Randall L. Sell earned his doctorate from the Harvard School of Public Health and is currently an Assistant Professor at the Joseph L. Mailman School of Public Health at Columbia.

Many different terms and definitions have been proposed over the last 130 years to describe the sexual orientation of subjects. One of earliest and most important sexual orientation classification schemes was proposed by [Karl Heinrich] Ulrichs in a series of pamphlets he published privately in the 1860s. Ulrichs's scheme, which was only intended to describe males, separated subjects into three basic categories: Dionings, Urnings, and Uranodionings. Arguably, these categories directly correspond with the scientific terms preferred today: heterosexual, homosexual and bisexual. Homosexual women, who were largely ignored by the early researchers, were referred to as Urningins and heterosexual women were referred to as Dioningins.

Xavier Mayne, a follower of Ulrichs, provided a definition of an urning in the first major work on homosexuality to be written by an American. He defined an urning as "a human being that is more or less perfectly, even distinctively, masculine in physique; often a virile type of fine intellectual, oral and aesthetic sensibilities: but who, through an inborn or later-developed preference feels sexual passion for the male human species. His sexual preference may quite exclude any desire for the female sex: or may exist concurrently with that instinct". Mayne's definition also encompasses male uranodionings by stating that desire for the female sex may exist concurrently. . .

Today the terms heterosexual (straight), homosexual (gay and lesbian), and bisexual are the most commonly used terms by researchers to describe sexual orientations. While not many other terms have been proposed to describe heterosexuality or bisexuality, a plethora of terms have been used by researchers to describe homosexuality, including uranianism, homogenic love, contrasexuality, homo-erotism, similsexualism, tribadism, sexual inversion, intersexuality, transexuality, third sex, and psychosexual hermaphroditism. Today's preferred terms and the term "sexual orientation" itself have a wide variety' of definitions in the literature but these generally comprise one or both of two components: a "psychological" component and a "behavioral" component. Not all definitions include both of these components, and as are discussed in detail below, definitions that include both components use either the conjunction "and" or "or" to join them.

Mayne's (1908) definition of the term Uming and Benkert's of the term homosexual (Robinson, 1936) only include a description of the psychological state. Mayne discussed how an individual's feelings of sexual passion determine their sexual orientation while Benkert talked of an "urge." Ellis, one of the most important writers on sexuality in late 19th and early 20th century England, also only talked of a psychological entity which he described as "sexual instinct." Ellis defined homosexuality as "sexual instinct turned by inborn constitutional abnormality toward persons of the same sex" (Ellis and Symonds, 1896). These definitions and other early ones generally omit any discussion of behavior (and in particular sexual behavior), except to say that the thought of it with the other sex is repulsive or horrifying to the homosexual. Krafft-Ebing, like his contemporaries, even makes the point to exclude behavior front the diagnosis of homosexuality. Krafft-Ebing (1886) stated that "the determining factor here is the demonstration of perverse feelings for the same sex; not the proof of sexual acts with the same sex. These two phenomena must not be confounded with each other."

More recent definitions often include both components. For example LeVay (1993) defined sexual orientation as "the direction of sexual feelings or behavior toward individuals of the opposite sex (heterosexuality), the same sex (homosexuality), or some combination of the two (bisexuality)," and Weinrich (1994) defined homosexuality "either (1) as a genital act or (2) as a long-term sexuoerotic status." Here the psychological states referred to are "sexual feelings" and "sexuoerotic status," and the behavioral outcome is "sexual behavior" as referred to by LeVay and a "genital act" as referred to by Weinrich. The psychological and behavioral components in both definitions are joined by "or" signifying that either one can be used to assess sexual orientation.

In A Descriptive Dictionary and Atlas of Sexology (Francoeur et al, 1991), homosexuality is broadly defined as "the occurrence or existence of sexual attraction, interest and genitally intimate activity between an individual and other members of the same gender." Here the psychological components are "sexual attraction" and "interest" and the behavioral outcome is described as "genitally intimate activity." Unlike the definitions of LeVay (1993) and Weinrich (1994), this definition joins the two components with the conjunction "and." Using the conjunction "and" makes it unclear as to whether both components are necessary for the assignment of sexual orientation classifications.

At the other extreme from the early definitions provided by Mayne and Benkert are definitions that only include discussions of the behavioral component. For example, Stedman's Medical Dictionary (1982) defined homosexuality as "sexual behavior, including sexual congress, between individuals of the same sex, especially past puberty." Here the psychological component does not seem to hold much if any importance for the assessment of sexual orientation. Beach (1950) is emphatic about only including sexual behavior in the definition of sexual orientation in his critique of the first English language translation of Gide's defense of homosexuality, Corydon. Beach (1930) states that "the term (homosexuality) means different things to different people . . . it is preferable to set forth the significance of the term as used in this discussion. Homosexuality refers exclusively to overt behavior between two individuals of the same sex. The behavior must be patently sexual, involving e! ! rotic arousal and, in most instances at least, resulting in the satisfaction of the sexual urge." According to Diamond (1993), it is this type of definition that is favored by researchers determining the size of the "homosexual" population in various countries. In the studies reviewed by Diamond, while all used some assessment of sexual behavior to determine the prevalence of sexual orientations, none used any assessment of a psychological state (such as sexual attraction).

Thus far I have discussed the two definitional components of sexual orientation as if the components themselves were uniform across definitions, but as is evident in the examples already provided, there are important variations. Psychological components of definitions may include the terms "sexual passion," "sexual urge," "sexual feelings," "sexual attraction," "sexual interest," "sexual arousal," "sexual desire," "affectional preference," "sexual instinct," "sexual orientation identity," and "sexual preference." Each of these terms may have a distinct meaning and not necessarily be indicative of the same phenomenon. That is, different terms in definitions may be describing slightly different phenomena despite the similar label for that phenomena.

Similarly, the behavioral component varies between definitions. Behavior can be stated simply as "sexual behavior" or it can be described, for example, as "genital activity," "sexual contact," or "sexual contact that achieves orgasm." Each one of these presents further challenges for researchers. That is, how do we define each of these terms within the definition itself and how would we operationalize them for measurement?

Obviously, definitions and preferred terms vary significantly from researcher to researcher and across time. We must concern ourselves with whether these definitions are describing the same phenomena and whether the measures of sexual orientation based upon these definitions do the same. . .

Researchers wanting to measure sexual orientation today have four basic choices of measurement tools. These are dichotomous measures, the Kinsey Scale, the Klein Scale, or the Shively and DeCecco Scale. None of these is completely satisfactory. First, dichotomous scales are unsatisfactory for the reasons outlined by Kinsey. In particular, as Kinsey et al. stated: "Not all things are black nor all things white. It is a fundamental of taxonomy that nature rarely deals with discrete categories." Second, the Kinsey Scale is unsatisfactory because it forces the artificial combination of psychological and behavioral components and perhaps incorrectly requires individuals to make tradeoffs between homosexuality and heterosexuality. Third, the Klein scale is unsatisfactory because the relative importance of each dimension in measuring sexual orientation has not been thoroughly investigated or grounded in theory, and like Kinsey, Klein required subjects to make trade-offs between heterosexuality and homosexuality on his scale. Finally, the Shively and DeCecco scale is unsatisfactory because its properties have not been thoroughly investigated and its' consideration of physical and affectional preference may be oversimplified or even inappropriate.

	Excerpted with permission from "Defining and Measuring Sexual Orientation: A Review" by Randall L. Sell in Archives of Sexual Behavior Vol. 26, No 6 (Dec 1997), pp. 643-658. Footnotes omitted.