available evidence casts serious doubt on the simplistic, popular notion that sexual orientation is biologically determined. There is simply no firm evidence to support that conclusion, only unproven theories. As the American Psychiatric Association’s latest statement on the issue summarizes: “Currently there is a renewed interest in searching for biological etiologies for homosexuality. However, to date there are no replicated scientific studies supporting any specific biological etiology for homosexuality.”

As the prominent scholars Richard C. Friedman and Jennifer I. Downey concluded in 2002, “the assertion that homosexuality is genetic is so reductionistic that it must be dismissed out of hand as a general principle of psychology.”

Sexual Orientation Can Change Over Time

There is significant evidence that sexual orientation is more plastic than commonly supposed. Changes in sexual orientation are difficult to measure simply because the concept itself is hard to define.

The American Psychological Association explains sexual orientation as “an enduring pattern of emotional, romantic, and/or sexual attractions to men, women, or both sexes. Sexual orientation also refers to a person’s sense of identity based on those attractions, related behaviors, and membership in a community of others who share those attractions” (emphasis added). The problem, though, is that many people who report having same-sex attractions are not consistent across these dimensions.

Researchers have found that three of these dimensions—attraction, behavior, and identity—are subject to change over time. Moreover, the presence of a large bisexual population is evidence that sexual orientation is, for some people and to some extent, fluid.

Research suggests that a person’s sexual orientation may be influenced by individual preference or choice. While this idea is highly controversial, some studies conclude that for some women self-identity as a lesbian is experienced as a personal choice rather than an immutable constraint. Indeed, a 10-year study of 79 non-heterosexual women, published by psychologist Lisa M. Diamond in 2008, reported that 67 percent changed their identity at least once, and 36 percent changed their identity more than once.

Some studies also show changes over time in the intensity of same-sex attraction. When asked to rate their attraction to members of the same sex on a scale, many individuals vary in their own estimation over time, with some becoming more “gay” and others becoming less “gay.” A 1997 study of the same-sex attraction of bisexual men reported that “homosexuality is not some monolithic construct one moves toward or from in a linear way; movement toward homosexuality fails to capture the fluid and contextual nature of sexuality.”

Scholarly surveys also show that a significant number of individuals in same-sex relationships previously have been married to someone of the opposite sex.

One might defend the immutability of sexual orientation by insisting that anyone whose sexual orientation changes over time is bisexual and that bisexuality is a discrete category of sexual orientation. But lumping everyone whose behavior does change into a broad residual category conveniently manipulates theory to blot out the primary evidence that the nature of sexuality is often mutable, not a fixed feature of someone’s personality and behavior.

If human sexual preference were generally fluid rather than fixed, one would expect that some individuals would fall at the tails of the bell curve where behavior, attraction, and identity remain exclusively homosexual or heterosexual, but that others would exhibit at least some variation along the sexual orientation continuum with respect to attraction, behavior, and identity. Research confirms this expectation.

A characteristic as mutable as sexuality is a poor candidate for inclusion in employment discrimination statutes, which are predicated on essential, durable characteristics of individuals bringing claims under federal law.


“Sexual orientation” should not be recognized as a newly protected characteristic of individuals under federal law. And neither should “gender identity” or any cognate concept.  In contrast with other characteristics, it is neither discrete nor immutable. There is no scientific consensus on how to define sexual orientation, and the various definitions proposed by experts produce substantially different groups of people.

Nor is there any convincing evidence that sexual orientation is biologically determined; rather, research tends to show that for some persons and perhaps for a great many, “sexual orientation” is plastic and fluid; that is, it changes over time. What we do know with certainty about sexual orientation is that it is affective and behavioral—a matter of desire and/or behavior.  And “gender identity” is even more fluid and erratic, so much so that in limited cases an individual could claim to “identify” with a different gender on successive days at work.  Employers should not be obliged by dint of civil and possibly criminal penalties to adjust their workplaces to suit felt needs such as these.

Despite the effort of ENDA’s legislative drafters to confine “sexual orientation” to homosexuality, heterosexuality, and bisexuality, the logic of self-defined “orientation” is not so easily cabined. When given the opportunity to do so by interviewers, many people who otherwise might be classified as either homosexual or lesbian or bisexual, instead describe themselves in less determinate terms such as: “mostly” (but not nearly entirely) attracted to members of the same (or opposite) sex. Even polyamory, “a preference for having multiple romantic relationships simultaneously,” has been defended as “a type of sexual orientation for purposes of anti-discrimination law” in a 2011 law review article.

And why not? If “sexual orientation” is—like “gender” and “sexuality”—a “social construct,” as many have urged us to view it, there is no stopping point in logic or justice where we can say that the law must not go in its protections. But ENDA is not the solution to this problem, whatever its precise dimensions might be.  It would instead lead to insurmountable enforcement difficulties, arbitrary and even whimsical results in many cases, and it would have an unjustified chilling effect upon all too many employers’ decisions.

No one who opposes ENDA need maintain that unjustified employment discrimination against persons on the basis of their mannerisms, pattern of sexual attraction, or the nature or objects of their sexual desires never happens. Indeed, everyone should resolve to oppose and to reduce such discrimination, which has all too often taken coarse and even gross forms in American experience.

We believe, too, that our common life would be more harmonious and less checkered by unjust discrimination if we reduced our focus—indeed, our current collective fixation—on matters of “sexual identity.”

This article is drawn from the amicus brief the authors filed with the US Supreme Court on marriage. That brief, with full citations to the studies discussed in this essay, can be found here.