Transphobia: Definition and Impact

Transgender (or trans) and Gender Diverse (TGD) people are those whose gender identity and/or expression differ from that which is usually associated with their sex assigned at birth (APA, 2014; Currah et al., 2006). TGD people present a higher prevalence of mental and physical health challenges when compared to cisgender people (i.e., those whose gender identity and sex are aligned) (e.g., Su et al., 2016), including higher levels of depression and anxiety (e.g., Borgogna et al., 2019), as well as higher risk of suicidality (e.g., Bränström et al., 2022). The higher vulnerability for TGD people to experience mental health difficulties is suggested to result from an interplay of general risk factors and the experience of trans-specific minority stressors (Hendricks & Testa, 2012), such as discrimination, harassment, and abuse (e.g., Drabish & Theeke, 2022). TGD people report high exposure to stigma (e.g., Bockting et al., 2020), which can hinder access to societal resources (e.g., employment, healthcare) (see Hughto et al., 2015) and to establishing a strong support network (e.g., Jackman et al., 2018), thus leading to social isolation (e.g., Wilson & Liss, 2022).

Transphobia is described as the irrational fear, hatred, and overall prejudicial attitude towards those whose gender identities and/or presentations are not aligned with societal gender expectations (Hill & Willoughby, 2005) and is mainly directed towards TGD people (Tebbe & Moradi, 2012). Research suggests that, although with growing acceptance, western countries (e.g., USA, UK) continue to endorse negative attitudes towards TGD people (Lewis et al., 2017; Storrie & Rohleder, 2018) and low support for pro-trans policies (Axt et al., 2021), including in samples of students of helping professions (e.g., Acker, 2017). Understanding transphobia and its underlying sociopolitical factors (here defined as social and/or political factors that are the root causes and/or consequences of social, economic, or political phenomena) provides much-needed information for policy-making and social justice (see Ciocca et al., 2020). For example, studies conducted in western countries (e.g., UK, Germany) suggest that higher education (until tertiary education) is associated with more positive views of homosexuality (e.g., La Roi & Mandemakers, 2018), perhaps due to the socialization with sexual and gender diverse colleagues (e.g., Carvacho et al., 2013), but to our knowledge, the role of education on trans attitudes has not been explored. Also, low personal contact with TGD people, high religiosity, and political conservatism seem to be correlated with more negative attitudes towards TGD people (Norton & Herek, 2013).

Sociopolitical Factors Underlying Transphobia

Allport’s (1954) seminal work on the impact of interpersonal contact on attitudinal change laid the ground for future studies on prejudicial attitudes towards sexual and gender minorities. Studies mostly conducted in the USA show that more interpersonal contact with sexual minority individuals (e.g., gay, lesbian, bisexual people) correlates with less prejudicial attitudes (e.g., see Smith et al., 2009). However, contact with TGD people is reportedly less common than with sexual minorities, which can limit contact to solely via media representation (Hoffarth & Hodson, 2018) that may lack positive and/or accurate depictions of TGD identities (e.g., McLaren et al., 2021). Overall, reduced exposure to and contact with TGD people may result in discrimination and transphobic bias (King et al., 2009). Contrarily, interpersonal contact with TGD people can reduce transphobia (e.g., Barbir et al., 2017; Flores et al., 2017), and positively leads to support of TGD rights (e.g., Flores et al., 2018; Tadlock et al., 2017). However, studies have focused more on frequency and quality of contact (positive versus negative) with TGD people (see Hodson et al., 2013), and less on categories of relationship (e.g., work colleagues, friends, family members), which seem to influence, for example, attitudes towards same-sex marriage in the USA (Dyck & Pearson-Merkowitz, 2014).

Studies mostly conducted in samples of Christian participants (e.g., Johnson et al., 2010; Rowatt et al., 2014), and more recently with Muslim participants (e.g., Yeck & Anderson, 2019), suggest that religiosity (i.e., the degree to which a person is committed to, or involved in, religious practices and/or beliefs) is one of the causal factors of prejudice. Although diversity within religious denominations and experiences exists (Cragun, 2013), religion seems to be at the core of the prejudice experienced by sexual and gender-minoritized people (Barton, 2012). Various measures of religiosity have been consistently correlated with transphobia (e.g., Cragun & Sumerau, 2015; Kanamori & Xu, 2022; Tadlock et al., 2017), and overall results suggest that self-identifying as religious is significantly correlated with transgender prejudice (see Campbell et al., 2019). However, the largest growing number of religion-related affiliation (or lack thereof) is the “nones” — atheist, agnostic, or religiously unaffiliated (Newport, 2022) — and to our knowledge, their attitudes towards TGD people have not been systematically studied in the general population (most studies are conducted in samples of college students), especially regarding their acceptance of, comfort with, and tolerance towards trans people. To our knowledge, studies have focused more on religious people’s transphobia/genderism. However, lower levels of transphobia/genderism do not necessarily equate to a positive view of TGD people. Notwithstanding, one transnational report concluded that, more than religion itself, transphobia and trans-acceptance are influenced by the social and cultural context (Balzer et al., 2015).

Research suggests that conservatism and right-wing authoritarianism are associated with sexual prejudice (e.g., Poteat & Mereish, 2012) and transphobia (e.g., Nagoshi et al., 2008) when compared to liberal political views (e.g., Barth & Parry, 2009; Hoyt et al., 2019), and predicts opposition to TGD rights (Miller et al., 2017), including in samples of gay men (e.g., Warriner et al., 2013). Authoritarians are known to value traditional gender roles and conformity with sexual norms, which can result in anti-TGD regulations (e.g., Whitley & Ægisdóttir, 2000) and lower acceptance of sexual and gender minorities (e.g., Hetherington & Jonathan, 2009). Transphobia has also been associated to one’s knowledge about sex and gender, showing that those who ascribe to bioessentialist beliefs about the origins of gender differences present higher levels of transphobia (Konopka et al., 2020). However, most studies have focused on the liberal-conservative binary of the American political system, which fails to grasp the nuances and heterogeneity of the social and fiscal ideological landscape of European multi-party systems (e.g., Alves & Porto, 2022; Feldman & Johnston, 2014). One study conducted in a sample of 302 Portuguese active politicians found statistically significant differences in sexual prejudice across the social and fiscal left–right spectrum, with the conservative right scoring higher and the left lower (Ferros & Pereira, 2021).

Transphobia in Portugal: Brief Sociocultural and Legal Context

Sociopolitical factors underlying transphobia have not been extensively explored in Portugal. In fact, few studies in Portugal have explored the attitudes of the Portuguese population towards TGD people (see Moleiro et al., 2022 for a review). To our knowledge, only two studies have directly explored the general population’s attitudes towards TGD people, both in samples composed of adolescents, and only one targeted exclusively attitudes towards trans people (Carrera-Fernández et al., 2020) instead of sexual and gender minorities (Costa & Davies, 2012). Indeed, studies on TGD people have been lacking in Portugal, perhaps due to the overall absence of visibility of transgender issues, only recently catalyzed by the legislative change on trans rights (see Saleiro, 2021).

Portugal was under a highly catholic right-wing dictatorship for almost half of the twentieth century, until the Carnation Revolution in 1974. Only after a twenty-year lull the first LGBTQI + (lesbian, gay, bisexual, transgender, queer, intersex) movement started to emerge in the 1990s (see Santos, 2016 for an overview). In 2011, the first legal gender recognition law was approved and implemented, allowing a person to change their legal name and gender in a Civil Registry if accompanied by a medical document attesting a “gender identity disorder” diagnosis. The medical gatekeeping of trans identities was recently overcome through the self-determination of gender Law (38/2018), which eliminated the need for a medical diagnosis to legally change a person’s gender and name (see Moleiro & Pinto, 2020; Saleiro, 2021).

According to the 2021 Census, more than 80% of the Portuguese population self-identifies as Roman Catholic, followed by a growing number of none-religious (more than 14%) (Instituto Nacional de Estadística, 2022). Contrary to what had been reported as Portugal’s inherent immunity against right-wing populist movements (e.g., Quintas da Silva, 2018), a precipitous rise of the ultra-right vote occurred in the 2022 legislative election, making the ultra-right party the third most voted (7.18%, i.e., 399.510 votes, with 12 seats in the Parliament), which fourfolded in parliamentary seats in the 2024 legislative election (i.e., 18.06%, with 48 seats). Despite Portugal’s progressive legal framework toward sexual rights (e.g., law 9/2010 of marriage equality, law 2/2016 of access to adoption by same-sex couples, law 38/2018 of gender self-determination), studies continue to show discrimination towards TGD people (e.g., Beatriz & Pereira, 2022; Gato et al., 2020), and comprehensive and systematic data on attitudes towards TGD people according to sociopolitical factors that may be related to discrimination are lacking.

The Pitfalls of Measuring Transphobia

It is commonly acknowledged that one major impediment to studying transphobia is the almost absence of psychometrically tested and sound instruments. One meta-analysis found 83 measures of transphobia and/or proxy constructs, none of which respecting sound psychometric requirements (Morrison et al., 2017). When it comes to attitudes towards TGD people, the existing four measures — the Genderism and Transphobia Scale (Hill & Willoughby, 2005); the Transgender Attitudes and Beliefs Scale (Kanamori et al., 2017); the Attitudes Towards Transgender Individuals Scale (Walch et al., 2012); and the Transphobia Scale (Nagoshi et al., 2008), although well designed and valid measures, did not include the contribution of TGD individuals (e.g., item generation, feedback on the items content) during the qualitative assessment of measures’ content validity.

Transgender Knowledge, Attitudes, and Beliefs Scale

The Transgender Knowledge, Attitudes, and Beliefs scale (T-KAB) was developed through a thorough literature review and included consult experts in transgender issues and TGD people during the content development (Clark & Hughto, 2020). This 22-item measure was designed and tested through an iterative and rigorous process, including resorting to qualitative interviews with TGD individuals and healthcare professionals. T-KAB has the advantage of separating acceptance of the gender spectrum (which might be influenced by information regarding transgender identities and scientific knowledge on the complex relationship between sex and gender) and social tolerance towards TGD people (which might tap into a more general tolerant attitude towards others). This provides a more nuanced lens through which the social sentiment towards transgender people might be studied, thus providing a more accurate understanding of which sociopolitical factors are associated with a more antagonistic versus accepting attitude towards TGD people.

Translated Versions of the T-KAB

The three-factor structure of the T-KAB has been corroborated in its Spanish version (Fresán et al., 2022). However, more research is needed to conclude its cultural generalizability, given that the Spanish version found a poor factor loading of item 1 (which was eliminated), and advised more testing of the multicollinearity of items 20 and 22. Additionally, convergent validity was either not tested (Fresán et al., 2022) or tested with general attitudes toward sexual minorities instead of with other transphobia measures (Clark & Hughto, 2020). Finally, although studies showed that the T-KAB is sensitive to finding changes in transgender acceptance after a brief video-based psychoeducation resource (Martin et al., 2022) and that it positively correlates with more liberal political views, and lower religiosity in a sample of dietitians/nutritionists (Buonaiuto et al., 2023) and of cisgender heterosexual Filipinos (Reyes et al., 2024), no study was yet conducted to test the T-KAB ability to differentiate between sociopolitical groups. This is paramount to better understand and overcome transphobia through psychoeducational programs tailored and targeted to specific sociopolitical segments of the population, given that even low levels of prejudice may result in social discrimination and great social consequences (Greenwald et al., 2015).

The current study aims to explore the psychometric properties of the Portuguese version of the T-KAB, as well as to test differences in attitudes towards TGD people according to sociopolitical factors (i.e., education, socioeconomic status, sexual orientation, contact with TGD people, religiosity, and political ideology).

Method

Participants

The current study was conducted using a convenience sample of N = 467 Portuguese adults from the general population, ages between 18 and 65 years old (M = 30.28, SD = 10.89). Sociodemographic characteristics of the participants were self-reported and are depicted in Table 1. Socioeconomic status (SES) was based on participants’ self-report of occupation, according to the Portuguese profession classification: low (e.g., undifferentiated workers, farmers, cleaning staff), medium (e.g., psychologists, schoolteachers, nurses), and high (e.g., university professors, medical doctors, judges) (Instituto Nacional de Estatística, 2011). Participants who were students, unemployed, and retired were also encoded (1 = low SES; 2 = medium SES; 3 = high SES; 4 = student; 5 = unemployed; 6 = retired; 7 = working student).

Table 1 Sociodemographic characteristics of participants (N = 467)
Full size table

Almost half of participants reported knowing a transgender person (N = 222; 47.5%), of which N = 67 reported knowing one transgender person (30.2%), N = 101 between two and five transgender individuals (45.5%), N = 33 between six and ten trans people (14.9%), N = 11 between eleven and twenty trans people (5.0%), and N = 10 reported knowing more than twenty trans individuals (4.5%). Regarding the relationship with trans people, participants reported knowing trans people from television (N = 7; 3.2%), by sight (N = 75; 33.8%), as co-workers (N = 11; 5.0%), friends (N = 113; 50.9%), and family members (N = 11; 5.0%). Five participants preferred not to answer (2.3%).

Procedure

The translation process of the T-KAB followed established guidelines on the validation of instruments (e.g., Sousa & Rojjanasrirat, 2011; Sperber, 2004), except for pre-testing and cognitive interviewing due to limited resources. The first translation into Portuguese was conducted by the first author, who is fluent in English and Portuguese. Then, two researchers and clinicians not involved in the study, fluent in English and Portuguese, with expert knowledge of validation studies and/or LGBT + health, back-translated the items into English. The accuracy of the items was assessed. Small changes in wording were conducted (e.g., item 9 from “Iria sentir-me confortável se soubesse que o meu vizinho é trangénero” to “Eu sentir-me-ia confortável se descobrisse que uma pessoa minha vizinha é transgénero”). The final version was included in the study protocol.

Before starting data collection, the current study was evaluated and approved by the Scientific Research Ethics and Deontology Committee of the School of Psychology and Life Sciences of Lusófona University (Lisbon, Portugal). The study was advertised on social media platforms through convenient sampling. In order to attain a sexually and gender-diverse sample, the study was also advertised in LGBT + social media pages and groups. Also, national LGBT + rights associations were invited to collaborate by advertising the study in their newsletters and mailing lists. Data collection was conducted online through Qualtrics software. The online protocol was accessed by N = 607 participants, of which 137 have not proceeded. Additionally, three participants were excluded from data analyses as their responses indicated bias and unreliable answers (e.g., adding “unicorn” to gender identity, answering “a lot of sex” to the question regarding biological sex). The final sample is composed of N = 467 participants. Informed consent was provided digitally by clicking on “I accept to participate”, after reading information on the study goals, target population, voluntary participation, anonymity, email of the research team to contact in case of doubts, average duration, and confidentiality of data. Databases were protected by a password only the research team could access. IP address and geolocation information were eliminated from the extracted databases.

Measures

Sociodemographic Questionnaire

Sociodemographic variables were assessed through self-report questions. In addition to questions regarding age, marital status, and education, participants responded questions regarding sociopolitical factors. Participants were instructed to choose one of the available response options to each question. Sexual orientation was self-reported according to the following options: heterosexual, gay, lesbian, bisexual, queer, undefined, other (which?), I prefer not to answer. Sex was assessed according to the following options: male, female, intersex, other (which?), and I prefer not to answer. Gender was assessed through the following response options: man, woman, non-binary person (e.g., gender queer, gender fluid), agender (i.e., no gender), I do not know, other (which?), and I prefer not to answer. The contact with TGD people was assessed using three questions: one binary question (yes/no): “Do you know any transgender person?”, and two multiple choice questions: “how many transgender people do you know” (options: one person, two to five people, six to ten people, eleven to twenty people, I prefer not to answer) and “what is your relationship with the transgender people that you know (you can choose more than one option)” (options: I know from television, I know by sight, We are work colleagues, We are friends, We are family, I prefer not to answer). Religion was assessed by self-reporting one of the following options: Catholic, Protestant, orthodox Christian, Jewish, Muslim, Buddhist, Hindu, Agnostic, Atheist, Other (which?), I prefer not to answer. Participants self-reported on political ideology by choosing one of the following options: Marxist or Trotskyist Left, Socio-democratic Left, Socio-democratic Right, Christian-democratic Right, Far Right or National Populism, Conservative, Liberal, Libertarian, Other (which?), I prefer not to answer.

Transgender Knowledge, Attitudes, and Beliefs Scale

This measure (T-KAB; Clark & Hughto, 2020) is a 22-item measure extracted from a pool of 44 generated items that was developed to assess societal attitudes towards transgender people, using a 4-point Likert-like scale (1 = strongly disagree; 4 = strongly agree). The original version found a three-factor structure that measures acceptance of the gender spectrum (8 items; e.g., “sex and gender are the same thing”), social tolerance (7 items; e.g., “Transgender people are a worthwhile part of our society”), and comfort and contact with transgender people (7 items; e.g., “I would feel comfortable if I learned a friend is transgender”). The total score of the T-KAB was highly and negatively correlated with homophobia (r =  − .80) and moderately with social dominance (r =  − .55). The original study found excellent levels of reliability: acceptance of gender spectrum α = .95, social tolerance α = .93, and comfort and contact α = .91.

Genderism and Transphobia Scale – Short Version

This measure (GTS-short version; Carrera-Fernández et al., 2014, 2020) is developed and validated by Hill and Willoughby (2005) to assess attitudinal and behavioral components of transphobia, in a 6-point scale (1 = strongly agree; 6 = strongly disagree). In its original long version, GTS is composed of 32 items that measure transphobia/genderism (25 items) and gender bashing (7 items). A short Portuguese version with 12-items was validated, which found the same two-factor structure, with good reliability: α = .82 for the gender bashing factor, and α = .86 for the transphobia/genderism factor (Carrera-Fernández et al., 2020). Given that the Portuguese validation study was conducted in a sample of adolescents, we have changed the words “boys” to “men”, and “girls” to “women”. The current study found good reliability results: α = .86 for the gender bashing factor and α = .90 for the transphobia/genderism factor.

Data Analysis

All statistical analyses were conducted using IBM SPSS Statistics (version 27.0). The factor structure of the T-KAB was explored through Exploratory Factor Analysis (EFA) with Principal Axis Factoring (PAF) as the estimation parameter method. Promax rotation was conducted as the resulting factors were expected to be significantly correlated (Tabachnick & Fidell, 2007). The factorability of the matrix was assessed through the Kaiser Meyer-Olkin (KMO) test, in which KMO > 0.80 is considered adequate (Kaiser, 1974; Sharma, 1996). The number of extracted factors was analysed using the eigenvalues, with eigenvalues > 1 suggesting factor retention (Johnson, 1998). Additionally, and given that eigenvalues may result in factor over-extraction (Hubbard & Allen, 1987; Zwick & Velicer, 1986), the PAF results were confirmed with a parallel analysis (PA). Factors are confirmed when the PAF eigenvalues are greater than the corresponding PA eigenvalues (Franklin et al., 1995). To estimate the variance of each item accounted for by its factor, extraction communalities (h2) were considered. Results indicating h2 < .30 suggested the items should be eliminated (Hair et al., 1998).

The reliability of each factor was assessed through Cronbach’s alphas (α). Results of α > .70 were indicative of acceptable reliability (Field, 2013). Item-total correlations > .50 also suggested acceptable factor reliability (Nunnally, 1978; Tabachnick & Fidell, 2007).

Convergent validity analyses were conducted through correlation analyses between T-KAB and GTS, given that the instruments measure related constructs (Cohen et al., 2003).

Differences between two groups on T-KAB were tested using t-test analyses. Significant differences between means were considered when p < .05. The effect sizes were assessed according to Cohen’s d, where magnitudes of d = .20 were interpreted as small, d = .50 as medium, and d = .80 as large effect sizes (Cohen, 1988). Differences between three or more groups were tested with one-way analyses of variance (ANOVA). Differences in T-KAB scores were calculated according to education, socioeconomic status, sexual orientation, contact with transgender people (number and relationship), religious affiliation, and political ideology. Effect sizes were calculated with eta squared (η2), where η2 = .01 indicates small effect size, η2 = .06 medium effect size, and η2 = .14 large effect size (Tabachnick & Fidell, 2007). To test pairwise group differences, post-hoc analyses were conducted using the Scheffe test, due to unequal group sample sizes, and the Games-Howell test when the homogeneity assumption was violated. A post-hoc power analysis for one-way ANOVA fixed effects was conducted in G*Power, considering an effect size of f = 0.25, α = .05, a sample size of N = 467 and number of groups = 10 (“political ideology” has the most categories: 10), and suggested a power of 98.20%.

Results

Preliminary Data Analyses

Preliminary analyses on T-KAB showed acceptable values of skewness (Sk <|3|; Skmin =  − 2.61, SKmax =  − .55) and kurtosis (Ku <|8–10|; Kumin =  − .71, Kumax = 6.99), which suggest that normal distribution was not violated (Kline, 2005). Thirty-six moderate univariate outliers were found throughout the 22 items of T-KAB, none of which showing a pattern throughout the measure. Our decision to maintain the outliers followed the assumption that in doing so the data would more likely resemble the variability of the population in the study (Tabachnick & Fidell, 2007). No missing values occurred since the online protocol did not allow participants to proceed and/or submit incomplete questionnaires.

Factor Structure Analysis

A Principal Axis Factoring was conducted, with Promax rotation, to explore the dimensionality of the T-KAB. Results showed the data was adequate [KMO = 0.954; χ2 (231) = 9309.04, p < 0.001], and the eigenvalues suggested the extraction of two factors explaining 56.10% (factor 1; eigenvalue = 12.34) and 5.46% (factor 2; eigenvalue = 1.20). To overcome potential over-extraction, a parallel analysis (PA) was conducted and corroborated these results, showing that the two components’ eigenvalues were greater than the corresponding randomly generated matrix in PA (factor 1: EigenvaluePA = .46; factor 2: EigenvaluePA = .39). The extracted communalities of all items were h2 > .30 (h2min = .33; h2max = .82), suggesting that all items should be retained (Table 2).

Table 2 Complete item pool, communalities, and factor loadings (N = 467)
Full size table

According to the clear pattern suggested by the factor loadings, factor 1 is composed of items that reflect acceptance and comfort with the gender spectrum (T-KAB_ACGS; items: 1, 2, 10, 11, 12, 13, 14, 15, 17, 18, 19, 20, 22), and factor 2 measures social tolerance (T-KAB_ST; items: 3, 4, 5, 6, 7, 8, 9, 16, 21).

Given that the original study (Clark & Hughto, 2020), as well as the Mexican validation (Fresán et al., 2022), found a three-factor structure (F1: acceptance with gender spectrum; F2: social tolerance; F3: comfort and contact), we have rerun the analysis fixing three factors, instead of using eigenvalues as criteria for factor extraction. Results suggested a poor factor structure as only two items (8 and 9) would load on factor 3. See Table 2 to compare the original and Portuguese factors.

Reliability Analyses

Considering the values of Cronbach alphas found, both factors of the T-KAB attained a good reliability: T-KAB_ACGS presented an α = .95 and T-KAB_ST an α = .94 (Table 3).

Table 3 Reliability analyses of the T-KAB (N = 467)
Full size table

Results showed good item-total correlations in both factors, from r = .54 to r = .87, and the factors’ reliability would not be improved by excluding any of the items.

Convergent Validity

To test convergent validity, Pearson correlation analyses were conducted between the T-KAB and the GTS. Results showed that T-KAB_ACGS was positively and significantly correlated with T-KAB_ST (r = .79, p < .001), with the gender bashing factor of GTS (r = .30, p < .001), and with the transphobia and genderism factor of GTS (r = .75, p < .001). Similarly, the T-KAB_ST was positively and significantly associated with the gender bashing factor of GTS (r = .39, p < .001), and with the transphobia and genderism factor of GTS (r = .74, p < .001).

Sociopolitical Trends and Attitudes: Differences Between Groups on T-KAB Scores

Sociopolitical differences in T-KAB scores were tested. Results showed that T-KAB_ST, but not T-KAB_ACGS (F(6) = 1.88, p = .083, η2 = .02), presented significantly different scores according to the level of education (F(6) = 9.80, p < .001, η2 = .04). The Games-Howell post-hoc test showed those with a high school degree (M = 31.64, SD = 5.34) presented significantly less social tolerance than those with a bachelor’s degree (M = 35.25, SD = .06, p = .004) and those with a doctorate degree (M = 35.31, SD = 1.32, p < .001). In terms of socioeconomic status, significant differences were found in T-KAB_ACGS (F(6) = 2.47, p = .023, η2 = .04), but not in T-KAB_ST (F(6) = 1.42, p = .205, η2 = .02). However, Scheffe’s post-hoc test did not find any pairwise statistical differences.

Significant differences were found in T-KAB_ACGS (F(8) = 21.38, p < .001, η2 = .19) and in T-KAB_ST (F(8) = 22.98, p < .001, η2 = .15) according to the participants’ sexual orientation. The Games-Howell post-hoc test showed heterosexual participants had significantly lower scores (M = 40.62, SD = 8.77) on T-KAB_ACGS than gay men (M = 47.48, SD = 5.69, p < .001), bisexual (M = 48.28, SD = 5.28, p < .001), queer (M = 49.83, SD = 1.60, p < .001), and pansexual people (M = 50.00, SD = 2.54, p < .001). Heterosexual participants also showed significantly lower scores (M = 31.38, SD = 5.45) on T-KAB_ST than gay men (M = 34.89, SD = 1.78, p < .001), lesbians (M = 34.58, SD = 3.08, p = .008), bisexuals (M = 35.17, SD = 1.91, p < .001), and pansexual people (M = 35.90, SD = .32, p < .001).

Results also show significant differences in T-KAB scores according to the participants’ contact with transgender people. Results from t-test analyses showed significant differences in T-KAB_ACGS (t(450.35) =  − 7.91, p < .001, d = .73) between those who know transgender people (M = 45.74, SD = 7.02) and those who do not (M = 39.81, SD = 9.11). Similarly, significant differences were found in T-KAB_ST (t(423.81) =  − 6.58, p < .001, d = .60) between those who report knowing transgender people (M = 33.92, SD = 3.78) and those who report not knowing a transgender person (M = 31.00, SD = 5.69). Significantly different scores were found in T-KAB_ACGS (F(4) = 13.58, p < .001, η2 = .15) and in T-KAB_ST (F(4) = 12.15, p < .001, η2 = .14) according to the number of transgender people participants report knowing. The Games-Howell post-hoc test showed those who reported knowing one transgender people have significantly lower T-KAB_ACGS (M = 42.28, SD = 8.09) than those who report knowing between two and five (M = 45.96, SD = 6.65, p = .021), six and ten (M = 49.45, SD = 3.15, p < .001), eleven and twenty (M = 50.18, SD = 3.25, p < .001), and more than twenty (M = 49.50, SD = 3.06, p < .001). Those who report knowing between two and five transgender people present significantly lower T-KAB_ACGS than those who know between six and ten (p < .001), eleven and twenty (p = .014), and more than twenty (p = .049). A similar pattern occurred in T-KAB_ST: those who report knowing one transgender person had significantly lower scores (M = 31.96, SD = 5.48) than those who report knowing between two and five (M = 34.32, SD = 2.67, p = .013), six and ten (M = 35.58, SD = .90, p < .001), eleven and twenty (M = 35.82, SD = .60, p < .001), and more than twenty (M = 35.60, SD = .70, p < .001). Those who report knowing between two and five transgender people also had significantly lower T-KAB_ST than those who know between six and ten (p < .001), eleven and twenty (p < .001), and more than twenty transgender people (p = .005). Significant differences were also found in T-KAB_ACGS (F(5) = 2.94, p = .014, η2 = .06) but not in T-KAB_ST (F(5) = 1.50, p = .233, η2 = .04), according to the relationship with transgender people. However, Scheffe’s post-hoc test did not find significant pairwise differences. Nonetheless, the difference between participants who report having transgender friends (M = 46.76, SD = 6.50) and participants who report knowing transgender people by sight (M = 43.39, SD = 7.98) almost reach significance (p = .058). It should be noted that a considerable amount of our sample (N = 113) reported being friends with at least a transgender person, which might not be a generalizable result, and thus suggesting possibly a self-selection bias resulting from the sample selection strategy (e.g., social media and LGBT + associations).

Significant differences were found between those who report being religious and those who report not being religious in both T-KAB_ACGS (t(430.69) = 6.81, p < .001, d = .65) and T-KAB_ST (t(353.28) = 7.67, p < .001, d = .72). Religious participants scored significantly lower on T-KAB_ACGS (M = 40.35, SD = 8.80) and on T-KAB_ST (M = 30.98, SD = 5.75) than non-religious ones (M = 45.58, SD = 7.18 and M = 34.26, SD = 2.90, respectively).

Significant differences in T-KAB_ACGS (F(9) = 24.91, p < .001, η2 = .30) and in T-KAB_ST (F(9) = 17.48, p < .001, η2 = .31) were found according to participants’ self-reported political ideology. Regarding T-KAB_ACGS, Games-Howell post-hoc tests found that those with a radical left ideology (Marxism/Trotskyism) scored significantly higher (M = 47.48, SD = 8.08) than those with a centre-right social democracy ideology (M = 37.55, SD = 7.97, p < .001), centre-right Christian democracy ideology (M = 28.36, SD = 7.24, p < .001), conservative ideology (M = 24.71, SD = 7.85, p = .001), and ultra-right ideology (radical right/extreme right) (M = 20.80, SD = 7.16, p = .004). Also, participants with a centre-left social democracy ideology scored significantly higher (M = 45.75, SD = 6.04) than those with a centre-right social democracy ideology (p < .001), centre-right Christian democracy ideology (p < .001), ultra-right ideology (p = .014), conservative ideology (p = .005), and liberal ideology (M = 42.37, SD = 7.49, p = .047), and significantly lower than those with a libertarian ideology (M = 50.00, SD = 1.87, p = .047). Those with centre-right social democracy ideology scored significantly higher than those with a centre-right Christian democracy ideology (p = .043), ultra-right (p = .049), and significantly lower than those with a libertarian ideology (p < .001). Participants with a centre-right Christian democracy ideology presented significantly lower scores than those with a liberal ideology (p = .001) and libertarian ideology (p < .001), as well as with those who report “other” political ideology (M = 40.05, SD = 10.87, p = .032) and who “prefer not to say” (M = 43.10, SD = 7.70, p = .001). Participants with a ultra-right ideology had significantly lower scores than those with a liberal (p = .019) or libertarian (p = .006) ideologies, as well as with those who respond “other” (p = .018) or “prefer not to say” (p = .019). Participants with a conservative ideology had significantly lower scores than those with liberal (p = .012) or libertarian (p = .001) ideologies, as well as with those who report “other” (p = .027) or “prefer not to say” (p = .010). Participants with a liberal ideology scored significantly lower than those with a libertarian ideology (p < .001).

Regarding T-KAB_ST, Games-Howell post-hoc tests showed that participants with a radical left ideology (M = 34.96, SD = 2.28) scored significantly higher than those with a centre-right social democracy ideology (M = 30.64, SD = 4.79, p < .001), centre-right Christian democracy ideology (M = 24.27, SD = 3.29, p < .001), ultra-right ideology (radical right/extreme right) (M = 14.60, SD = 7.64, p = .037), and those who “prefer not to say” (M = 32.21, SD = 4.84, p = .002). Participants with a centre-left social democracy ideology (M = 34.22, SD = 2.94) scored significantly higher than those with a centre-right social democracy ideology (p < .001), centre-right Christian democracy ideology (p < .001), ultra-right ideology (p = .044), and with those who “prefer not to say” (p = .003). Finally, participants with a centre-right Christian democracy ideology had significantly lower scored than those with a liberal ideology (M = 32.89, SD = 3.63; p < .001) and those who “prefer not to say” (p < .001).

Discussion

To our knowledge, this is the first psychometric study of the European Portuguese version of the T-KAB, and the second translation and validation study of the measure along with the Spanish version (Fresán et al., 2022). This study suggests that the Portuguese version of the T-KAB is a psychometrically valid measure of acceptance, comfort, and social tolerance towards TGD people and that it is sensitive to finding differences in attitudes towards TGD people between different sociopolitical groups.

Factor Structure of the Portuguese Version

Our study did not find, through Principal Axis Factoring and Parallel Analysis, the three-factor structure attained in the studies of the original (Clark & Hughto, 2020) and Spanish versions (Fresán et al., 2022). Instead, we found a two-factor structure: acceptance and comfort with the gender spectrum (ACGS = items 1, 2, 10, 11, 12, 13, 14, 15, 17, 18, 19, 20, 22) and social tolerance (ST = items 3, 4, 5, 6, 7, 8, 9, 16, 21). Except for items 8 and 9 (which are loaded in the ST factor), all items previously conceptualized as “contact and comfort” with trans people in the original and Spanish versions, loaded in the same factor as those conceptualized as “acceptance of the gender spectrum” in our study. This suggests that there is an underlying shared factor for both accepting the gender spectrum and being willing to contact and be comfortable interacting with trans people. This not only echoes the high correlation (r = .85) that was found between acceptance of the gender spectrum and contact and comfort (Fresán et al., 2022) but also tentatively seems to be aligned with the literature on the relationship between contact with trans people and trans acceptance (e.g., Hoffarth & Hodson, 2018). Differences between the factor structure of the Portuguese version and the Spanish one might result from differences in the statistical procedures followed: by using confirmatory factor analysis, Fresán et al. (2022) have defined a priori a three-factor structure, without exploring how the data would freely be grouped.

Reliability

The two-factor Portuguese version found good reliability, with all items presenting acceptable item-total correlations, and all contributing to the factor internal consistency. Regarding convergent validity, and as reported in the Spanish version study (Fresán et al., 2022), acceptance and comfort with the gender spectrum were significantly correlated with social tolerance, which suggests a potential bidirectional relationship between accepting the complex relationship between sex and gender, and being tolerant towards trans people (Martin et al., 2022).

Convergent Validity

To our knowledge, this is the first study to test the association between the T-KAB’s dimensions and a measure of transphobia. Results suggest that those who present higher acceptance and comfort with the gender spectrum, and who are more socially tolerant towards trans people, are less likely to endorse social beliefs of sex and gender being the same and/or biologically determined (i.e., genderism), present less aversion towards trans people (i.e., transphobia), and are less likely to engage in discriminatory behaviors and harassment against trans people (e.g., gender bashing). It should be noted that the correlation between acceptance and comfort with the gender spectrum and gender bashing was significant but small (r = .30), which might result from a reduced variance of gender bashing (i.e., few people endorsing statements of violence against TGD people). These results suggest the construct validity of T-KAB, as well as seem to be in line with previous studies that point to the relationship between knowledge on/of transgender issues and transphobia (e.g., Flores, 2015; Stroumsa et al., 2019).

Differences Between Sociopolitical Groups

The current study sought to explore whether the T-KAB was sensitive to detect differences between groups according to sociopolitical factors. Results found that those with a higher level of education (i.e., bachelor’s degree and doctorate) presented significantly more social tolerance towards trans people than those with a high school degree. These results are in line with previous studies suggesting that those who have not graduated college endorse more negative views of trans people (e.g., Norton & Herek, 2013). However, it should be noted that no significant differences were found in acceptance and comfort with the gender spectrum according to the level of education. Although previous studies have found that those who endorse bioessential views on gender (genderism) tend to have lower levels of education (e.g., Konopka et al., 2020), our results suggest that those with a higher education do not present significantly more acceptance and comfort with gender diversity. It may be the case that acceptance of the gender spectrum involves knowledge on the complex relationship between sex and gender, which is not necessarily a function of education. For example, one study found that although students with a higher education present a more favorable view of trans people, they present little knowledge on transgender identities (Ozamiz-Etxebarria et al., 2020).

Transgender attitudes were not significantly different according to socioeconomic status (SES), which is in line with the few existing studies exploring this variable (e.g., Clark & Hughto, 2020). Although SES has been widely explored as a risk factor for transgender people’s social precarity (e.g., Masa et al., 2023), to our knowledge SES has been almost absent from studies on trans attitudes (e.g., Kanamori et al., 2017). Given that SES entails not only income but also education (e.g., Morgan et al., 2009), it seems that the level of education is a more prominent factor associated with trans attitudes than SES.

Also, results suggest that heterosexual individuals present lower acceptance, comfort, and social tolerance towards trans people comparatively to gay, lesbian, bisexual, and other sexually minoritized individuals (LGB +). These results echo previous studies that showed heterosexual individuals present less favorable views of trans people (e.g., Warriner et al., 2013), which relates to their higher endorsement of binary conceptions of gender (e.g., Norton & Herek, 2013). It may also be the case that heterosexual individuals have less contact with transgender individuals, thus presenting a binary conception of gender and fewer positive attitudes towards trans people (e.g., Barbir et al., 2017). An interesting result was that although lesbians had significantly more social tolerance towards TGD people than heterosexual people, they did not statistically differ from them in terms of acceptance and comfort with the gender spectrum. This should be further explored in future studies, firstly by controlling for the effect of gender (given that our sample of heterosexual participants was composed of both men and women).

Indeed, those who reported being acquainted with transgender people presented significantly more acceptance and comfort with the gender spectrum, and more social tolerance towards trans people, than those who did not know any trans person. Although causality cannot be drawn from these results (i.e., more trans-accepting individuals might thus be more willing to engage interpersonally with trans people; and interpersonal contact with trans people might contribute to more positive views on trans people), they seem to corroborate studies that found transphobic bias (e.g., Barbir et al., 2017; Flores et al., 2017) and anti-trans rights (e.g., Flores et al., 2018; Tadlock et al., 2017) are correlated with less interpersonal contact with trans people. In fact, results seem to suggest that those who report knowing a larger number of trans people are significantly more accepting of and comfortable with the gender spectrum, and are more socially tolerant towards trans people, than those who know fewer trans people, which echoes previous studies (e.g., Hodson et al., 2013). No differences were found in trans attitudes according to the type of relationship with trans people, although differences in acceptance and comfort with the gender spectrum between those who report knowing a trans person by sight and those who are friends with trans people almost reached significance (p = .058). This calls for an in-depth examination not only of the frequency and type of relationship with trans people but also of the quality and closeness of the interpersonal contact.

Self-identifying religious participants were significantly less accepting of, comfortable with, and tolerant towards trans people when compared to non-religious participants, which is in line with a vast literature suggesting a correlation between religiosity and transphobia (e.g., Campbell et al., 2019). Although the current study did not explore the level of religiosity (e.g., fundamentalism, literal interpretation of scripture), but rather used a self-identifying category of religion, it is possible to hypothesize that those who have a religious affiliation present less positive attitudes towards trans people due to a religiously-fueled world view of gender and sex as coherent and immutable god(s)-given features, thus following specific doctrinal ordinances that advise against gender variance and non-conformity (Whitley, 2009). In the specific case of Portugal, although a growing number of people self-identify as non-religious, the vast majority (more than 80% of the population) are catholic (INE, 2022). Given that religion is consistently correlated with conservative ideologies across countries and religious affiliations (see Caprara et al., 2018), one might hypothesize that the religious pillar of the 48-year dictatorship (1933–1974) has contributed to what has been suggested as a conservative tilt and authoritarian nostalgia of the Portuguese society (e.g., Ramos & Magalhães, 2021; Santana-Pereira et al., 2016).

Results showed statistical differences in trans acceptance and social tolerance according to political ideology. Those who ascribed to left-wing ideologies were significantly more accepting of, comfortable with, and tolerant towards TGD people than those with right-wing ideologies. Studies suggest that right-wing ideologies are rooted in and/or offer structured views of gender, which are resistant to sociocultural changes (e.g., Meeusen & Dhont, 2015; Whitley & Ægisdóttir, 2000), including sexual and gender diversity (e.g., Hetherington & Jonathan, 2009). No differences were found within the left spectrum (radical left – centre left social democracy) nor within the right spectrum (ultra-right – centre right social democracy – Christian democracy). Self-identified libertarians were the ones who presented higher scores of acceptance and comfort with the gender spectrum, including significantly higher than centre-left social democrats, and with no significant difference with those who endorse radical left ideologies. Also, no differences were found in social tolerance between liberals and those with a radical left ideology, which calls for a more nuanced approach to the relationship between political ideology and trans acceptance and tolerance than the binaries left–right and liberal-conservative, for example, by considering both the economic/fiscal and social/cultural factors underlying political views (e.g., Alves & Porto, 2022; Feldman & Johnston, 2014).

Limitations and Future Directions

Limitations of the current study warrant a cautious interpretation of results. First, although the sample was demographically diverse, participants were recruited through convenience sampling, thus not being representative of the Portuguese population. Most of the sample was composed of cisgender women (67.0%), thus preventing from concluding on the gender measurement invariance of the T-KAB. Future studies should conduct multigroup confirmatory factor analyses to confidently establish its structure invariance for all genders. Also, there was an unequal number of participants in sociopolitical groups, which might have affected statistical power and type I error, thus producing bias due to unequal variances. Additionally, the dropout rate (76.9% of those who accessed the online protocol did not complete it) might have influenced results due to self-selection bias. Future studies should consider participant reimbursement as a strategy to reduce attrition. Finally, although results on religiosity and T-KAB are in line with vast literature suggesting that religious people present higher levels of transphobia, our results should be carefully interpreted. “Religiosity” is a complex construct that entails not only affiliation but also belief (e.g., fundamentalism), scripture interpretation (e.g., literalism), and praxis (e.g., religious attendance). The low representativeness of non-Catholic religious participants in our sample prevents us from conducting a nuanced exploration of religiosity and trans acceptance, comfort, and tolerance.

Conclusion and Social-Policy Implications

The current study corroborates the robust validity of the T-KAB as a measure of trans acceptance, comfort, and social tolerance, suggesting a two-factor solution for its European Portuguese version. Results echo previous studies by suggesting that those who have a lower education, a heterosexual orientation, no contact/less contact with transgender people, a religious affiliation, and/or a right-wing or conservative political ideology present lower levels of acceptance of, comfort with, and tolerance towards TGD people. These results further inform future studies on the potential benefits of developing and implementing tailored psychoeducational programs that target specific sociopolitical segments of society to reduce anti-trans attitudes and promote trans-inclusiveness. Our results tentatively echo the suggestion that the increase of contact with TGD people, for example by highly religious people, may contribute to reduce their negative attitudes towards TGD people (Kanamori & Xu, 2022). Nonetheless, future research should focus on the type of contact that is more effectively conducive of attitudinal changes towards TGD individuals (e.g., information-based: focused on psychoeducational content, such as the distinction between sex and gender; or attitudinal-focused: promoting a compassionate stance towards TGD experiences).