Exploring the supervisory relationship in the context of culturally responsive supervision: a supervisee’s perspective

Abstract

Clinical supervision is a relationship-based education, considered crucial in providing clinicians with emotional support, skill development and improving client outcomes. Culturally responsive supervision assumes that culture permeates clinical practice and supervision. Culturally responsive supervisors promote the development of cultural competence in supervision, through modelling, reflective discussion and responsivity. Research has demonstrated that greater perceived cultural responsivity in supervision may result in greater satisfaction for supervisees, particularly those from racially or ethnically minoritised (REM) backgrounds. The current study explores supervisee perceptions of culturally responsive supervision and supervisory relationships between different supervisory dyads, comprising supervisees from REM and White backgrounds. This was a cross-sectional design incorporating a between-groups comparison. Trainee and qualified clinical psychologists, counselling psychologists and CBT therapists (n = 222) completed an online survey. Perceptions of cultural responsivity and the supervisory relationship were explored. Participants provided information about their supervisor’s race and ethnicity and their own, and were organised into four supervisory dyads. Participants from REM backgrounds in dyads with White supervisors perceived their supervision as significantly less culturally responsive, with significantly lower quality supervisory relationships. Greater perceived cultural responsivity in supervision significantly predicted better supervisory relationships (regardless of participant cultural background). Findings suggest that culturally responsive supervisory practices may play an important role in developing cultural competence and strengthening the supervisory relationship, particularly in cross-cultural supervisory dyads. Findings present important clinical and theoretical implications.

Key learning aims

1. (1) To understand the need for cultural responsivity within the context of clinical supervision.

2. (2) To explore the differences between cross-cultural and culturally similar supervisory dyads in perceptions of cultural responsivity in supervision.

3. (3) To explore the differences between cross-cultural and culturally similar supervisory dyads in perceptions of the quality of the supervisory relationship.

4. (4) How does culturally unresponsive supervision impact supervisee experiences?

Introduction

Clinical supervision aims to help supervisees develop and integrate therapeutic and professional skills into clinical practice, to improve client outcomes (Fleming and Steen, 2012). It is considered crucial in developing vital clinical skills, such as cultural competency among mental health professionals (Falender and Shafranske, 2004). It can enable clinicians to better understand culturally diverse clients and establish effective therapeutic relationships, leading to better therapeutic outcomes (Gainsbury, 2017). Culturally responsive supervision assumes that the practitioner’s cultural background and experiences permeate their clinical practice and clinical supervision (Arthur and Collins, 2009). Culturally responsive supervisors promote cultural competence in supervision through responsivity, reflective discussion and modelling their own cultural competence (Ancis and Marshall, 2010; Burkard et al., 2006; Sue and Sue, 2008). Cultural competence can be defined as a practitioner’s acquisition of cultural awareness, knowledge, and skills required to provide effective and responsive treatment for all cultural groups (Sue and Sue, 2008).

Culture is a heterogeneous construct, often conceptualised by numerous contextual variables (e.g. race, gender, class, age) (Small et al., 2010). This paper focuses specifically on race and ethnicity when referring to aspects of culture, as these aspects remain largely under-researched in supervision literature; in particular, the quality of supervision when there is a visible difference between the supervisor and supervisee compared with when there is not (Sukumaran, 2016).

Race and ethnicity can influence individual worldviews and beliefs, shaping how problems are perceived and interpreted by clients, clinicians and services (Gainsbury, 2017). Hence, it is imperative that services offer culturally responsive support, consistent with the values and unique life experiences of racially or ethnically minoritised (REM) individuals (Patel and Keval, 2018). The term ‘minoritised’ provides a social constructionist approach to understanding that social processes shaped by power are responsible for ‘minoritisation’ and that individuals do not naturally exist as racial or ethnic minorities (Gunaratnam, 2003; Predelli et al., 2012).

The existing research literature repeatedly highlights the supervisory relationship (SR) as the most important factor for successful supervision, regardless of the supervision model adopted (Beinart, 2012; Inman and Ladany, 2008; Ladany et al., 1999; Milne, 2009). Bordin’s (1983) Model of the Working Alliance suggests three aspects contribute to the supervisory relationship: mutual agreement on supervision goals, mutual agreement on tasks to achieve goals, and the bond between supervisor and supervisee. Bordin suggested that shared experience and fostering trust within the supervisory relationship can aid development of bonds. In support, Clohessy (2008) proposed relational factors such as trust influence openness and investment into the supervisory relationship. Contextual factors such as supervisor and supervisee values and cultural identity can shape the flow of supervision and the development of trust and safety in the supervisory relationship.

Existing developmental supervision models based on psychological theories often fail to explicitly mention the importance of cultural variables within the supervisory process (Banks, 2001; Patel, 2011). Hawkins and Shohet (2006) proposed ‘The Seven-Eyed Model’ of supervision, which integrates relational and systemic aspects of supervision into a model. While this is one of the few models to reference the importance of cultural variables on the SR, it does not explicitly focus on developing cultural responsivity in supervision.

Additionally, the supervisor’s cultural power (power and privileges benefitting some cultural groups over others) often remains unmentioned in supervision literature (Patel, 2011; Ryde, 2000). Power relations may manifest in the SR in different ways. Personal identity and cultural experiences may play a particularly important role here (Clohessy, 2008). For instance, when a REM supervisee is supervised by a White supervisor, differing histories and experiences of privilege and oppression may amplify the power imbalance inherent within the SR. It therefore becomes important to address power and privilege differentials in supervision and their potential impact on the SR (Cook et al., 2018). Addressing power dynamics may foster greater bonds and help build connection and trust within the SR (Bordin, 1983; Clohessy, 2008).

In the United Kingdom, psychological therapists remain predominantly from White backgrounds (Kline, 2014; Turpin and Coleman, 2010). Consequently, REM therapists are more likely to experience cross-cultural supervisory relationships than White therapists.

The empirical literature to date suggests many supervisors may lack cultural awareness, cultural competence and access to appropriate training (D’Andrea and Daniels, 1997). Some studies suggest that White supervisors may experience greater reluctance to bring up cultural issues due to a perceived lack of awareness and cultural competence (Constantine, 1997; Day-Vines et al., 2007; Ladany et al., 1997). Some supervisors may fear offending supervisees by getting things wrong, which can lead to an avoidance of cultural discussions. This may result in supervisees feeling invalidated or the dismissal of cultural concerns brought forward by supervisees (Pieterse, 2018).

Hird et al. (2004) found that White supervisors reported less perceived cultural competence than REM supervisors. However, the latter group spent more time discussing cultural issues in supervision, regardless of their supervisee’s race or ethnicity. Additionally, White supervisors reported discussing cultural issues more with REM supervisees than with White supervisees (Schroeder et al., 2009).

To develop cultural responsivity, practitioners must reflect on and confront personal cultural norms, values, assumptions and biases in supervision (Patel, 2011). Research demonstrates that higher perceived levels of supervisor cultural responsivity may lead to more cultural discussion and self-disclosure by supervisees (Hutman and Ellis, 2020; Mori et al., 2009), better case conceptualisation and better treatment outcomes (Inman, 2006).

Wilcox et al. (2022) found supervisees rated greater satisfaction with supervision when supervisors demonstrated greater cultural humility. Greater satisfaction was also reported when supervisors attended to opportunities for cultural discussion, directly relating to supervisees. Similarly, a qualitative meta-analysis of 29 studies revealed that supervisees are more satisfied in supervision when supervisors demonstrate cultural sensitivity (Coleiro et al., 2022).

Likewise, Burkard et al. (2006) found that REM and White supervisees receiving culturally responsive supervision felt encouraged to further explore cultural issues, with positive effects on the SR and client outcomes. However, REM supervisees reported more cultural unresponsiveness and adverse effects than White supervisees.

Duan and Roehlke (2001) found that supervisors reported making more efforts to address cultural issues than supervisees perceived. In contrast, supervisees said that they had a greater sensitivity to cultural issues than their supervisors. These findings imply that direct and explicit engagement in cultural competence development in supervision may be required.

Despite the importance of this area, there is a distinct lack of UK research exploring supervisees’ perspectives of culturally responsive supervision and its impact on the SR (Patel, 2011). Many studies to date (e.g. Burkard et al., 2006; Constantine and Sue, 2007) have originated in North America, making it difficult to generalise findings to the UK’s differing demographics, social and cultural histories (Cherry et al., 2000). Furthermore, the research has overwhelmingly focused on supervisory dyads consisting of a White supervisor and a REM supervisee from counselling backgrounds (Ladany et al., 1997).

The current study aims to explore supervisees’ perspectives of culturally responsive supervision within the UK. It is the first study to explore relationships between the supervisee’s perspectives of culturally responsive supervision and the perceived quality of the SR within cross-cultural and culturally similar supervisory dyads. Supervisees provided information about their current supervisor and were placed in ‘supervisory dyads’ consisting of White supervisees (WSE) and White supervisors (WSR); REM supervisees (REMSE) and WSR; WSE and REM supervisors (REMSR); and REMSE and REMSR. This study considered qualified clinicians and trainees from clinical and counselling psychology and cognitive behavioural therapy (CBT) backgrounds. It is hoped that this research may provide a deeper insight into the current culturally responsive supervisory practices within the UK, informing future practice and training.

It was hypothesised that higher levels of perceived supervisor cultural responsivity would be associated with higher levels of satisfaction with the supervisory relationship (irrespective of the supervisees’ cultural background). A hierarchical multiple linear regression model was also used to explore predictors of greater quality SRs for REMSE and WSE.

Method

Design

A cross-sectional design incorporating a between-groups comparison was used. An online survey was used to collect questionnaire data from trainee and qualified clinical psychologists, counselling psychologists and CBT therapists who met the eligibility criteria and consented to take part in the study.

Participants

Participants were recruited from across the UK from: NHS sites, professional training programmes, professional bodies and social media sites. The inclusion criteria were as follows: (1) trainee or qualified: clinical psychologists, counselling psychologists, or CBT therapists; (2) aged over 18 years; (3) currently working in clinical practice within the UK; (4) receiving regular clinical supervision (minimum of once a month for qualified supervisors); (5) receiving supervision from their current or most recent supervisor for longer than 4 months.

Measures

Demographic data

Demographic data for participants’ age, gender, ethnicity and information about their current supervisor’s ethnicity were collected and based on self-report. If their supervisor’s racial/ethnic identity was unknown, supervisees were asked to take a best guess.

The Race-Ethnicity Supervision Scale (RESS) (Burkard and Hartmann, 2012)

The RESS is a 29-item self-report measure that examines supervisee perspectives of culturally responsive and unresponsive supervisory practices. It is based on supervisee ratings within four domains of culturally responsive supervision, on a seven-point Likert scale (1, never; 4, neutral; 7, always). Domains include: (1) promoting supervisee race-ethnicity cultural competence, (2) development and responsivity to cultural identity within supervision, (3) perceived supervisor cultural competence, and (4) harmful supervisory practices. The RESS demonstrates good internal consistency and reliability; the total scale Cronbach’s alpha coefficient was 0.97 (Bartell, 2016). In the current study, the scale was also found to have a strong internal consistency of α = 0.97.

The Short Supervisory Relationship Scale (S-SRQ) (Cliffe et al., 2016)

The S-SRQ is an 18-item self-report scale that assesses supervisee’s perspectives of their supervisor based on three subscales: safe base, reflective education and structure. Participants rate their agreement with statements on a 7-point Likert scale from ‘strongly disagree’ to ‘strongly agree’. The scale has previously shown a strong internal consistency of α = 0.96 and has been validated for use with UK trainee clinical psychologists. In the current study, the scale was found to have a strong internal consistency of α = 0.93.

Procedure

The present study employed a cross-sectional design using online-based self-report measures. A live link was shared within recruiting correspondence inviting participants to complete an online survey on Qualtrics XM survey software. Initially, participants were directed to an online participant information page and consent statement before proceeding with the study. Participants then completed a demographic questionnaire. Next, participants were asked to take part in a short visualisation exercise asking them to think about their current supervisor and current experience of supervision. It was hoped that bringing their supervisor to mind would encourage greater accuracy within self-reported measures (McAvinue and Robertson, 2007). The participants completed measures in the following order: S-SRQ, RESS, followed by questions regarding personal views towards culturally responsive supervision. Following this, participants were presented with an online debriefing statement, ending their participation. Participants were given the option to ‘opt-in’ to a prize draw of vouchers.

Results

Demographic characteristics

A total of 231 participants took part in the study; of these, nine participants’ data were removed for the following reasons: six did not select their job role; one did not provide information about their ethnicity; and two only received group supervision. Table 1 displays the included number of participants across professional roles, whilst Tables 2 and 3 display participant demographic information.

Table 1. Participant characteristics across professional roles

Table 2. Participant ethnicity and group identity

Table 3. Participant demographic characteristics by supervisory dyad

Preliminary data analyses

Data analyses were conducted using SPSS (version 27; IBM Corporation). Exploration of the data revealed the presence of significant outliers and a non-normal distribution. Data were also unbalanced in each of the four supervisory dyads, with a smaller sample for ‘WSE–REMSR’ and ‘REMSE–REMSR’ supervisory dyads, as expected given the disproportionate under-representation of REM supervisors within the profession (Turpin and Coleman, 2010). For these reasons, a non-parametric, Kruskal–Wallis test was selected for analysis (Field, 2018).

Differences in the perceived quality of the SR between supervisory dyads

A Kruskal–Wallis test showed a significant difference in S-SRQ scores between the four supervisory dyads, H(3) = 22.82, p = <0.001. Pairwise comparisons of supervisory dyads suggested significant differences in S-SRQ scores (after Bonferroni adjustment) between REMSE-WSR and WSE-WSR supervisory dyads (p = <0.001) and WSE-REMSR and REMSE-WSR supervisory dyads (p = <0.05). The effect size was calculated as d=Z√N (Pallant, 2007). No other comparisons were significant after Bonferroni adjustment (all p > 0.12). Ranked S-SRQ scores can be found in Table 4.

Table 4. Mean ranked S-SRQ and RESS scores across supervisory dyads

Post-hoc Mann–Whitney tests were used to further explore differences in S-SRQ subscale scores between supervisory dyads. Supervisees in the WSE-WSR dyad scored significantly higher than supervisees in the REMSE-WSR dyad in the ‘Safe Base’ subscale, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 3005, z = –4.34, p < .00, d = 0.31; ‘Reflective Education’ subscale, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 3181, z = –3.88, p < .001, d = 0.27 and ‘Structure’ subscale of the S-SRQ, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 3913, z = –2.04, p < .05; d = 0.14.

The tests also revealed that supervisees in WSE-REMSR supervisory dyads had significantly higher scores than REMSE-WSR supervisees, in the ‘Safe Base’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 252, z = –2.08, p < .05, d = 0.22 and ‘Structure’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 187.5, z = –2.92, p < .01, d = 0.31. However, there was not a significant difference between dyads in the ‘Reflective Education’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 262, z = –1.95, p = .051).

Differences in perceptions of cultural responsivity between supervisory dyads

A Kruskal–Wallis test showed a significant difference in RESS scores between the four supervisory dyads, H(3) = 24.76, p = <0.001. Pairwise comparisons of supervisory dyads suggested significant differences in RESS scores (after Bonferroni adjustment) between REMSE-WSR and WSE-WSR supervisory dyads (p = <0.001) and WSE-REMSR and REMSE-WSR supervisory dyads (p = <0.01). The effect size was calculated as d=Z√N (Pallant, 2007). No other comparisons were significant after Bonferroni adjustment (all p > 0.06). Ranked RESS scores can be found in Table 4.

Post-hoc Mann–Whitney tests were used to further explore differences in subscale scores between significantly different supervisory dyads. Supervisees in the WSE-WSR dyad scored significantly higher than supervisees in the REMSE-WSR dyad in the ‘Promotion of Supervisee Cultural Competence’ subscale, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 2786, z = –4.86, p < .001, d = 0.34; ‘Perceived Supervisor Competence’ subscale, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 3201, z = –3.83, p < .001, d = 0.27 and reversed scored ‘Harmful Supervision’ subscale, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 3658.5, z = –3.35, p = 0.01, d = 0.24. However, there was no significant difference between dyads on the ‘Development of Cultural Identity’ subscale, U(N _WSE-WSR = 126, N _REMSE-WSR = 75) = 4187, z = –1.35, p = 0.176).

Supervisees in the WSE-REMSR dyad scored significantly higher than supervisees in the REMSE-WSR dyad in the ‘Promotion of Supervisee Cultural Competence’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 144.5, z = –3.47, p = .001, d = 0.37 and ‘Perceived Supervisor Competence’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 154, z = –3.35, p = .001, d = 0.36. However, there was no significant difference between dyads on the ‘Development of Cultural Identity’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 266, z = –1.91, p = 0.56) and ‘Harmful Supervision’ subscale, U(N _WSE-REMSR = 11, N _REMSE-WSR = 75) = 295, z = –1.74, p = 0.082.

Cultural responsivity and the supervisory relationship

Kendall’s Tau correlation was used to calculate whether there was a correlation between RESS scores and SSRQ scores (Field, 2018). A significant correlation was found between perceptions of culturally responsive supervision and the quality of the SR; greater RESS scores were related to greater SR scores (τ_b = .443, p < .001).

As seen in Table 5, a significant positive relationship was found for greater perceived cultural responsivity predicting greater quality of the supervisory relationship, irrespective of the supervisee’s cultural background.

Table 5. Regression results for predictors of the quality of the SR for REM supervisees

sr ² small effect size = 0.02, medium effect size = 0.15, large effect size = 0.35.

Predictors of higher quality SRs for REMSE and WSE

Hierarchical multiple linear regressions were carried out among the REMSE and WSE samples separately to explore similarities and differences in the predictors of higher quality SRs (e.g. length of time with supervisor and RESS subscales). In the first block, age and gender were entered to control for demographics. The length of time working with their current supervisor was entered in the second block to be controlled for. The final block consisted of all the controlled variables and the remaining RESS subscales (‘Promotion of supervisee cultural competence’; ‘Development of supervisee cultural identity’; ‘Perceived supervisor cultural competence’ and ‘Harmful practice in supervision’).

Results for REM supervisees are shown in Table 5 and indicate a significant regression equation for the final block, (F _6,79 = 23.22, p <.001) with an R ² of .64. Three RESS subscale variables [development of supervisee cultural identity, perceived supervisor cultural competence and harmful practice (higher scores indicate less harmful practice experienced)] were significantly positively associated with SR quality for REM supervisees. Results for White supervisees are shown in Table 6 and indicate a significant regression equation for the final block (F _6,128 = 9.287, p <.001) with an R ² of .30. Perceived supervisor cultural competence appeared to be significantly positively associated with the quality of the supervisory relationship for White supervisees, with a small effect size.

Table 6. Regression results for predictors of the quality of the SR for White supervisees

sr ² small effect size = 0.02, medium effect size = 0.15, large effect size = 0.35.

Discussion

The current study aimed to explore supervisees’ perspectives of culturally responsive supervision within the context of the supervisory relationship. In addition, existing differences between cross-cultural and culturally similar supervisory dyads were further explored. To our knowledge this is the first quantitative UK study to explore supervisee perspectives of culturally responsive supervision. This study found that White supervisees (with White supervisors) self-reported significantly higher quality SRs and cultural responsivity in supervision than REM supervisees (with White supervisors). Additionally, White supervisees (with REM supervisors) self-reported significantly higher quality SRs and cultural responsivity in supervision than REM supervisees (with White supervisors).

REM supervisees (with White supervisors) self-reported significantly less safety and structure in supervision when compared with White supervisees (with White supervisors and REM supervisors). Furthermore, REM supervisees (with White supervisors) perceived their supervisors as significantly less culturally competent and perceived them to be promoting supervisee cultural competence less than White supervisees did. REM supervisees (with White supervisors) also reported experiencing significantly more harmful culturally unresponsive supervision than White supervisees (with White supervisors); however, no significant difference was found between REM supervisees (with White supervisors) and White supervisees (with REM supervisors).

These findings are consistent with a previous study that found REM supervisees were more likely to experience culturally unresponsive supervision, less safety and difficulties within the SR than White supervisees (Burkard et al., 2006). REM supervisees experienced their White supervisors as less culturally responsive, particularly if they did not engage in cultural discussions. This could be due to a lack of awareness, training, or fears of revealing incompetence (Constantine and Sue, 2007; Desai, 2018).

However, if White supervisors appear less willing to engage in critical cultural discussions, REM supervisees may potentially feel unsafe and disclose less, leading to less development and greater dissatisfaction (Patel, 2011). In support, Hutman and Ellis (2020) found supervisee working alliance mediated the relationship between supervisor cultural competence and supervisee non-disclosure. This is further supported by Wilcox et al. (2022) who found supervisors’ cultural humility and willingness to engage in cultural opportunities was associated with greater supervisee satisfaction. Interestingly in the present study, no significant differences were found between dyads on their perceptions of cultural identity development. Due to the overall low self-reported scores across supervisees, it appears that this occurred infrequently in supervision.

Higher RESS scores significantly predicted higher S-SRQ scores, irrespective of supervisee cultural background. In support, Burkard et al. (2006) reported that greater cultural responsivity in supervision might be related to higher quality SRs. Likewise, Coleiro et al. (2022) found that, from the supervisee perspective, helpful aspects of supervision involved greater cultural sensitivity and awareness. Supervisees reported unhelpful aspects of supervision involved a lack of supervisor cultural responsivity.

The regression analyses demonstrated that perceived supervisor cultural competence was a predictor of SR quality for both White and REM supervisees. This suggests that a culturally competent supervisor may benefit supervisees and the SR, regardless of supervisee cultural background. Additionally, for REM supervisees, ‘development of cultural identity’ in supervision and fewer experiences of ‘harmful culturally unresponsive practices’ were significant predictors of SR quality, thus highlighting the need for cultural responsivity in the development of high-quality supervisory relationships for REM supervisees (Clohessy, 2008).

Duan and Roehlke (2001) found that both White and REM supervisees in cross-cultural supervisory dyads felt it was essential for their supervisors to express an interest in their cultural background. Conversely, the current study found that White supervisees may not perceive the development of cultural identity to be as crucial in the SR as REM supervisees. This could be partly explained by the perpetuation of ‘Whiteness’ within the profession, where ‘culture’ is sometimes regarded as a dimension reserved for REM groups only (Wood and Patel, 2017). White supervisees may naturally experience less dissonance and cultural issues than REM supervisees in cross-cultural SRs (Constantine and Sue, 2007). Therefore, a greater level of introspection and awareness may be required to understand the influence of their cultural identity on themselves and others (Prajapati et al., 2019). Within the context of developing cultural responsivity, all practitioners must actively work towards developing their self-awareness regardless of race or ethnicity; it can be harmful if this is not seen as a priority for all (Patel, 2011).

The findings from the current study offer a range of important theoretical and clinical implications. Helms’ (1990, 1995) Racial Identity Development theory proposed that racial identity is developed through processing and working through various stages which exist on a continuum. Cook (1994) expanded on this theory and described progressive, regressive and parallel dyads. In progressive dyads, the supervisor is further along in their cultural identity development than their supervisee. In regressive dyads, the supervisee is further along in their cultural identity development than their supervisor. Furthermore, in parallel dyads, the supervisor and supervisee have similar cultural identity statuses; this could be a high or low development status (Pillay, 2013).

Findings from the present study provide some support for this theory. Due to the predominance of White supervisees within the profession (Turpin and Coleman, 2010), REM supervisees are more likely to be placed in cross-cultural dyads than White supervisees. If White supervisees and White supervisors are similarly at lower stages of cultural identity development, they may experience parallel dyads, where both may be unaware of deficits in cultural responsivity. This might negatively affect cultural competence development and treatment for REM clients (Ladany et al., 1997).

REM supervisees may be further along in their cultural identity development due to their own lived experiences and awareness of cultural differences within the UK; however, this is subject to individual difference (Constantine and Sue, 2007). If paired with a White supervisor at a lower cultural identity development stage, a regressive SR may be experienced. Ladany et al. (1997) suggested that regressive SRs are often the most difficult to navigate due to inherent power dynamics in supervision. White supervisors may not be aware of the need to develop their racial/ethnic identity, particularly if it is often seen as the ‘norm’ within the UK.

Both supervisors and supervisees require mutual trust and safety to develop emotional bonds and create an effective supervisory working alliance (Bordin, 1983). Clohessy’s (2008) theory additionally highlights the influence of contextual factors (i.e. personal experiences and stressors) and relational factors (i.e. trust and safety) on quality of the SR. Although race/ethnicity is not explicitly mentioned within these existing theories, findings in this study imply greater cultural responsivity has beneficial outcomes for the SR, particularly for REM supervisees (with White supervisors). An explanation for this, based on the findings of Bordin (1983) and Clohessy (2008), may be that greater supervisor cultural responsivity fosters a trusting, safe and open environment for supervisees, particularly supervisees from REM backgrounds.

For REM supervisees to feel greater safety and security in the SR, White supervisors may need to demonstrate greater cultural responsivity and model openness, honesty and reflection. They may do so by frequently engaging in cultural discussions and prioritising the development of their own cultural identity, which may require uncomfortable self-assessment and reflection. It becomes vital for power relations and their impact on all individuals within the supervisory triad to be addressed in supervision (Holloway, 1995; Patel, 2011). It is also recommended that supervisors should take responsibility in addressing racial/ethnic power dynamics to promote culturally responsive supervision (Pieterse, 2018). This may aid the development of trust, security and mutual empowerment, particularly in cross-cultural dyads (Bordin, 1983; Clohessy, 2008; Martínez and Holloway, 1997).

The results highlight the discrepancy in perceptions of culturally responsive supervision and the SR between REM and White supervisees. This highlights the need for therapy training programmes, professional bodies, supervisors and service leads to prioritise culturally responsive supervision development. These institutions have the systemic power to promote culturally responsive practices and better support supervisors to do the same (Thrower et al., 2020). There is a need for evidence-based supervision models and protocols that facilitate this development.

This study incorporated the RESS, a validated quantitative measure with good psychometric properties, specifically exploring the constructs of race and ethnicity in supervision (Bartell, 2016). Furthermore, the study incorporated a range of professionals from clinical and counselling psychology and CBT therapists who were either in training or qualified practitioners to ensure the data reflected a range of therapy-related backgrounds. An additional strength was the number of participants (n = 222) that took part, with 39% self-identifying as belonging to a REM group.

In consideration of limitations, this study relied entirely on self-report data which can be subject to bias (Rosenman et al., 2011). Some participants may have provided more socially desirable responses. Additionally, some responses may be based on inaccurate recall or interpretation of retrospective events (Buchanan, 2007). Furthermore, a selection bias may have been present, as participants who volunteered in the study may have had a personal interest in the study topic. Supervisory dyads also remained heavily unbalanced, with very few supervisees in REMSE-REMSR and WSE-REMSR dyads, which reflects the under-representation of REM supervisors within the profession (Turpin and Coleman, 2010).

Supervisees were allocated to supervisory dyads based on the demographic information provided about their supervisor; if unknown, supervisees were asked to take a best guess which may have been inaccurate. The study heavily relied on supervisee perspectives, with no clarification of supervisor perspectives to complement findings. Although there may be differences in perceptions between the two groups, differences in supervision styles and expectations across professions could potentially impact the SR (Fleming, 2004). Further research may benefit from a greater WSE-REMSR and REMSE-REMSR sample size and additional measures (e.g. acculturation measure and a measure of cultural identity development status) to further determine influences of SR quality.

In summary, the present study found differences between supervisory dyads on their perceptions of culturally responsive supervision and SR quality, with REM supervisees (with White supervisors) self-reporting the least culturally responsive supervision and lower quality SRs. The findings suggest that greater cultural responsivity in supervision may strengthen the SR. Overall, this study concludes that self-reflection of cultural identity, power and privilege are important pre-requisites to providing culturally responsive supervision. The profession must prioritise the development of cultural responsivity in supervision and additional training. This could benefit the SR, supervisee well-being and outcomes, ultimately improving the quality of culturally responsive care offered to clients.