ScienceDirect

E-mail addresses: [e-mail] (T.D. Steensma), [e-mail] (P.T. Cohen-Kettenis).
In recent years, adolescents who experience gender incongruence with their birth-assigned gender received much clinical and media attention. A sharp increase in the number of referrals to gender identity clinics and a decline in age at which medical interventions, aiming at gender reassignment, are requested, are observed in Europe (de Vries and Cohen-Kettenis, 2012), as well as in Northern America (Wood

0018-506X/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.yhbeh.2013.02.020

et al., 2013). One of the reasons might be the availability of puberty sup- pression as an aid to provide adolescents with gender incongruence with time and without the accompanying distress caused by the physi- cal changes of puberty before a more definite decision regarding gender reassignment is made (e.g. Olson et al., 2011). This increase in attention raises questions regarding what knowledge we have on gender identity development in adolescence and what factors are of theoretical and clinical relevance during this critical developmental phase. This article provides a historical overview of the terminology and describes what we know about developmental pathways and contributing factors.

The concept of gender identity

The term identity comes from the Latin noun “identitas,” which means the same. The term, referring to a person's mental image of him or herself thus implies some sameness with others in a particular way. Each individual may have a number of identities, such as an eth- nic identity, a religious identity, or a national identity (Kroger, 2007). A very fundamental identity, however, is one's gender identity. Gender identity refers to the extent to which a person experiences oneself to be like others of one gender. One's sense of being male or female largely determines how people view themselves and provides an important basis for their interactions with others.

Over the years the terms gender identity and also gender role (behav- iors, attitudes, and personality traits which, within a given society and historical period, are typically attributed to, expected from, or preferred by persons of one gender) have been used in different ways. In the 1950s, the terms were introduced in the clinical literature when psychol- ogists working with individuals with disorders of sex development (DSD; previously called intersex conditions), and with gender dysphoria started to study gender identity development.
In most cases, gender identity will develop in accordance with phys- ical gender characteristics. A baby with XY sex chromosomes and male genitalia will generally be assigned to the male gender, will show male typical behaviors, and have a male gender identity. Discordance between these gender aspects does occur, however, in some conditions. DSD are congenital conditions in which the development of chromosomal, go- nadal, or anatomical sex is atypical (Hughes et al., 2006). For instance, in DSD, external male appearing genitalia may not correspond with the gonads, and/or sex chromosomes. Gender identity may be in line with the chromosomes and gonads, but not with the external genitalia. Gender dysphoria refers to the distress resulting from incongruence between experienced/expressed gender and assigned gender. In gender dysphoric individuals, a gender identity may develop that does not match with sex chromosomes, gonads and genitalia, although the physical sex character- istics all correspond with each other.
Albert Ellis was one of the first to report on gender identity and sexual orientation variations in adults with DSD (Ellis, 1945). A decade later, this line of research was continued and elaborated by John Money, a modern sexologist who worked with children with DSD (Money, 1994). He proposed to make a clear distinction between the terms sex and gender, because, particularly in the field of DSD, sex is a confusing concept. For instance, does a 46,XY person with a complete androgen insensitivity syndrome (CAIS), characterized by high testosterone levels, undescended testes, and a vulva belong to the male or female sex? Money also introduced the dual concept of gender identi- ty/role (GI/R). He considered gender identity to be the private manifesta- tion of gender role, and gender role the public manifestation of gender identity. However, in gender dysphoric persons, the gender role, which, according to Money is the public expression of one's gender identity, is at least for some period, seriously blocked. Their gender identity, but not their gender role, may thus be different from their assigned gender. For this reason, and because in research the concepts are often dealt indepen- dently, gender identity and gender role are currently used separately.
Some decades ago, Stoller (1968) introduced the concept of core gen- der identity. He considered it the “…essentially unalterable core of gender
identity (e.g., I am male) to be distinguished from the related but differ- ent belief, I am manly (or masculine)….” (p 40) and speaks of an “inner conviction that the sex of assignment was right.” (Stoller, 1985, p.11)
Gender identity has not only been investigated in clinical research. Cognitive developmental psychologists also made use of the concept. For a few decades, they mainly focused on cognitive components of gender identity (Fagot and Leinbach, 1985; Kohlberg, 1966; Ruble and Martin, 1998). For instance, Kohlberg (1966, p. 88) defined gender identity as the “cognitive self-categorization as boy or girl” and Fagot and Leinbach (1985, p. 685) considered gender identity to be “the con- cept of the self as male or female.” More recently, researchers in this field gave more attention to affective components of gender identity, such as feelings of contentment with one's gender (Egan and Perry, 2001), and they started studying its relationship with mental health. They also considered felt pressure for gender conformity and felt com- patibility aspects of gender identity. Tobin et al. (2010) proposed a five dimensional model, subdivided into membership knowledge of a gender category, gender centrality (the importance of gender to other identities), gender contentedness, felt gender conformity, and felt gen- der typicality, to conceptualize gender identity. In their conceptualiza- tion of gender identity, the recent cognitive developmental researchers are much closer to clinical theorists than their predecessors.
In clinical psychology and psychiatry, individuals who do not identify with their assigned gender, became known as transsexuals (WHO, 1992) or individuals with a gender identity disorder (GID; APA, 2000). Clinically they were, and still are, categorized according to criteria as formulated by the APA and WHO. If they fulfill the criteria for the diagnosis and are able to live in the preferred gender for a period of time, supported by pre- scribed cross-sex hormones, and are capable to handle the complex issues surrounding treatment, they are referred for gender reassignment surgery. In DSM-IV-TR's accompanying text, terms such as “the other sex” are frequently used, and within the DSM GID criteria, the term “cross-gender identification” also suggests that there are only two gender identity categories, male and female. For long, gender identity, gender role, and gender problems were conceptualized dichotomously rather than dimensionally.
During the last decade, the dimensionality and diversity of gender identity and gender problems have received increasing attention and criticism in the literature (e.g. Fausto-Sterling, 2000). It is argued that individuals who experience gender problems do not necessarily experi- ence a complete cross-gender identity and do not always need clinical attention (e.g. Diamond and Butterworth, 2008; Lee, 2001). Bockting (2008) showed that the gender identification of individuals covers a wide spectrum of gender identity labels, such as; “shemale,” “third gen- der,” “pan-/poly-/or omnigendered,” “gender fluid,” instead of male and female or even transsexual. These individuals may or may not experi- ence distress and they may or may not want to live as “the other gender” (see Cohen-Kettenis and Pfäfflin, 2010, for an overview). Regarding treatment, some only want parts of the classical gender reassignment, consisting of hormone treatment and gender reassignment surgeries. For example, in a specific condition, men desire to obtain chemical or surgical castration (in some with additional penectomy) without the desire to transition to the female gender, because they do not identify as females but as eunuchs. They are referred to as Male-to-Eunuch indi- viduals (e.g. Johnson and Wassersug, 2010; Wassersug et al., 2004).
As an umbrella term for aspects of gender that are gender non-conforming or non-normative the term gender variant is often used. With regard to gender identity development, a normative (or conforming) gender identity development and a variant (or non- normative) gender identity development can be distinguished.

Adolescent identity development

Developmental psychologists like Erikson (1968) and Marcia (1966), Marcia et al. (1993), have demonstrated that adolescence serves as an important period for the formation of a personal identity. A

personal identity includes values, principles and roles an individual has adopted as his or her own. Identity formation is an individual process in which adolescents explore and commit to identity-defining roles and values in a variety of life domains (politics, occupation, religion, intimate relationships, friendships, and gender roles). The variation in styles through which this process evolves will lead to differences in identity development and identity outcome. For example; in early ado- lescence commitment to a domain may be made without any prior explorations (this is called foreclosure), often based on parental values, or commitment may not be formed at all because of disinterest in find- ing personally expressive adult roles and values (called diffusion). Later in adolescence, commitment may (still) not yet be formed but the adolescent searches for meaningful adult roles and values (called moratorium), eventually followed by a style where commitment is based on thoughtful exploration (called identity-achieved) (Kroger, 2008; Marcia et al., 1993).
More recently, the focus on identity development has been expanded and directed at the role of context (e.g. Adams and Marshall, 1996; Yoder, 2000). Here, identity development is presented as an individual as well as a social process at which identity shapes and is shaped by the surrounding milieu (Adams and Marshall, 1996). Also, the role of gender and possible gender differences regarding identity structure, the importance of identity domains and the process of identity formation have been examined (Kroger, 1997). Kroger (1997) reviewed the litera- ture up to 1995 reporting on these topics and concluded that there was little evidence for gender differences. However, instead of gender differ- ences, empirical evidence suggests a potential role of gender-role orien- tation (masculine, feminine, androgynous) on the identity-formation during adolescence (e.g., Bartle-Haring and Strimple, 1996; Sochting et al., 1994). Whether gender-role orientation may affect gender identity- formation, lead to actual gender identity fluctuations, gender identity changes, or result in gender role experimenting alone, without an influ- ence on gender identity is, however, currently unclear.

Gender identity development

Cognitive developmental researchers studying gender identity development have almost exclusively focused on the role of cognitive factors in young children (Ruble et al., 2006). They found that gender learning starts early, is a gradual process taking many years, and passes through various stages (Kohlberg, 1966). Most children devel- op the ability to label their own and others' gender between 18 and 24 months. This ability is related to increased gender typed prefer- ences such as the preference for stereotyped toys (e.g. boys preferring trucks and girls preferring dolls) (e.g. Serbin et al., 2001; Zosuls et al., 2009), the preference for certain play behaviors (rough-and-tumble play in boys, cooperative play in girls) (e.g. Ruble and Martin, 1998), and the gradual increase in the preference for same-sex play- mates (e.g. Lobel et al., 2000). There are indications for gender differ- ences at the end of childhood with boys having a stronger gender identity than girls, reporting to be more content with their gender, viewing themselves as more gender typical than other boys, and plac- ing more pressure on themselves to conform to the expected gender role, than girls (Egan and Perry, 2001). For most children, gender identity is largely congruent with their gender role behaviors.

Hill and Lynch (1983) proposed that in adolescence, gender inten-