Registered social workers’ supervision across areas of practice in Aotearoa New Zealand

INTRODUCTION: This article examines the supervisory experiences and views of registered social workers across the practice areas of statutory services, health and non-government organisations (NGOs.) The article aims to establish a baseline for supervision across areas of practice in Aotearoa New Zealand and discuss the implications any differences have for the supervision of registered social workers. METHODS: Demographic data from 266 postal survey respondents was descriptively analysed. A one-way ANOVA and Tamhane T2 post hoc tests were applied using IBM SPSS 24 to explore variances in means for the independent variable of Area of Practice across 10 scales about the respondents’ supervision experiences. FINDINGS: Differences were identified in the workforce profile of each area, and there were significant differences in supervisees’ experiences of supervision across areas of practice which reflected each area’s different supervision culture, policy, and practices. The findings show that supervision in health and NGO areas was more professional, clinical, cultural, reflective and involved more positive content within a more constructive supervision climate than supervision in the statutory area. CONCLUSIONS: A significant difference was found between the quality of supervision experienced by social workers in health and NGOs and their statutory social work colleagues. This needs to be addressed through changing the supervision climate, developing supervisor capability and the uncluttering of supervision through separating professional/clinical supervision from line management. This study provides a foundation for further research that compares supervision across practice areas.

This article is focused on registered social workers' experiences and views of their supervision across statutory (mostly public child welfare), health and non-governmental organisations (NGOs) in Aotearoa New Zealand and aims to identify any differences in registered social workers' supervision across these areas. The data presented are from a national survey of registered social workers' supervision, which aimed to establish a baseline for their supervision and compare it with the Social Workers Registration Board's (SWRB) policy and guidelines. This is the third article from the survey and presents an additional analysis across areas of practice which ORIGINAL ARTICLE QUANTITATIVE RESEARCH was too substantial to be included in the first article, which explored registered social workers' personal and professional characteristics and identified differences in supervisees' supervision experiences according to ethnicity, experience as a social worker, recognised qualification, sexual orientation, type of registration and gender (see O'Donoghue, 2019a). The second article examined the 138 supervisors' responses to supervisor-specific questions and found significant differences about the area of practice, ethnicity, experience as a social worker and supervisor, social work qualifications and supervisory education and training (O'Donoghue, 2019b). This article examines registered social workers' experiences and views across the three largest areas of practice in Aotearoa New Zealand. It aims to establish a baseline regarding supervision experiences across each area, identify any differences, and discusses the implications of the differences as they apply to the supervision of registered social workers.

Background
Research on social work supervision demonstrates that high-quality supervision improves worker and organisational outcomes and is associated with improvements in client outcomes (Benton et al., 2017;O'Donoghue, 2021;O'Donoghue & Tsui, 2015;Tsui et al., 2017). Few previous studies have specifically explored the differences in supervision across specific practice areas (O'Donoghue, 2021;O'Donoghue & Tsui, 2015;Sewell, 2018). Recent research reviews identified only one study that compared social work supervision across different areas of practice, namely, Scott and Farrow (1993); O'Donoghue (2021); O'Donoghue and Tsui (2015); and Sewell (2018). Scott and Farrow (1993) surveyed statutory child welfare social workers and hospital social workers in Victoria, Australia. From a descriptive analysis of the supervision functions, they identified minor differences between the statutory social workers and hospital social workers, with the latter conforming more with the local professional supervision standards than the former.
In the last decade, there has been specific research about supervision in child welfare, health and NGO settings ( Geißler-Piltz, 2011;McPherson et al., 2016;Rankine et al., 2018;Robinson, 2013;Sewell et al., 2021;Wilkins et al., 2017). British research about child welfare supervision has identified what happens in supervision, how it is recorded, a lack of time and space for reflection and supervision having primarily a managerial focus ( Beddoe et al., 2021;Saltiel, 2017;Wilkins et al., 2017). British researchers have also explored the relationship between supervision, practice, and clients and found that supervision focused on social work practice was associated with improvements in client outcomes Bostock, Patrizo, Godfrey, Munro et al., 2019;Wilkins et al., 2018). In North America, researchers have identified that child welfare supervision relationships varied according to agency culture, that supervisees were satisfied with strength-based supervision, and there was a relationship between work self-efficacy and supervision ( Julien-Chinn & Lietz, 2015;Lietz & Julien-Chinn, 2017;Zinn, 2015). An Australian qualitative study identified effective child welfare supervision needed to have a safe supervisory relationship supported by knowledge and leadership within an empowering organisational and community context (McPherson et al., 2016). A recent local study involving four statutory social work supervisors in a learning community found that providing protected space for evaluating, critiquing and developing supportive supervision practice within a pressured statutory environment can enhance supervisors' critical reflection and practice (Rankine & Thompson, 2021). Overall, this research highlights that, while there have been advances and improvements in child welfare supervision such as strength-based supervision, practice-focused supervision, and the creation of safe spaces for critical reflection, agency culture and a

ORIGINAL ARTICLE
managerial focus dominate and impact the provision of quality supervision.
In health, the supervision of social workers occurs in a multidisciplinary context and amid the politics of the health professions. Geißler-Piltz's (2011) study on the supervision in the health care system in Germany found that the medical domination of health institutions impacted supervision to the extent that social workers had an ambivalent view of supervision in which they valued its contribution to their professional socialisation, yet also felt it did not have a solid connection to their work. A recent study in Ontario, Canada, with health social workers found that most were engaged in administrative and supportive supervision. However, only half participated in clinical supervision, with half of these participants receiving interprofessional supervision from a supervisor from another health profession (Sewell et al., 2021). This study also found that newly qualified social workers had more frequent supervision than their more experienced colleagues and that the frequency of supervision reduced over time. In short, these studies identify a challenging environment for social work supervision in health systems and settings. Turning to supervision in NGO settings, a comparative qualitative study in NGO organisations in Australia and the UK found that among social workers who worked with refugees and migrants that over half of UK participants and a few Australians did not receive formal or informal supervision. In addition, the approach to supervision across NGOs was inconsistent and did not provide adequate support to workers (Robinson, 2013). Similarly, a local, Aotearoa New Zealand study identified that social workers in child and family support NGOs struggled to safeguard sufficient space for reflective supervision in a demanding neoliberal contracting environment (Rankine et al., 2018). In a further study, Rankine (2019) compared internal and external supervision amongst dyads from child and family support NGOs, with five of the eight dyads participating being external supervision relationships. Rankine (2019, p. 44) concluded that external supervision was "a valuable space for participants to openly discuss practice and critically reflect on their work". He also noted that external supervision varied across NGOs in both quality and availability. This was due to its cost and the funding available to agencies. In summary, supervision in the NGO setting also occurs in a challenging environment and is inconsistent in its provision.
Within Aotearoa New Zealand, it has been noted that "supervision differs across fields of practice" and in the "form, functions, models and approaches" (O'Donoghue & Tsui, 2012, pp. 12-13). Statutory child welfare supervision has been traditionally provided by a supervisor with line management responsibility and involved the traditional administrative, educative and supportive functions (Field, 2008). In contrast, supervision in Health involves a dual model of a peer colleague providing the professional supervision and a manager or team leader providing the administrative supervision (Shepherd, 2003). In the NGO field, mixed provision was common with professional supervision provided by an external contractor and managerial supervision provided by a team leader or manager (O'Donoghue, 2010). Despite these known differences, there has been no research in Aotearoa New Zealand that has compared the experiences and views of supervisees across different areas of practice.

Method
The questionnaire was based on a previous instrument used in 2004 and was updated following a review of supervision research (see: O'Donoghue, 2019a;O'Donoghue & Tsui, 2015). It consisted of multi-choice questions about the respondents' background and 5-point semantic differential and Likert type scales. The internal reliability was assessed with Cronbach's Alpha, and 9 of the 10 scales had adequate internal consistency (i.e., >0.5) (see Table 1). Six were

QUANTITATIVE RESEARCH
greater than 0.7, generally accepted as good internal reliability (Helms et al., 2006). The participation in forms of supervision scale, which had the lowest score (0.425), did not have any implications for the use of the data collected because it aligned with Schmitt's (1996, p. 352) criteria of a measure that "has other desirable properties, such as meaningful content coverage", with the content, in this case, being participation in a range of forms of supervision across 12 months. The questionnaire had content, criterion and face validity because it addressed the content and criteria about social work supervision and its constitutive elements as described in the supervision literature (Kadushin & Harkness, 2014;O'Donoghue & Tsui, 2015). It also built on the constructs from a previous instrument used in a 2004 postal survey of supervision, which was conducted before the establishment of social worker registration in Aotearoa New Zealand (De Vaus, 2014;O'Donoghue et al., 2005).
A postal survey was chosen over other methods because the publicly available Social Workers Registration Board Register contained details of each person's workplace but not email addresses. A random sample of 708 social workers was drawn from 4388 registered social workers in 2014. The data collection occurred between December 2014 and February 2015; 20 questionnaires were returned undelivered, and from the 688 questionnaires presumed to have been received, 278 were completed and returned. The overall response rate was 40.4%, lower than the expected response rate of 50%, based on the 2004 survey (O'Donoghue, 2010). The overall sampling error was calculated to be 5.7% at the 95% confidence level, which is within the parameters of 4% and 8% at the 95% confidence level, which is deemed acceptable (Field, 2018). Some 96% (266) of the respondents identified their area of practice as statutory, health, or NGO. The remaining 4% (12) consisted of 10 whose area of practice was education and training and two in private practice. Because of the small number within the education and training and private practice areas, these areas were excluded from this analysis. The questionnaires were analysed using IBM SPSS 24. This analysis involved descriptive statistics and a one-way ANOVA to compare the mean results from the 10 scales with the independent variable of Area of Practice. Where significant differences were identified, Tamhane T2 post hoc tests were applied to identify the differences between the statutory, health and NGO groups. Tamhane T2 tests are a conservative test used when the variances are unequal, and samples differ, which was the case as the area of practice groups are unequal in number, and the standard deviations are varied. Sauder and DeMars (2019, p. 37), recommended the use of the Tamhane T2 test as part of the "better safe than sorry" approach because it is one of four tests that controls for Type 1 error (i.e., the mistaken rejection of a true null hypothesis) in "real-data research (i.e., groups are often unequal, and population variances are almost never equal for demographically based groups)" (Sauder & DeMars, 2019, p. 37). The null hypothesis is that there are no significant statistical mean differences between the area practice groups. The eta squared coefficient (η2) was used to measure the effect size. The effect is deemed small at 0.01, medium at 0.06 and large at 0.14 (Pallant, 2013, p. 264). The alpha level was set at 0.05.

ORIGINAL ARTICLE
Massey University Human Ethics Committee approved the study.
The limitations of the survey are the reliance on the respondents' reports, social desirability bias, and missing data bias. Missing data was addressed by leaving the cells in IBM SPSS 24 blank and reporting the number of respondents throughout the article (Pallant, 2013).

Respondents' characteristics
The overall distribution of the 266 survey respondents across the areas of practice was that 40.9% (n = 109) worked in statutory social work (most working in public child welfare), 36.1% (n = 96) were in health, 23% (n = 61) NGOs. A comparison with SWRB 2014/15 annual report showed 27% of registered social workers worked in public child welfare, 25% in health, and 23% in NGOs (SWRB, 2015a). When the 17% who were not practising were discounted, 32% in public child welfare, 30% in health, and 28% in NGOs. This means that the survey sample was over-representative of those in statutory and health settings and under-representative of the NGO sector.
The respondents' personal characteristics as they relate to each practice area are presented in Table 2. It is difficult to ascertain how  representative their characteristics are of the wider social worker population due to a lack of reliable workforce data at the time of the survey. The comparisons made with the 2013 New Zealand Census Social Work sub-group have limitations. For example, the census asked a binary question about sex identity rather than gender and was not responsive to sexual and gender diversity (Statistics New Zealand, 2013). In addition, there were no questions about sexual orientation in the 2013 and 2018 New Zealand census, and previous estimates of prevalence are unreliable statistically and problematic (Henrickson et al., 2007;Statistics New Zealand, 2021). The respondents' professional characteristics are detailed by area of practice in Table 3. Likewise, it is difficult to ascertain how representative these characteristics are of registered social workers at the survey time. There was limited reliable workforce data available that compared these characteristics amongst each area group.

QUANTITATIVE RESEARCH
Overall, the respondents' personal and professional characteristics provide background about the respondents and each area of practice group.

Findings
The findings reported concern the differences for the forms of supervision participated in, the overall emphasis, logistics, types of contact, the supervision climate, focus, methods and processes, their supervisors' use of ideas from supervision models or approaches, the aspects and contents of sessions, as well as the supervisees'

ORIGINAL ARTICLE
overall satisfaction and evaluation of their supervision.

Forms of supervision
The respondents rated on a scale their level of participation over the previous 12 months (where 1 = "none" and 5 = "high") in each of the forms of supervision. The 12 forms included represented the differing ways supervision is construed and practised in Aotearoa New Zealand (O'Donoghue & Tsui, 2012). Table 4 presents the means and count for each area group, the overall mean and count. The one-way ANOVA and the effect size are shown where mean differences were statistically significant.
The mean differences for individual supervision indicate that supervisees working in health and NGOs participated in more individual supervision than their statutory social work colleagues.
For clinical/professional supervision, the differences show that supervisees in health participated in more clinical/ professional supervision than their NGO and statutory colleagues. NGO supervisees also participated in more clinical/professional supervision than their statutory colleagues. For external supervision, supervisees in NGOs participated in external supervision more than their colleagues in health and statutory, and supervisees in health participated in external supervision more than those in statutory.

Overall emphasis
The respondents rated the overall emphasis of their supervision on a scale (where 1 = "not at all" and 5 = "almost always"). Table 5 shows there were differences in the emphasis on practice with clients, wellbeing and development as a worker, and the environment of the workplace. For all  three items, the results reveal that greater overall emphasis was put on the supervisee's practice with clients, the supervisee's wellbeing and development as a worker, and the workplace environment within supervision in health and NGOs than statutory social work.

Logistics
Several multi-choice questions concerned the logistics involved and included the number of supervisors, the type of supervision agreements or contracts, the frequency of supervision contact and the average length of supervision sessions. Table 6 presents the logistics by area of practice and indicates that over half of supervisees in statutory and health have one supervisor. In contrast, over two-thirds of supervisees in NGOs have two or more supervisors. For supervision agreements, health and NGOs had a higher percentage of supervisees with written supervision agreements than their statutory colleagues. Conversely, the statutory area had a greater percentage who did not have a supervision agreement or had an oral agreement than NGOs and health.
The results for frequency of contact show a greater percentage (43.5%) of supervisees in statutory had at least fortnightly supervision contact compared to those in NGO (37.7%) and health (16.8%). The situation was

Types of supervision contact
The respondents indicated on a scale (where 1 = "not at all" and 5 = "almost always") their experience of a range of types of supervision contact. Table 7 shows differences for formal individual meetings and ad-hoc, informal, opendoor consultations. For formal, individual meetings, these meetings occurred more frequently for NGO supervisees than for their statutory colleagues. Whereas, for ad-hoc, open-door consultations, statutory social workers consulted more frequently with their supervisors on an ad-hoc basis than their colleagues in health.

Climate
The respondents recorded their level of agreement (where 1 = strongly disagree and 5 = strongly agree) for nine statements that concerned their views about their supervision climate pertaining to safety, trust, choice, and relational and power dynamics. Table 8 shows differences for

QUANTITATIVE RESEARCH
all of the statements, with the apparent theme being that NGO and health had higher means and therefore a higher level of agreement than statutory. The higher means across all climate statements for NGO and health indicate a more supportive supervision climate for NGO and health supervisees than for their statutory colleagues.

Focus, methods and processes
The respondents recorded their level of agreement (where 1 = strongly disagree and 5 = strongly agree) for five statements about the focus of supervision. Table 9 shows that health and NGO supervisees had a higher level of agreement about the focus of their supervision on safe and ethical practice, the supervisee's needs, and learning and development than their statutory colleagues. Health supervisees also had a higher level of agreement on the focus given to client issues in their supervision than their colleagues in statutory social work.
The respondents recorded their level of agreement (where 1 = strongly disagree and 5 = strongly agree) for the eight methods and process statements. Table 10 shows that health and NGO social workers had a higher level of agreement about the extent to which their supervision was antioppressive, linked theory and practice, strength-based, reflected on client-worker interactions and used a problem-solving process than their statutory social work colleagues.

ORIGINAL ARTICLE
*Level of agreement ranged from 1 ("strongly disagree") to 5 ("strongly agree") **Mean differences are significant (p<.05) *Level of agreement ranged from 1 ("strongly disagree") to 5 ("strongly agree") **Mean differences are significant (p<.05)  Table 11 details the supervisees' rating of their views about their supervisor's use of aspects or ideas from a range of supervision models/ approaches on a scale (where 1 = "not at all" and 5 = "almost always"). The differences identified across areas for strength-based, reflective, feminist, eclectic, cultural and narrative were that health and NGO supervisees experienced these approaches more than statutory social workers. While the differences for adult learning and solution focused were that the NGO supervisees had a greater experience of these approaches than their statutory colleagues. Table 12 displays the results for the occurrence of specific aspects of the supervision sessions, which the respondents rated on a scale (where 1 = "not at all" and 5 = "almost always"). The specific aspects were based on Morrison's (2005) elements of a session, which O' Donoghue et al. (2005) adapted to the Aotearoa New Zealand context and developed into an 11-item scale to measure their occurrence within supervisees' sessions. There were mean differences across almost all aspects except for the prioritisation of items. Health and NGO had higher means than statutory for preparation, checking in, discussion, summarisation and review, evaluation and closure. NGO also had higher means than statutory for the occurrence of karakia, action planning, agenda setting and decision-making. Overall, the findings suggest that NGO and health supervisees experience a greater occurrence of more aspects of a supervision session than their colleagues in statutory.

Content of Supervision Sessions
The respondents indicated on a scale (where 1 = "not at all" and 5 = "almost always") the occurrence of specific items that were discussed in their supervision sessions. There were significant mean differences for 10 of the items discussed in supervision sessions items (see Table 13).
The differences between groups were that NGO and health had higher means than statutory, for ethical issues, supervisees' concerns or matters, boundaries, professional development, stress, cultural matters, complex and challenging cases and the supervision relationship. NGO had a higher mean than statutory for success stories and statutory had a higher mean than health for supervisors' concerns. These results show that the occurrence of discussions about ethical issues, the supervisees' concerns or issues, boundaries, professional development, stress, cultural matters and complex and challenging cases were more common in supervision for health and NGO supervisees than their statutory colleagues. The discussion of success stories occurred more often for NGO supervisees than statutory. For statutory supervisees, the occurrence of discussions about their supervisor's concerns or matters was more common than for supervisees in health.

Overall satisfaction and overall evaluation
The respondents rated their overall satisfaction with their supervision on a scale (where 1 = "not at all" and 5 = "completely satisfied"). Their overall evaluation of supervision was rated on a scale (where 1 = "poor" and 5 = "excellent"). Table 14 presents the results for overall satisfaction and evaluation. The significant mean differences for overall satisfaction and *Occurrence ranged from 1 ("not at all") to 5 ("almost always") ** Mean differences are significant (p<.05)

ORIGINAL ARTICLE
evaluation show that health and NGO had higher means than statutory. This means that supervisees in health and NGOs are more satisfied with their supervision than their colleagues in statutory social work. They also had a more favourable evaluation of their supervision than their colleagues in statutory social work.

Discussion
Generally, the results have identified that supervisees in health and NGO areas had more satisfying, supportive and practice focused supervision than their statutory colleagues. Health social workers' greater participation in clinical/professional supervision, and NGO higher participation in external supervision, together with a greater overall emphasis on the supervisee's practice with clients, the supervisee's wellbeing and development as a worker and the environment of the workplace within the supervision, than statutory social workers. This difference appears to reflect the administrative and case management supervision emphasis within child welfare supervision (Dill & Bogo, 2009;Wilkins et al., 2017). The results concerning the number of supervisors showed that the majority of supervisees in statutory and health had one supervisor. In contrast, those in NGOs had two or more, which was indicative of their greater participation in external supervision. For supervision agreements, there was a greater percentage of supervisees who did not have a supervision agreement, or had an oral agreement in the statutory area than health and NGOs, which perhaps suggests there was less checking and follow-up concerning supervision agreements and compliance with the supervision policy in statutory social work than the other areas. This arguably would pose a potential challenge to the social workers' ability to comply with any request made by the SWRB for their supervision agreement when renewing their practising certificate and raises questions about organisations' awareness and compliance with SWRB's expectations concerning supervision (SWRB, 2015b).
The differences for the types of supervision contact showed firstly that NGOs had more individual formal meetings than their colleagues in statutory social work and secondly that statutory supervisees consulted their supervisor more on an adhoc basis. These results seem to indicate that supervision contact for statutory social workers was less planned or more ad-hoc and appears to indicate a supervision culture driven by crises (Hawkins & McMahon, 2020). These findings also reflect the British, North American, and local literature, which emphasises the lack of time and space for reflection and supervision, the influence of managerialism and the variability of agency culture (Beddoe et al., 2021;Rankine & Thompson, 2021;Zinn, 2015).
The mean differences across all climate statements further emphasises this with health and NGO having a more supportive supervision climate than statutory. The less supportive climate for supervision in statutory social work was characterised by less choice of supervisor, lower trust of the supervisor, who was perceived to have less expertise in supervision and practice. These environments lead to a less open, honest and constructive relationship in which the power dynamics were not as well managed, and it was less safe for supervisees to discuss both ethical issues and their emotions. Arguably, this climate parallels that, within this field of practice, which was highly influenced by politically driven child protection reforms which promulgated a risk-focused, investigative, child rescue practice climate in the face of variable public confidence and high scrutiny (Hyslop, 2021). Such an environment is not ideal for quality supervision and contributes to low trust and supervisees' experiencing challenges in the use of power, authority and relational dynamics within supervision (Young, 1994).
The findings detailing what is focused on in supervision reflect health's professional or clinical supervision focus through having a greater focus on safe and ethical practice, the supervisee's needs, learning

QUANTITATIVE RESEARCH
and development, and client issues than statutory. Similarly, NGOs' greater participation in external supervision is apparent in their greater focus on safe and ethical practice, the supervisee's needs, learning and development than statutory. The methods and process statements findings showed that supervision for health and NGO supervisees was more antioppressive, linked theory and practice, more strength-based, more reflective on clientworker interactions, and had greater use of a problem-solving process than statutory.
These results further reinforced that supervision for health and NGOs was more professional or clinical and more reflective than that of statutory. The greater reported use of supervision approaches and models by their supervisors amongst health and NGOs supervisees added to the previous findings of linking theory and practice in supervision and further illustrated the extent to which supervision was more professional or clinical and reflective than statutory. The results concerning the aspects of supervision session where NGO and health supervisees had more experience of most aspects of a supervision session than their statutory colleagues suggests that statutory supervisees experience less structure within their supervision sessions and arguably supports the findings related to less experience of individual sessions and greater experience of ad-hoc contact or meetings.
The differences concerning the occurrence of what was discussed in supervision reinforced that for health and NGO supervisees, there was a greater occurrence of professional, clinical, cultural, reflective, supervisee focused, and positive content discussed than their statutory colleagues. Contrastingly, the only content that was more commonly discussed in statutory supervisees' supervision was their supervisor's concerns or matters, which arguably suggests a more managerial supervisor-led content.
Given all the differences discussed above, it was not surprising that health and NGO supervisees were more satisfied and evaluated their supervision more favourably than their statutory colleagues. Overall, the findings indicate that supervision of social workers in health in Aotearoa New Zealand, is in reasonably good shape and is professional, clinical, reflective, and conducted in a supportive supervision climate. The results also show a wider difference between supervision of health social workers and statutory social workers than the minor differences found by Scott and Farrow (1993). The results of this survey are markedly different from the ambivalence reported in the German study of supervision within health settings and the lack of participation in clinical supervision reported in the Ontario study (Geißler-Piltz, 2011;Sewell et al., 2021). The NGO supervision findings show a more positive experience of supervision than those reported in the comparative UK and Australian study (Robinson, 2013). The results are also more positive about the experience of reflective supervision than Rankine et al. (2018). The reason for this is likely due to the widespread participation in external supervision, which ensures that NGO supervisees were supported and had space to reflect on their well-being and professional development (Rankine, 2019).

Implications
The findings show that the supervision experiences of NGO and health supervisees were more satisfying and better than their statutory colleagues. This was because they experienced more professional, clinical, reflective and supervisee focused supervision. They also had greater participation in both clinical/professional supervision and external supervision and a more supportive supervision climate than their statutory colleagues. The implications of these findings concern how the experiences of statutory social work supervisees might be improved through learning from NGOs and health. Given the passage of time since 2015, when the survey was conducted, questions arise  Tribunal, 2021). These reports detail that, despite Oranga Tamariki updating its supervision policy and providing training for supervisors in 2018, supervision is provided inconsistently and is predominantly task focused rather than engaging in critical reflective practice. In other words, they resonate with recent research into child welfare services in the UK (Beddoe et al., 2021;Saltiel, 2017;Wilkins et al., 2017). Recent local research involving statutory social work supervisors in a learning community identified that supervisors inhabit a "cluttered supervision space" influenced both managerial demands and expectations and the need to ensure safe practice and support practitioners (Rankine & Thompson, 2021, p. 98). Rankine and Thompson (2021) asserted that the learning community approach may be a pathway for statutory supervisors to create an environment conducive to more critically reflective conversations in supervision and provide a starting point for the development of supervisor capability and a culture change in statutory social work. Their research highlights the importance of prioritising space for reflection to support a professional approach to supervision rather than a managerial led one. In contrast, the experiences of health and NGOs supervisees from this survey show a more professional, clinical, reflective and supervisee focused supervision occurring within a supportive supervision climate. Notably, supervisees in health and NGOs have a greater ability to choose their supervisor because their supervision is either clinical/professional with a peer or with an external supervisor. If statutory social work supervision is to improve its supervision climate, providing supervisees with a greater ability to choose their supervisors would build trust and strength in the supervision alliance, which is foundational to effective supervision (Davys, 2002;O'Donoghue et al., 2018). It would also necessitate separating linemanagement from professional practice supervision and require an investment in building supervisor capability so that the professional supervision model and training aligns with the recommendations of He Pāharakeke, he Rito Whakakīkīnga Whāruarua: Oranga Tamariki Urgent Inquiry concerning better and more consistent supervision across all sites (Waitangi Tribunal, 2021). Recent UK research suggests that professional supervision focused on improving the practice of social workers is more likely to be associated with improvements in the outcomes for children and families Bostock, Patrizo, Godfrey, Munro et al., 2019, Wilkins et al., 2018. For such a significant change to occur for statutory social workers, the support of the professional and regulatory bodies in Aotearoa New Zealand is needed. They need to unequivocally state in their respective supervision policy documents that supervision of social work practice is a specialist role and is separate and different from line management.

Conclusion
This article aimed to establish a baseline regarding supervision experiences across each area of practice, identify any differences across areas, and discuss the implications of the differences as they apply to the supervision of registered social workers. The strengths of the findings are that a baseline for each of the three areas has been established, which is something that future researchers can build upon. Differences in supervision across the three areas have been identified, showing that supervision in health and NGOs was more professional, clinical, cultural, reflective and involved more positive content within a more constructive supervision climate than statutory social work. For health and NGO social workers, the findings differ from the international literature and present a more positive portrait of supervision. The statutory social work findings align with recent reports from the Ombudsman, Waitangi Tribunal, and child welfare supervision research in which supervision is described as inconsistently provided and is predominantly task focused rather than engaging in critical reflective practice (Beddoe et al., 2021;Saltiel, 2017;Wilkins et al., 2017). The limitations of the findings are that the data is from 2015, and it is difficult to ascertain how representative it is of registered social workers. Nonetheless, it is recommended that line-management and professional practice supervision be separated in the supervision of statutory social workers and that they have a greater ability to choose their supervisors. In addition, it is also recommended that the professional and regulatory bodies in Aotearoa New Zealand revise and emphasise in their supervision policy documents that the supervision of social work practice is a specialist role that is separate and different from line management.