Australian Mental Health Care Practitioners’ Cultural Competence, Therapeutic Alliance, Desirable Responding and Racial/Ethnic Attitudes

Racial, ethnic, religious, and cultural diversity in Australia is rapidly increasing. Although Indigenous Australians account for only approximately 3.5% of the country’s population, over 50% of Australians were born overseas or have at least one migrant parent. Migration accounts for over 60% of Australia’s population growth, with migration from Asia, Sub-Saharan African and the Americas increasing by 500% in the last decade. Little is known about Australian mental health care practitioners’ attitudes toward this diversity and their level of cultural competence. Given the relationship between practitioner cultural competence and an effective therapeutic alliance with diverse clients, this study aimed to identify factors that inuence non-White and White practitioner cultural competence and therapeutic alliance. An online questionnaire was completed by 139 Australian mental health practitioners. The measures included: the Multicultural Counselling Inventory (MCI); the Color-blind Racial Attitudes Scale (CoBRAS); and the Balanced Inventory of Desirable Responding (BIDR). Descriptive statistics were used to summarise participants’ demographic characteristics. One-way ANOVA and Kruskal-Wallis tests were conducted to identify between-group differences (non-White compared to White practitioners) in cultural competence, therapeutic alliance, and racial and ethnic blindness. Correlation analyses were conducted to determine the effect of participants’ gender or age on cultural competence and therapeutic alliance. Hierarchical multiple regression analyses were conducted to predict cultural competence and therapeutic alliance. older age, greater attendance of cultural competence-related trainings and increased awareness of general and pervasive racial and/or ethnic discrimination. Practitioners with higher MCI total scores were also likely to have higher self-deceptive positive enhancement scores on the BIDR than those with lower MCI total scores. signicant 125.5, 14.6 M 121.5, p = 0.069). Subscale 4: Multicultural Counselling (M = 33.9, SD 5.4 vs. M = 33.7, SD = 4.5, p = 0.433). Subscale 1: Counselling M p signicant. for Subscale 2: Multicultural Awareness (M = 29.5, SD = 5.0 vs. M = 26.4, SD = 4.6, p < 0.001) and Subscale 3: Multicultural Counselling Relationship (M = 25.4, SD = 3.5 vs. M = 24.2, SD = 3.3, p signicantly in the non-White group compared to group.

of Whiteness clarify who reaps bene ts from Whiteness and those who do not. Hence, in this paper, the term non-White refers to individuals who are excluded from being bene ciaries of Whiteness as a result of their racial, ethnic, cultural, religious, linguistic, or national identities (10).
We recognise that not all Whites encounter the same privileges or to similar levels, or that Whiteness has maintained the same parameters over time, or that it is impermeable and in exible (11,12). Additionally, we recognise that Whiteness (and how those who bene t, or are excluded, from it are de ned) is constantly under revision (13). Even within this consistent state of ux, Whiteness and its consequences endure (14) and are present across all areas of Western (if not global) environments, systems, and practices (15).
One such system is the mental health care system. Mental health is a core element of human health and wellbeing. In 2020, the Australian Institute of Health and Welfare (AIHW) (16) indicated that 45% of all Australians aged 16 to 85 years -8.7 million people -will experience mental illness at some point in their life. According to 2015 statistics, mental illness and substance misuse were the second largest contributor (23%) of the non-fatal burden of disease in Australia (17). During the 2017 and 2018 nancial year, $9.9 billion (16) was spent on mental health, with spending increasing in 2018-2019 to $10.6 billion (18). Given the prevalence and social and economic costs of mental illness, access to and effectiveness of mental health services is crucial. Mental health is therefore an area in which systems and practitioners need to adapt and demonstrate exibility to support clients (19).
However, health practitioners and systems struggle to adapt or to be exible when working with clients who do not t within rigid frameworks of Whiteness in health care (4,5). Importantly, the negative health outcomes and experiences of non-Whites within (mental) health care settings are well documented and related to systems and practitioners that harbor constructions of health and wellbeing based on frameworks of Whiteness (20)(21)(22)(23). For instance, ndings from our systematic review, representing 5,870 mental health care practitioner constructions of non-White people, indicate that practitioners who perceived non-White people as having 'backwards beliefs' about mental health were less likely to engage in culturally competent health care approaches while working with non-White people (20). It is therefore important to consider and identify Australian mental health practitioners' attitudes to Whiteness and their impact on cultural competence and the therapeutic alliance. In doing so, advances can be made in practitioner training and clients' mental health outcomes.
In Australia, little is known regarding mental health workers' cultural competence or attitudes towards non-White people and how these factors in uence the therapeutic alliance. Contemporary understandings of health care provider attitudes can be implied from well documented negative health experiences and outcomes within minority populations and across variety of health care settings (24). Overall, the crux of these experiences are linked to health systems and providers internalising Anglo-centric and racist constructions of health and wellbeing (25). As a solution, practitioners and researchers advocate for anti-racism and improved cultural competency (15). Cultural competency is de ned as "a set of congruent behaviours, attitudes and policies that come together in a system, agency or among professionals and enable that system, agency or those professionals to work effectively in cross-cultural situations" (27). International evidence indicates that when practitioners internalise and enact the principles of cultural competence, they are able to cultivate and maintain a positive therapeutic alliance between themselves and their clients (15,(28)(29)(30).
Therapeutic alliance is broadly de ned as "the collaborative and affective bond between therapist and patient … [and] is an essential element of the therapeutic process" (31). The therapeutic alliance is exempli ed by bi-directional and mutual collaborations between mental health care providers and clients, are dependent on a person-centred health care ethos (32) and based on a mutually nurtured psycho-socio-emotional connection (33). With, the concept of the therapeutic alliance has its origins in early psychoanalytic theories (e.g., 34) and is now a staple of contemporary evaluations and conceptualisations of the therapeutic process. Notably, international evidence indicates that the therapeutic alliance is strongly associated with clients' mental health engagement and outcomes (31,35).
Given the limited literature on the topic speci c to the Australian context this study aimed to identify the factors that in uence practitioner constructions of non-White people and their relationship to practitioners' perceived cultural competence and therapeutic alliance. We therefore asked the following research questions: Additional recruitment took place via social media including Facebook, Twitter, Instagram, and LinkedIn. Participants were invited to complete the survey online or request a paper copy.
One hundred and eighty-two practitioners commenced the survey. The majority engaged in the survey online (n=175) while seven completed the survey in hardcopy. After screening for completeness of all the demographic, MCI, CoBRAS and BIDR items, 139 participants remained. The study was approved by and conducted within the guidelines of the Human Research Ethics Committee (Approval Number 2017/105) at the University of Wollongong.

Quantitative Survey Measures
The survey included the following measures:

Demographic Questionnaire
The demographic questions collected information about participants' age, gender, ethnicity and country of origin, highest degree achieved, year highest degree achieved, mode of professional practice/training (private and/or public), years of experience and engagement with post-quali cation cultural competency training. Participants were also asked where they completed or are engaged in their psychology/counselling training and how many non-White clients they see in a month.

Multicultural Counselling Inventory (MCI)
This 40-item self-report inventory (36) assesses behaviours and attitudes related to four multicultural competencies on a 4-point Likert scale from very inaccurate (1) to very accurate (4). For each item, a score of 1 indicates low multicultural competence and a score of 4 indicates high multicultural competence (37). Scale scores are obtained by adding the items speci c to each subscale. The MCI total score ranges from 40 to 160 with higher scores indicating greater multicultural competence. Higher subscale scores also indicate greater multicultural competence in the respective subscale areas (see below). As explained in Ottavi et al. (38), the four areas of multicultural competency are as follows: 1. Skills-11 items measuring general counselling and speci c multicultural counselling skills. Sample items include "When working with all clients, I am able to be concise and to the point when re ecting, clarifying, and probing" and "When working with minority clients, I monitor and correct my defensiveness." Minimum score = 11, maximum score = 44.
2. Knowledge-11 items measuring treatment planning, case conceptualization, and multicultural counselling research. Sample items include "When working with minority clients, I keep in mind research ndings about minority clients' preferences in counselling" and "When working with minority clients, I apply the socio-political history of the clients' respective minority groups to understand them better." Minimum score = 11, maximum score = 44.
3. Awareness-10 items measuring multicultural sensitivity, interactions, and advocacy in general life experiences and professional activities. Sample items include "I am involved in advocacy efforts against institutional barriers in mental health services for minority clients" and "When working with international students or immigrants, I understand the importance of legalities of visa, passport, green card and naturalization." Minimum score = 10, maximum score = 40.
These three subscales allowed us to measure cultural competence in line with similar principles in the cultural competence model developed by Campinha-Bacote (39).
4. Relationship-8 items measuring the counsellor's interaction process with the minority client (e.g., comfort level, worldview, and counsellor's trustworthiness). Sample items include "When working with minority individuals, I am con dent that my conceptualization of individual problems do not consist of stereotypes and biases" and "When working with minority clients, I perceive that my race causes the client to mistrust me." Minimum score = 8, maximum score = 32.
We used this subscale to measure therapeutic alliance in line with principles of this concept as outlined in Escudero,Friedlander (40

Color-Blind Racial Attitudes Scale (CoBRAS)
This study used an adapted version of the Color-Blind Racial Attitudes Scale (41) to assess practitioners' constructions of non-White people. The CoBRAS consists of 20 items to assess attitudes with a 6-point Likert scale of 1 (strongly disagree) to 6 (strongly agree). The CoBRAS is designed to assess cognitive dimensions of colour-blind racial attitudes including the degree to which respondents distort, deny, and/or minimize the existence of institutional racism (e.g., "Everyone who works hard, no matter what race they are, has an equal chance to become rich," and "Racism may have been a problem in the past, but it is not an important problem today"). Total scores can range from 20 to 120, with higher scores representing greater colourblind racial beliefs (41)  Although the CoBRAS is based on the U.S. context and uses terminology relevant to racial dynamics in the U.S., many of these dynamics (e.g., institutional racism and systemic discrimination) are also present within the Australian context. As such, the content of the items did not require amendment, however their context 'American' versus 'Australian' or 'United States' versus 'Australia' or 'African American' versus 'Afro-Australian' did require revision. In place of the term 'White' the term 'Anglo-Australian' was used as many individuals who identify as being of an ethnic minority may appear White but may experience prejudice or discrimination due to other manifestations of their ethnicity (wearing a hijab, having a non-Australian accent, or having non-Anglo family members for example). Finally, for those items which only mention race, and not race and ethnicity, the word ethnicity or ethnic has been added to re ect the realities of multiculturalism in Australia where ethnicity may be mutually exclusive of race (e.g., "Racial problems in the U.S. are rare, isolated situations" was changed to "Racial problems in Australia. are rare, isolated situations". American spellings (e.g., color versus colour) were also amended in line with Australian spelling conventions.
Neville et al. (41) reported that the coe cient alpha for the total CoBRAS was 0.91. Our study demonstrates the coe cient alpha for the total CoBRAS

Balanced Inventory of Desirable Responding
The Balanced Inventory of Desirable Responding (BIDR) (42) measures the tendency to respond and exhibit behaviours or thoughts that are viewed as socially desirable but are not accurate representations of an individual's real attitudes (43). Scholars in counselling have used BIDR to control counsellors' social desirability in their self-report data (44). Sample items include "My rst impressions of people usually turn out to be right" and "I always know why I like things." BIDR consists of 40 items with two subscales of 20 items each evaluating impression management and selfdeception. The scale is composed of a response format consisting of a 7-point Likert scale ranging from 1 (not true) to 7 (very true); 1 point is scored for each extreme answer (6 or 7) with a total score ranging from 0 to 40. Higher scores indicate a greater tendency to respond and exhibit behaviours or thoughts that are viewed as socially desirable. BIDR has been used successfully with various racial/ethnic and cultural groups (44). Chao (45) reported a coe cient alpha of 0.85 with a sample of graduate students in psychology and mental health professions. Our study reports the coe cient alpha of total score was 0.88, Subscale 1: Self Deceptive Enhancement was 0.73, and Subscale 2: Impression Management was 0.86.

Data Analysis Strategy
Sample size was estimated to be N=86 (at 90% power), adequate for anticipated effect size on the basis of previously published literature (46). A sample of N=139 was recruited in this study, which also appears to be appropriate according to Cohen's guideline which suggests N=97 is adequate for medium effect size (47). Reliability analyses were also conducted to assess internal consistency of each measure and to ensure their reliability within this novel sample. Descriptive statistics were used to summarise participants' demographic characteristics and the primary dependent variables. Kolmogorov-Smirnov test and Shapiro-Wilk test were conducted to check the normality of the data. One-way ANOVA and Kruskal-Wallis test were performed to identify the between-group (White and non-White practitioners) differences of the perceptions of cultural competence, therapeutic alliance, and racial and ethnic blindness. A correlation analysis was conducted to determine the effect (if any) of participants' gender or age on cultural competence and therapeutic alliance (MCI total score). These analyses addressed research questions 1 and 2.
A hierarchical multiple regression analysis was performed to predict practitioner cultural competence and therapeutic alliance (MCI total score). In the rst step, social desirability (measured by BIDR subscales) was entered. In the second step, participants' age, ethnicity (for the two groups-White and non-White[1]-entered with dummy codes), and participants' speaking a language other than English were entered. In the third step, cultural competency variables including the number of cultural competence-related workshops, conferences or training sessions attended since beginning practicing mental health care, number of non-White clients [2] seen a week, and whether their formal training prepared them to work with non-White clients were entered. In the fourth step, racial and ethnic blindness attitudes (measured by CoBRAS subscales) were entered. This analysis addressed research questions 3 and 4.
[1] Participants were asked to select any and all relevant ethnic identi ers including Asian, Paci c Islander, African, Latin American, Western European, Eastern European, Aboriginal and/or Torres Strait Islander, Indigenous Canadian and/or American, Middle Eastern, Mediterranean and Caucasian. Those who chose Caucasian and/or Western European were categorised as White for the purposes of analysis in line with our previous research in which participants of Western European descent self-identi ed as White. [2] To ensure conceptual clarity survey participants were presented with the de nition of non-White people indicated in the Background section of this paper. Table 1 summarises the demographic, professional, and cultural competency related characteristics of the participants. The mean age of the participants was 37.3 (SD = 11.3) years, with majority being female (89.2%). Most of the participants were White (64.7%), and an Australian citizen (91.4%). Most participants had completed a postgraduate degree (80.6%) and had completed their highest level of academic quali cation within last ve years (59.7%). All participants received their academic quali cations from Commonwealth countries, and the majority did not speak a language other than English (66.2%). Of the 139 participants, 76.7% were psychologists with varied duration working in mental health and types of practice. In terms of cultural competency related characteristics, most participants (93.5%) were seeing at least one non-White client per week, 82% of the participants had attended cultural competence-related workshops, conferences, or training sessions since the beginning of their mental health care practice, and over half of the participants (51.5%) agreed that their qualifying mental health training su ciently prepared them to provide culturally competent services for non-White clients. Table 1 Practitioners' demographic, professional, and cultural competency related characteristics (n = 139). To examine whether the dependent variables (cultural competence and therapeutic alliance) varied as a function of participants' gender and age, we conducted an analysis of variance (ANOVA) of gender and a correlation analysis of age with the dependent variables. The ANOVA results revealed no signi cant main effects for participants' gender on MCI total score, F(2,139) = 0.863, p = 0.424. A correlation analysis showed that age was weakly but statistically signi cantly correlated with the total score of MCI (r = 0.183, p = 0.032) and thus was considered a predictor variable in the nal analysis.

Differences In Desirable Responding Across White And Non-white Practitioners
The average total BIDR score of all participants was M = 10.4 (SD = 6.8) ( Table 2). The total score was not statistically different across the two groups despite the non-White group having a higher score than the White group (M = 11.8, SD = 7.2 vs. M = 9.6, SD = 6.3, p = 0.094). This nding suggests that the groups did not differ in terms of desirable responding.

Differences in Cultural Competence and Therapeutic Alliance across White and non-White Practitioners
The average total MCI score of all participants was M = 122.9 (SD = 14.0) ( Table 2

Difference of Color-Blind Racial Attitudes Scale (CoBRAS) across White and non-White Practitioners
The average total CoBRAS score of all participants was M = 43.8 (SD = 13.6) ( Table 2 Table 3 summarises the variables predicting participants' cultural competence and therapeutic alliance. Self-deceptive enhancement (SDE: Subscale 1 of BIDR), the age of participants, attendance of cultural competence-related workshops, conferences, or training sessions since the beginning of mental health care practice, and Unawareness of Blatant Racism Issues (UBRI: Factor 3 of CoBRAS) were observed to contribute signi cant variance to multicultural competence. The total proportion of variance for the dependent variable, which is explained by the independent variables in the model at

Predictors Of Cultural Competence And Therapeutic Alliance
Step 4, was calculated as total R 2 = 0.374. and R 2 change = 0.065. The ethnicity of the participants (White and non-White) and the ability of the participants to speak a language other than English did not contribute to variance in cultural competence and therapeutic alliance. Similarly, in Step 3, variables related to practitioners' cultural competence experience were added to social desirability (BIDR subscales), age, ethnicity, ability to speak another language other than English. Cultural competence experience variables included: attendance at cultural competence-related workshops, conferences, or training sessions; number of non-White clients seen per week; and perceived preparedness to work with non-White clients following qualifying mental health training.
Step 3 yielded additional signi cant variance in cultural competence and therapeutic alliance, F (3,124) change = 4.102, p = 0.008, R 2 = 0.278, and R 2 change = 0.072. However, the number of non-White clients per week and the belief that professional mental health training su ciently prepared them to provide culturally competent services for non-White clients did not contribute any signi cant effect on cultural competence and therapeutic alliance. In the nal step (Step 4), the three subscales of CoBRAS were added in addition to all the variables in the previous steps. The analysis revealed that Unawareness of Blatant Racism Issues, together with the variables added in Steps 1 to 3, contributed to additional signi cant variance in cultural competence and therapeutic alliance with a large effect size, F (3,121) change = 6.142, p < 0.001, R 2 = 0.374, and R 2 change = 0.095.

Discussion
The purpose of this study was to identify the factors that in uence Australian mental health practitioners' cultural competence and therapeutic alliance. The ndings indicate that practitioners in this study valued cultural competence training given their high level of engagement in this type of continuing professional development. This may be in response to the fact that the vast majority engaged with at least one non-White client a week and wanted to maintain and/or improve their ability to provide culturally competent services to their non-White clients (24,39). lives, resulting in a high level of awareness of Whiteness and its impact on non-White people (30,(49)(50)(51). Consequently, non-White people are more likely to build a more robust connection with non-White clients due to their increased ability to be empathetic to their clients' experiences particularly with regards to discrimination-related distress (49,52,53).
In line with international samples of mental health practitioners, the current sample demonstrated low to moderate levels of colour-blind racial and ethnic attitudes with average CoBRAS total scores of M = 43.8 (SD = 13.6). For instance, Neville et al. (41) reported that their sample of 152 American mental health workers and trainees "held low to moderate levels of color-blind racial beliefs" (p. 483) with mean CoBRAS score of M = 48.59 (SD = 12.79). Interestingly, non-White practitioners in the current study were more likely to be in denial about racial privilege and the pervasiveness of racial/ethnic issues in Australia compared to their White counterparts. This may be because many White people from low socioeconomic backgrounds experience signi cant systemic disadvantages as a result of the pressures of Whiteness (11,12,54). Non-White practitioners working with a diversity of clients may therefore see parallels between White and non-White clients from low socioeconomic backgrounds and be less likely to attribute disadvantage to experiences of racial/ethnic discrimination (15). Another plausible explanation may be that non-White practitioners may be responding according to perceptions of their own privileges and the advantages they experience given their high levels of education, career capital and social positioning. Additionally, non-White practitioners may not want to buy-in to the idea that Australia suffers from pervasive and insurmountable institutional racism as this may hinder their own wellbeing and ability to encourage their non-White clients towards a ful lling life.
The results of the hierarchical multiple regression indicate that highly favourable perceptions of self and protection of self-esteem (BIDR Subscale 1: Self Deceptive Enhancement SDE) was positively associated with practitioner cultural competence and therapeutic alliance (MCI total score). As such, an increase in the levels of self-deceptive enhancement, or an honest but overly positive responding, was found to be linked with greater multicultural competence. This may be because practitioners who perceive themselves more positively may have a higher sense of con dence in their abilities and therefore rate themselves as more culturally competent (39). It could also be that participants protected their self-esteem due to the fear of being perceived as a culturally incompetent practitioner. This fear may be so strong, that practitioners may be inclined to even engage in self-deception to avoid it. These interpretations align with other literature which indicates the identity threat experienced by mental health care providers when confronted with information about Whiteness and their possible role in its perpetuation (45,(55)(56)(57).
The relationship between social desirability, cultural competence, and therapeutic alliance was unexpected given previous research which found no signi cant association between these variables. While some researchers (37) proport that social desirability is an important issue in relation to selfreport in cultural competence, others have not found any relationship between these variables (41,45). Following Chao (2013), the BIDR was included in this study. However, unlike Chao (2013) or Neville (2006) who found no relationship between these variables, the current study found a signi cant relationship between the self-deceptive enhancement subscale and higher MCI scores.
The hierarchical analysis also showed that participant's age as well as the number of cultural competence-related workshops, conferences, or training sessions attended since the beginning of mental health care practice were found to be positively associated with cultural competence and therapeutic alliance. This nding suggests that cultural competence and therapeutic alliance grows with clinical experience, regular engagement in training and maturity as a practitioner.
The hierarchical multiple regression also revealed that higher levels of unawareness of general and pervasive racial/ethnic discrimination (CoBRAS Factor 3: Unawareness of Blatant Racism Issues (UBRI) are negatively associated with cultural competence and therapeutic alliance (total MCI score).
This nding mirrors those in other studies (41,58,59) and highlights that addressing race and racism is central to cultural competence training. Notably, the lack of training on race and racism may be why 48.9% of the sample were either neutral or disagreed that their qualifying training had prepared them to work with non-White people.
In summary, the study results demonstrate that higher cultural competence and therapeutic alliance are associated with being non-White, practitioner's age, greater attendance of cultural competence-related trainings and increased awareness of general and pervasive racial and/or ethnic discrimination.
Practitioners with higher cultural competence and therapeutic alliance are likely to perceive of themselves more favourably than those with lower levels of competence.

Implications And Recommendations For Training And Practice
The ndings of this study reinforce the need for continued engagement in cultural competence training across the full trajectory of a practitioner's mental health career. The current study reiterates the importance of continuously addressing Whiteness by explicitly discussing White supremacy, privilege and power (for example) in Western mental health care settings and clearly identifying their impact on all people, not only those identifying as non-White (60-62). This re ects a growing national focus on key components of cultural competence training and implementation, namely multicultural awareness, and relationships. For instance, the Australian Heath Practitioner Regulation Agency (AHPRA) encourages health practitioners not only to learn about why difference is a signi cant factor in therapeutic relationships but also how to engage with diverse clients in culturally competent ways (63). While their current guidelines are focused on Indigenous Australians, many of the principles apply to other non-White groups. Given the lack of cultural competency guidelines from AHPRA relevant to the many other non-White groups in Australia, the development of such training and best-practice is needed to ensure mental health practitioner's preparedness to work with diverse populations. This recommendation is reinforced by the fact that cultural competence, a precursor to an effective therapeutic alliance, is an integral part of all 15 health professions regulated by AHPRA and their professional codes of conduct.
Training should, however, be focused on addressing the particular needs of practitioners at various levels of experience and consider practitioner race and ethnicity. Pragmatically, practitioners could, for example, begin by completing demographic items as well as questions related to their existing counselling competence and therapeutic alliance skills. Their responses would then determine allocation to training relevant to their age, racial/ethnic background as well as their professional experiences and expertise. For instance, younger, less experienced, White practitioners could engage in training that included discussions about Whiteness and anti-racism in mental health practice and services. Additionally, those practitioners with higher levels of cultural competence and therapeutic alliance can work on developing, implementing, or enhancing therapeutically safe mental health practises and those with lower levels can be introduced to the concepts and implications for their practice.
Aligned with the study ndings, mental health care practices should actively engage with non-White populations to increase opportunities for cultural encounters between mental health practitioners and non-White clients (64, 65). This can be done by offering home visits or providing services within community centres. Such engagement activities can also serve to decrease ethnic and racial colour-blindness amongst practitioners and offer opportunities for clients to develop effective therapeutic alliance with mental health care providers (56). Given that many mental health practitioners, particularly psychologists, work in one-on-one settings engagement with their professional networks, case conferencing and clinical practice supervision can assist in increasing their awareness of general and pervasive racial and/or ethnic discrimination experienced by their clients (66-68). Such forums would be especially effective where topics of race and Whiteness are directly addressed so that practitioners can be regularly challenged by their peers to reduce their racial and ethnic blindness and minimise the harm of practitioner ignorance on non-White clients.

Limitations And Future Research
There are limitations to this study. First, the majority (76.7%) of the sample were psychologists which may limit the generalisability of the results to other mental health practitioners given that psychologists often work one-to-one with mental health clients. Further research including a more diverse sample of mental health practitioners would provide a more robust representation of cultural competence and therapeutic alliance within the mental health sector. In addition, because of the limited number of non-White practitioners from diverse groups they were combined into a single group.
Consequently, the results cannot be generalised to any particular group of non-White practitioners (e.g., Asian Australians, African Australians, Indigenous Australians). Research including a more diverse sample of mental health professionals may provide more information about the relationships between cultural competence, therapeutic alliance and other variables like sex, age, duration of practice, and ethnicity. With a more diverse sample between groups difference can be explored more rigorously.
Second, the collected data drew on mental health practitioners' self-report. It is well documented that self-report may not re ect actual levels of cultural competence (69-71). Given that this study found a positive correlation between self-deceptive enhancement and cultural competence, additional methods for determining practitioner cultural competence and therapeutic alliance are needed. This may include a comparison between clients' perceptions of their therapist's cultural competence versus their practitioners' self-perceptions (72)(73)(74). Further, observations and/or analysis of recordings of mental health sessions can be used to determine the provider's cultural competence. While these recommendations present their own methodological limitations, they may be more in line with a culturally competent approach (75). Self-report may also contribute to the surprising relationships found between socially desirable responding and cultural competence and therapeutic alliance. As such, future studies with larger and more diverse samples should examine the association between social desirability and cultural competence and therapeutic alliance.

Conclusion
This is the rst Australian study to explore the complex relationships among mental health care practitioners' sociodemographic variables, cultural competence, therapeutic alliance, multicultural training, colour-blind racial attitudes, and social desirability. Additionally, it is one the few studies to use a series of regressions to analyse moderating variables and their effect on cultural competence and therapeutic alliance. The study sheds light on the attributes of mental health practitioners in Australia working with diverse clients while also navigating their own diversity and difference. The ndings highlight that the current one-size-ts-all and skills-development approach to cultural competence training ignores the signi cant role that practitioner diversity and differences play in the therapeutic alliance. The recommendations from this study can inform clinical educators and supervisors about the importance of continuing professional development relevant to practitioners' age, professional experience, and ethnic/racial background.