The perception that trans children are ‘too young’ to know their gender identity is used as a basis for denying them a suite of rights, and has long been a corner stone of arguments against social transition or timely provision of puberty blockers.
Cisgender (not trans) children are generally considered to know if they are a boy or a girl by a young age. But, not so for transgender or gender questioning children, according to the latest paper written by experts from the UK Children’s Gender Service, including the head of service Carmichael.
Costa, R. Carmichael, P.; Colizzi, M. (2016) To treat or not to treat: puberty suppression in childhood-onset gender dysphoria Nature Reviews Urology 13, 456–462 (2016)
UK Children’s Gender Service experts’ view of Gender Constancy
The Costa (2016) paper has a section on gender identity development. It states that “research has shown that around the age of 3 years, children show a basic sense of self as male or female, owing to their inner experience of belonging to one gender”.
It goes on to note:
“Some research suggests that a developmental lag exists in gender constancy acquisition in children with gender variant behaviour” (reference number 16).
According to the paper “achieving gender constancy represents a cognitive-developmental milestone in gender identity development and is due to the understanding that being male or female is a biological characteristic that cannot be changed by altering superficial attributes, such as hairstyle or clothing”
The belief in trans children having a ‘developmental lag’ in gender constancy leads to this statement “treating prepubertal individuals with gender dysphoria is particularly controversial owing to their unstable pattern of gender variance compared with gender-dysphoric adolescents and adults”
The belief in trans children having ‘a developmental lag” in gender constancy feeds directly into the Tavistock’s treatment protocols, such as proposing puberty suppression only be prescribed to those aged at least 12 “safely above the gender constancy achievement”.
Only one reference is provided for the claim that transgender children achieve ‘gender constancy’ later than cisgender children, reference 16, which is the source of this key statement:
“Some research suggests that a developmental lag exists in gender constancy acquisition in children with gender variant behaviour (reference 16)
If this single reference underpins the Tavistock’s belief that trans children do not understand their gender at the same age as cisgender children, and if this claim has direct implications on the Tavistock’s approach to treating trans children, then it is vital we review this paper.
The paper in question is:
Zucker, K. J. et al.(1999) Gender constancy judgments in children with gender identity disorder: evidence for a developmental lag. Arch. Sex. Behav. 28, 475–502 (1999).
Gender constancy in the Zucker paper is defined as “the understanding that ‘superficial’ or surface transformations in gender behaviour such as activity preferences or clothing style” do not change a person’s gender. The paper concludes that children referred to a Gender Clinic for ‘problems in identity development‘ have a ‘developmental lag in gender constancy‘. This conclusion merits further scrutiny.
Zucker et al.’s study focuses on a group of children who were referred to the Toronto Gender Clinic between 1978 and 1995.
The majority of the Gender Clinic children in this study were assigned males (207/236 = 88%). There were a small number of assigned females in the sample (12%). In order to simplify this blog post I have decided to focus the examples throughout on assigned males (noting that this editorial simplification perpetuates historical erasure of trans boys / assigned females).
The children registered at the Gender Clinic I will hereafter refer to as the ‘clinical sample’, to contrast with the study’s ‘control sample’ (a sample of children of the same age who were not registered at the gender clinic and were not known to have any gender issues).
It is known (and acknowledged in Zucker’s paper) that some of the clinical sample of assigned males were non-conforming boys rather than trans girls. How many were gender non-conforming (GNC) rather than trans is unknown as historical diagnoses focused on behaviour and interests more than on identity and Zucker did not believe in distinguishing between young gender non-conforming boys and trans girls.
The children in the clinical sample, together with a control group (aged 4-8 – average age 6 and a half) were put through three different types of test, which they either ‘passed’ or ‘failed’.
Zucker 1999, the tests
We will now look at the three tests, and see whether they do provide convincing evidence that transgender children (or children treated in the gender service) have a ‘developmental lag’, and understand their gender identity later than cisgender children.
Zucker 1999: Test 1: Slabey & Frey test
Test 1 Part A focused on Gender Discrimination
The children were shown dolls and photographs depicting a boy, girl, man, woman and asked to identify them. The children ‘passed’ if they got at least 12 out of 16 ‘correct’. 93% of the clinical sample ‘passed’ this test, compared to 98% of the control group.
Test 1 Part B: Gender Identity
The children were asked their own gender. The assigned-male-at-birth (amab) children ‘passed’ the test if they answered ‘boy’.
93% of the clinical sample ‘passed’ this test compared to 98% of the control group.
(The very high ‘pass’ rate for the clinical sample at first glance seems high as transgender children like my daughter would certainly ‘fail’ this test.
Perhaps the high ‘pass’ rate may add weight to suggestions that a large proportion of children referred to the Toronto gender clinic in the 1970s, 1980s and early 1990s were there for gender non-conformity (proto-gay cure….) rather than children with a gender identity different to their assigned sex.
The fact that a trans girl was considered to have ‘failed’ in her understanding of gender identity if she said she was a girl is an indication of the bias of the researchers.
Test 1 Part C: Gender Stability
The children were asked if their gender can change over time, for example if they were a different gender when they were born to their current gender. The children ‘passed’ if they said gender can never change over time.
80% of the clinical group ‘passed’ compared to 92% of the control group
Test 1 Part D: Gender Consistency
The amab children were asked questions like ‘if you wear a dress, are you a girl?’ ‘If you played with a doll would you be a girl?’. (the exact script, and the exact phrasing, is not provided so we cannot be sure exactly how the questions were worded)
66% of the clinical sample ‘failed’ this test, by stating that playing with dolls makes you a girl.
46% of the control group also ‘failed’, also thinking that playing with a doll made you a girl.
The fact that nearly half the control also think playing with a doll makes you a girl seems more an indication of the segregated and gendered restrictions on toys of Canadian children in the 70s, 80s and early 90s than any conclusion about gender identity. Given very few of the clinical group identified as trans in this study, it also brought to mind the limited freedom for boys to be feminine or play with perceived girls toys, and made me wonder how many assigned males had been told to ‘stop being a girl’ when playing with dolls or putting on a dress.
Zucker 1999 Test 1 – Conclusion
The data from test 1 parts A-D, and the fact that the clinical sample had a slightly lower ‘pass’ rate than the control sample, was interpreted by Zucker et al. as evidence that children at the gender clinic were more ‘confused’ about gender.
The researchers then take a further leap of faith, into a conclusion that the lower pass rate of the clinical group compared to the control group implied a ‘developmental lag’ in understanding of gender. However, the clinical sample and the control sample were the same age (ages 4-8, average age 6.5), and the clinical sample were not re-tested at a later point in time. How therefore can they claim a developmental lag? It is simply not possible to claim a ‘developmental lag’ based on this data. The assertion of a ‘developmental lag’ (with the implication that the clinical sample reach a similar level of understanding but at a later age than children not referred to a gender clinic) is pure speculation/fabrication.
Test 1 provides zero evidence that transgender children (those with a consistent, insistent, persistent identity different to their assigned sex) have a delayed understanding of gender.
Zucker 1999: Test 2 Boy-Girl Identity Test
The assigned male children were then shown a drawing of a boy. They were asked to give the child in the drawing a name. If they chose a girl’s name for the drawing they were corrected and given a boy’s name, eg ‘this is Tom’.
Zucker 1999 Test 2 Part 1:
The children were asked ‘If Tom really wants to be a girl, can he be a girl?’.
The children ‘fail’ this test if they say yes Tom can be a girl.
32% of the clinical sample ‘fail’ this task by stating Tom can be a girl vs 21% of the control group
(Interestingly there is not a huge difference between the clinical group and the control group, with 21% of the control group, children with no gender issues, thinking that yes, Tom can be a girl – Canadian 5 year olds from the 1970s, 1980s and early 1990s showing more sophisticated understanding of gender than their specialists…)
The children are asked to give a reason for their answer, and are defined as having reached ‘operational constancy’ if they justify their answer ‘No Tom can’t be a girl’ answer with the justification ‘because he was born a boy’.
This unethical line of questioning Zucker et al put trans children through is a form of coercive persuasion, tantamount to brainwashing, where the ‘wrong answer’ is corrected with the ‘right answer’ – ‘BECAUSE TOM WAS BORN A BOY’ until the child submits and agrees.
Zucker 1999 Test 2 Part 2:
A dress and or long hair is added to the picture of ‘Tom’ and the children are asked ‘If Tom puts on a dress, is he a girl?’ The expected ‘correct’ answer is ‘No’ and the expected justification is ‘because he was born a boy’.
71% of the clinical group ‘fail’ this test, as do 64% of the control group.
Again the control group is pretty similar to the clinical group in their openness to the possibility of Tom being a girl.
Zucker 1999 Test 2 Conclusion
Zucker 1999 Test 2 provides no evidence of a ‘developmental lag’ in understanding gender for trans children.
Zucker 1999 – Test 3 looks at ‘sex-typed behaviour’
Zucker 1999 Test 3 Part 1 asks children to draw a person, and then asks them if the person they have drawn is a boy or a girl.
The assigned male at birth children ‘pass’ if they opt to draw a boy and ‘fail’ if they opt to draw a girl.
66% of the clinical sample ‘fail’ by drawing a girl (in case of the assigned males). 54% of the control sample also ‘fail’ this test by drawing a girl (in case of the presumed cis boys).
Again the results of the control are fairly similar to the clinical group.
Again there is no evidence of trans children having a developmental lag in understanding gender.
Zucker 1999 Task 3 Part 2 has the children watched through a 1-way mirror in a room with ‘gender specific’ toys and or clothes and the researchers assess the amount of time the children spend with ‘appropriately gendered toys or clothes’. They are deemed to have ‘failed’ if they play too much with the ‘wrong’ gender toys or clothes.
This task has no place in today’s society in which children are not constrained by outdated gender stereotypes.
Zucker 1999 Task 3 Part 3 assesses what it calls ‘affected confusion’, assessing a child’s ‘desire’ to be a boy or a girl (rather than their identity). It asks assigned males (who have been referred to a gender clinic for non-conforming behaviour) questions like ‘is it better to be a boy or a girl’ and ‘do you ever wish you were a girl’. Assigned males are deemed ‘deviant’ if they state any wish to be a girl (perhaps because they are a trans girl or perhaps because they are a non-conforming boy who wants to be able to play with his dolls in peace without being taken to Dr Zucker every month/week…). Assigned males are also deemed ‘deviant’ if they acknowledge anything positive about girls or think that there are any ways in which it is better to be a girl than a boy.
64% of the clinical group are labelled ‘deviant’ due to their answers in this part, as are 50% of the control group (reminder the control group are ‘random’ children not being seen by the gender clinic and who are not known to have any gender issues and yet half are deemed by this test to be ‘deviant’).
Zucker 1999 Test 3 Conclusion
What on earth is going on, and how the heck is this research still being quoted in a 2016 journal article by the leading experts at the UK’s Children’s Gender Service!
The high ‘failure’ and ‘deviancy’ rate’ not only in the clinical group but also in the control group is perhaps indication that Canadian 5 years olds in the 70s, 80s and early 90s were did not have such ingrained stereotypes of gender, nor sexism, nor misogyny, as the ‘gender specialists’ who subjected them to such awful tests.
The Zucker 1999 article ends with a ‘blame the parents’ conclusion, proposing that parents of gender non-conforming boys or trans girls must have ‘actively’ encouraged ‘cross-dressing’ or appeared to ‘tolerate’ cross-gender behaviour. It highlights a view that “parental reinforcement of same-sex play was positively related to gender constancy in pre-schoolers”. It is pure ‘drop-the-barbie’ Zucker, more focused on installing out-dated gender conformity in non-conforming boys than any concern for how to help children who may be transgender.
Zucker 1999 in summary
To summarise, the Zucker 1999 research is deeply outdated, transphobic, stereotyped, homophobic, normative and unethical.
The clinical sample is known to contain at least some children who are non-conforming rather than trans and no effort is made to focus specifically on trans children.
The difference in answers between the clinical group and the control group are very small (the paper conducts regression analysis on a wide number of variables until it finds some that are considered statistically significant – this is an unsound approach to valid statistical analysis).
The study looks at a range of things that do not relate to gender identity (including toy preferences).
And most critical of all – the paper looks at the children at one moment in time – comparing the clinical group to a control group of children the same age (age range 4-8). There is no follow up at a later age and no comparison of children of different ages. Any claim to transgender children having a developmental lag (which I interpret to mean achieving a similar understanding of gender at a later age) is pure fabrication.
Given the obvious weaknesses of the the Zucker paper – why is it a core reference in the 2016 Costa paper from leading experts of the UK Children’s Gender Service
UK Children’s Gender Service
There are three major problems with the Tavistock’s view of gender constancy as shown in the Costa 2016 paper.
Problem 1 – Quoting fabricated conclusions as though evidence
The Costa 2016 paper states “Some research suggests that a developmental lag exists in gender constancy acquisition in children with gender variant behaviour”, referencing the Zucker 1999 paper.
As we’ve seen above, the Zucker 1999 research does not provide any evidence for this claim.
Problem 2 – Broadening the relevance of those unsound conclusions and applying them to clinical practice
The Zucker paper mentions a ‘developmental lag’, but does not mention any age at which transgender children reach ‘gender constancy’.
The Costa 2016 paper moves beyond even the conclusions claimed in Zucker 1999, taking a series of assumptions to extrapolate wider conclusions (for which no specific reference is provided). They move from the idea of ‘a developmental lag in gender constancy’ to the claim that “children with gender dysphoria are more likely to express an unstable pattern of gender variance”. They move from that unsupported statement to the claim that “treating prepubertal individuals with gender dysphoria is particularly controversial owing to their unstable pattern of gender variance compared with gender-dysphoric adolescents and adults”. And they shift further to arrive at the final statement that puberty suppression is unwise until at least age 12 “safely beyond the age of gender constancy”. It is important to recall that Zucker 1999 provides no information on the age at which trans children ‘achieve gender constancy’ and focuses on children aged 4-8 (where the Costa paper get the age 12 figure from for gender constancy is anyone’s guess).
The Zucker 1999 research bears no relevance to the question of whether trans children understand their gender. Yet Costa et al 2016 use this study as their basis to suggest that transgender children do not know their gender. They use it to support an argument that transgender children have ‘unstable gender variance’ up until puberty, and they extend this to argue against pubertal suppression for those starting puberty under the age of 12.
Problem 3 – Omitting reference to critical recent research
The third, and perhaps the biggest error, is one of omission. The Costa 2016 paper, in its section on gender identity development / gender constancy in transgender children, only mentions the Zucker research, omitting mentioning any other research on transgender children’s gender identity.
The Zucker 1999 research paper is now 19 years old. Surely in the last two decades there has been some other research on the gender identity of transgender children, ideally research that makes an effort to focus on transgender (rather than gender non-conforming) children, and research that focuses on the children’s gender identity rather than toy preference? The answer is a clear yes. There are important studies on this topic that the Costa 2016 paper fails to even mention.
So let’s take a brief review of recent research on this topic which are noticeable by their absence:
New research on transgender children’s gender identity
Olson, K.; Key, A.; Eaton, N. (2015) Gender Cognition in Transgender Children Psychological Science OnlineFirst, published on March 5, 2015
The introduction to the Olson el al 2015 paper describes historical (and current) scepticism to the idea of transgender children knowing their gender:
“This scepticism takes many forms: concerns that these children are “confused” and that they therefore need therapy, that these children are “delay[ed]” in their understanding of gender in part because of the behaviour of their parents (Zucker et al., 1999: Gender constancy judgments in children with gender identity disorder: evidence for a developmental lag), or that these children are merely saying they are the “opposite” gender, much as they might say on any given day that they are a dinosaur or princess.”
Olson et al.’s research aimed:
“to investigate whether 5- to 12-year-old prepubescent transgender children (N = 32), who were presenting themselves according to their gender identity in everyday life, showed patterns of gender cognition more consistent with their expressed gender or their natal sex, or instead appeared to be confused about their gender identity.”
“When the transgender children’s responses were considered in light of their natal sex, their responses differed significantly from those of the two control groups on all measures. In contrast, when transgender children’s responses were evaluated in terms of their expressed gender, their response patterns did not differ significantly from those of the two control groups on any measure.”
“Using implicit and explicit measures, we found that transgender children showed a clear pattern: They viewed themselves in terms of their expressed gender and showed preferences for their expressed gender, with response patterns mirroring those of two cisgender (nontransgender) control groups. These results provide evidence that, early in development, transgender youth are statistically indistinguishable from cisgender children of the same gender identity.
Our findings refute the assumption that transgender children are simply confused by the questions at hand, delayed, pretending, or being oppositional. They instead show responses entirely typical and expected for children with their gender identity.
The data reported here should serve as evidence that transgender children do indeed exist and that their identity is a deeply held one.”
See here for background and further details
Fast, A & Olson, K. (2017) Gender Development in Transgender Preschool Children, Child Development
“An increasing number of transgender children—those who express a gender identity that is “opposite” their natal sex—are socially transitioning, or presenting as their gender identity in everyday life. This study asks whether these children differ from gender-typical peers on basic gender development tasks. Three- to 5-year-old socially transitioned transgender children (n = 36) did not differ from controls matched on age and expressed gender (n = 36), or siblings of transgender and gender nonconforming children (n = 24) on gender preference, behavior, and belief measures. However, transgender children were less likely than both control groups to believe that their gender at birth matches their current gender, whereas both transgender children and siblings were less likely than controls to believe that other people’s gender is stable.”
So what do we know about gender constancy/ gender identity in transgender children?
We know that some claims are balderdash (junk science):
- The Zucker 1999 study holds no value in informing us about the gender identity development or constancy of transgender children.
- The statement “a developmental lag exists in gender constancy acquisition in children with gender variant behaviour” is unsubstantiated and shouldn’t be quoted in future articles
- Conclusions in the Costa (2016) report on “children with gender dysphoria are more likely to express an unstable pattern of gender variance” and “treating prepubertal individuals with gender dysphoria is particularly controversial owing to their unstable pattern of gender variance compared with gender-dysphoric adolescents and adults” are unsubstantiated and should be disregarded
- Policy recommendations in the Costa (2016) report on withholding pubertal suppression until “at least the age of 12, safely beyond the age of gender constancy” are unsubstantiated and should be disregarded
We know that recent research (Olson 2015 and Fast 2017) shows that:
- “Transgender children do indeed exist and their identity is a deeply held one.”
- “Three- to 5-year-old socially transitioned transgender children did not differ from controls or siblings on gender preference, behaviour, and belief measures.”
- “Transgender children aged 5-12 viewed themselves in terms of their expressed gender and showed preferences for their expressed gender, with response patterns mirroring those of two cisgender control groups.”
- “These results provide evidence that, early in development, transgender youth are statistically indistinguishable from cisgender children of the same gender identity.”
- Research “findings refute the assumption that transgender children are simply confused by the questions at hand, delayed, pretending, or being oppositional. They instead show responses entirely typical and expected for children with their gender identity.”