Online Group Supervision in Graduate Psychology Training During the COVID-19 Pandemic

Group supervision, a common method in graduate psychology training, shifted abruptly to online learning due to the COVID-19 pandemic. This study aimed to increase understanding of psychology graduate students’ perception of online group supervision during COVID-19 by focusing on the group process and the students’ professional identity formation. Data were collected through an online survey comparing an online COVID-19 group sample with a pre-COVID-19 in-person sample. Our findings showed no difference between online group supervision during COVID-19 and in-person group supervision prior to COVID-19 in students’ reports of group processes and the prevalence of professional identity statuses. However, group processes differed according to students’ professional identity statuses while accounting for the supervision format. We discuss the results of our study and offer several theoretical and practical implications regarding online supervision.

COVID-19 brought profound changes to higher education which shifted abruptly to online learning (Crawford et al., 2020).Like other disciplines, psychology programs moved to online group supervision for their students' practical training (e.g., Nadan et al., 2020).However, knowledge about how the online supervision format works is lacking, and there is a need for more research (Parks, 2020).The current study is guided by theoretical and empirical literature that has proposed that group processes contribute to both group supervisees' learning experience (Alschuler et al., 2015) and professional identity (Ayo et al., 2010).It aims to understand psychology graduate students' perception of online group supervision during COVID-19 by comparing their perceived group processes and professional identity formation to those of students who underwent in-person group supervision training before COVID-19.This understanding may establish supervisors' knowledge of the associations between online participation and group dynamics which, in turn, may enhance their effective adoption of online modality.

Group Supervision: In-Person and Online
Group supervision is an integral part of novice psychologists' training and an essential component of learning and professional identity development (Alschuler et al., 2015;Hanetz Gamliel et al., 2020).Beyond its pragmatic advantages (Fleming et al., 2010), group supervision has promising benefits for participants' professional development (Ögren & Jonsson, 2003).Group supervision broadens the frame of clinical reference by observational learning and exposes trainees to a variety of psychopathologies and treatment approaches (Alschuler et al., 2015).It thus increases their knowledge, skills (Linton & Hedstrom, 2006), and self-efficacy (Christensen & Kline, 2001), and contributes to the development of their identity as therapists (Ayo et al., 2010;Hanetz Gamliel et al., 2020).
Online group supervision is defined as a group using digital platforms through a synchronous audio-video format in which supervisor and trainees are not located in the same physical location (Pennington et al., 2019).The flexibility, convenience, cost-effectiveness, and accessibility of online group supervision are considered major advantages, especially for isolated students and clinicians seeking advanced training (Elliott et al., 2016).Furthermore, the empirical research conducted on online group supervision, albeit limited, has demonstrated that it is a feasible and effective setting as in-person group supervision (Abbass et al., 2011;Traube et al., 2021).For example, satisfaction and relationship with the supervisor and reported selfefficacy of counseling psychology students were found to be similar in both remote and inperson formats (Reese et al., 2009).
However, online group supervision presents unique challenges that might hinder the potential contribution of the supervision.Specifically, all digital platforms are susceptible to technical difficulties and are characterized by two-dimensional interaction.Therefore, interpreting data is limited and prone to miscommunication and misunderstanding (Rousmaniere et al., 2014).Moreover, as online supervision excludes informal encounters of supervisors and supervisees before and after the meeting, interpersonal connections may be further reduced (Weinberg, 2021).
The COVID-19 pandemic added a unique aspect to the experience of online group supervision.The rapid spread of the virus and subsequent quarantine policy had adverse economic, social (Shigemura et al., 2020), and psychological consequences worldwide (e.g., Torales et al., 2020), including in Israel (Israeli Central Bureau of Statistics, 2020).College students have, specifically, experienced additional distress due to the uncertainty and abrupt disruption of the semester and schools' closures (Zhai & Du 2020).This, in turn, has affected their relationships with the groups, e.g., familial, educational, and professional, which students rely on as important sources of self-esteem, meaning in life, and life satisfaction (Pyszczynski et al., 2021).Yet, the powerful human need to belong (Baumeister & Leary, 1995), especially in times of distress, may have increased the need to seek comfort, support, and enhanced selfesteem (Barnes, 2021) through social connections to these groups (Marmarosh et al., 2020).The scarce research data on online group supervision during COVID-19, while mainly qualitative and illustrative in nature, have supported this argument.For example, reports on online group supervision for counselors in China (Chen et al., 2021) and social workers in Italy (Cabiati, 2021) demonstrated that participation in such groups helped supervisees to process their feelings of frustration and helplessness, enhanced their capacities to cope more effectively with stressful life events, and fostered a strong sense of community among them.A report from the United States indicated that the transition to remote group supervision training created both anxiety and greater self-efficacy among trainees (Scharff et al., 2021).Finally, a reflective article from Israel on online group supervision for family therapy trainees following the COVID-19 outbreak found that the trainees reported increased responsibility and involvement in the group which enriched group discussions and enhanced their learning experience (Nadan et al., 2020).
In sum, as online group supervision and quarantines during COVID-19 may have framed students' experience of the group, the present study aimed to also assess students' COVID-19 related worries and perceived social support, which might have related to their perception of group processes in this supervision format.

Group Processes
The long-established literature relating to group processes during group supervision indicates three central group processes-group cohesion (Fleming et al., 2010), group climate (Hawkins & Shohet, 2012), and supervisor/supervisee working alliance (Livni et al., 2012)-as contributing to the group supervisees' learning experience and professional identity formation (Hanetz Gamliel et al., 2020;Mancini et al., 2015).
Group cohesion is defined as a sense of belonging and a belief that the group is important to the individual members' outcomes (Burlingame et al., 2011).Cohesion in the group supervision of novice psychologists was found correlated with their learning about patients, their identity as therapists (Fleming et al., 2010), and their experience of the supervision as significant and effective (Livni et al., 2012).Group climate, indicating the atmosphere in a group, is a multidimensional construct comprising members' perceptions of their own engagement with the group, avoidance of important or difficult topics, and conflict among group members (Gullo et al., 2015).Positive group climate in group supervision was found to promote learning (Fleming et al., 2010;Hawkins & Shohet, 2012).Working alliance, though dyadic in its original definition, has been applied to the group format and involves the presence of an emotional bond between the group leader and group members as well as agreement among them regarding the group's goals and tasks needed to achieve them (Bakali et al., 2013;Bordin, 1983).Positive supervisory working alliances were found to be significantly linked to professional development and job satisfaction for helping professionals (Livni et al., 2012), and positively associated with group members' self-disclosure and the overall group experience (Robak et al., 2013).
The limited research on group processes within online groups in general and online group supervision in particular has suggested that group cohesion and group climate can develop in online groups but at a slower pace than in in-person groups (Weinberg, 2021).Furthermore, a working alliance with the supervisor was found to be a key factor in the effectiveness of online group supervision (Rousmaniere et al., 2014) and sometimes even stronger than that experienced in in-person supervision (Elliott et al., 2016).A recent qualitative study on online group supervision among psychological counselors found that feedback from the supervisor and peers in online group supervision contributed to professional development (Amanvermez et al., 2020).

Professional Identity
Professional identity, which is a dominant aspect of adults' self-identity, consists of personal motives, interests, experiences, and competencies that are associated with a person's professional role.Professional identity also implies adopting the associated norms and values of one's profession (Pratt et al., 2006).Hence, professional identity is constructed via both intraindividual process and intergroup processes (Tajfel, 1982), according to which individuals define their own identities within the context of their membership in social groups.It has been argued that the processes emerging in group supervision, such as group cohesion, group climate and, especially, the working alliance with the supervisor, define members' experience of the group.This experience, in turn, contributes to the formation of their professional identity (Hanetz Gamliel et al., 2020).
Various studies have shown that professional identity among undergraduate psychology students is linked to job-related and academic factors (e.g., Mancini et al., 2015).Other studies have suggested that when professional identity status is described as committed to the profession, individuals' well-being, emotional adjustment (Crocetti et al., 2011), and job-related outcomes (Crocetti et al., 2014) are more positive in comparison to less committed statuses.Contemporary research has shown that the COVID-19 pandemic was related to professional identity formation in the helping professions, such that medical students reported that, despite the challenges, their professional identity formation remained unchanged (e.g., Findyartini et al., 2020).Similarly, among nursing students, COVID-19 was identified as contributing to a higher level of commitment to the profession (Shengxiao et al., 2021;Zhang et al., 2021).
To conclude, despite the growing body of literature on online group supervision (e.g., Miller, 2020), there is a lack of empirical quantitative research addressing online group supervision in higher education programs.Given the ever-growing use of online groups, the likely continuation of online elements in higher education (Crawford et al., 2020), and our recent experience during the COVID-19 pandemic, the current study seeks to begin filling this gap.It aimed to deepen understanding of the online group supervision during COVID-19 by focusing on students' perceptions of group processes and their professional identity statuses.Since the nature of our study design does not allow determining causality, we focus on group differences and associations.Specifically, the study's questions were: (1) Does the professional identity statuses' prevalence differ between online and inperson group supervision?
(2) Do the group processes (group cohesion, group climate, working alliance with supervisor) differ between online and in-person group supervision when considering professional identity statuses?
(3) Are COVID-19 related worries and social support associated with online group processes?

Method Participants
A total of 250 psychology graduate students from universities and colleges around Israel were recruited after finishing both their practical training (practicum) and their group supervision that was held in the academic institutions.The study's cohort was divided between: (a) the inperson pre-COVID-19 sample, which comprised 129 students (106 females), mean age 29.73 years (SD = 3.64), practicum duration 11.57months (SD = 3.69), and mean group supervision size 6.46 members (SD = 1.52) and (b) the online (via Zoom) COVID-19 group sample, which comprised 121 students (96 females), mean age 29.66 years (SD = 3.80), practicum duration 12.42 months (SD = 4.12), and mean group supervision size 6.20 members (SD = 1.31).There were no significant differences between the two samples in gender, age, practicum duration, and supervision group size.
The primary task of the group supervisions, above and beyond specific theoretical orientations, is to broaden group members' clinical orientation and to equip them with skills for the practice of psychotherapy.In both samples group supervision meetings were held weekly for the entire academic year with supervisors who are experts in their field.

Group Climate
The Group Climate Questionnaire-Short Form (GCQ; MacKenzie, 1983) is a 12-item self-report questionnaire assessing individual perceptions of the group environment.Each item ranges from 0 (not at all) to 6 (extremely).The GCQ comprises three-factor analytically derived subscales: engagement (Cronbach's alphas for the current samples was .76);avoidance (Cronbach's alphas = .69);and conflict (Cronbach's alphas = .67,after removing item 5).Higher scores indicate higher levels of engagement, avoidance, and conflict.

Group Cohesion
The 9-item cohesion subscale of the Therapeutic Factors Inventory (TFI; Lese & MacNair-Semands, 2000) ranges from 1 (strongly disagree) to 7 (strongly agree).The total score is the sum of the nine responses with higher scores indicating higher levels of cohesion.Cronbach's alpha for the current sample was .90.

Working Alliance
The Working Alliance Inventory/Supervision-Short (WAI/S-S; Ladany et al., 2013) is a 12-item self-report questionnaire used to assess trainees' perceptions of the working alliance with their supervisor.Each item ranges from 1 (not at all true) to 7 (very true), comprising one general scale and three subscales: goal, task, and bond.Higher scores indicate higher levels of goal, task, bond, and a general score.Cronbach's alpha for the current sample was .93 for the general score, .72 for the goals subscale, .90 for the task subscale, and .87 for the bond subscale.

Professional Identity
The Professional Identity Status Questionnaire (PISQ-5d; Mancini et al., 2015) is a 20item self-report questionnaire ranging from 1 (not at all) to 5 (very much) that was proved to be a useful tool for the evaluation of professional identity formation among university students (Mancini et al., 2015) The PSIQ-5d comprises five subscales: identification with commitment (α = .81);affirmation (α = .78);practice (α = .57);in-depth exploration (α = .42);and reconsideration of commitment (α = .71).These reliability levels conform with those of Mancini et al. (2015).Due to the relatively low internal consistency for practice and in-depth exploration, we eliminated these two scales from further analyses.Mancini et al. (2015) suggested a factor analysis of these subscales that formed five identity status clusters: (1) achievement-individuals who have made a professional commitment following a period of exploration; (2) foreclosureindividuals who have made a strong professional commitment without having explored alternatives; (3) moratorium-individuals who have yet to make a professional commitment but are still actively exploring alternatives; (4) diffusion-individuals who have yet to make a professional commitment and who have not engaged in exploration; and (5) searching moratorium-individuals who are vacillating between the moratorium and achievement statuses and who seek to revise commitments that have already been acted on.

COVID-19 Related Worries and Social Support
Students' worries related to the COVID-19 outbreak were measured using a 3-item questionnaire designed specifically for the current study.The questions addressed the students' reports of the impact of the pandemic on their economic state, social interactions, and academic studies ranging from 1 (no implications) to 5 (destructive implications).A higher score indicated greater experience of worries.Social support during the pandemic was measured using a 3-item questionnaire addressing the degree of support received from friends, family, and online friends with answers ranging from 1 (no support) to 5 (high support).A higher score indicated greater experience of worries.Demographic information included age, gender, practicum setting, and the size and duration of group supervision.

Procedure
The in-person group supervision sample was recruited between 2017 and 2018 and was approved by the institution's ethics committee (# 2017080).The online group supervision sample was recruited between July and December 2021 (# 2021058).Participants for both groups were mostly recruited using a snowball sample through social networks (Facebook and student WhatsApp groups).Participants were provided with a link to a Qualtrics survey (www.qualtrics.com)which they completed online.Some participants of the in-person sample were approached personally by a research assistant and thus completed the questionnaires manually and returned them in a closed envelope.An informed consent form was completed by all participants prior to completing the questionnaires.

Data Analysis
Descriptive statistics are presented as M (SD) or counts, as appropriate.Pearson correlations or Chi-square tests were used to test for correlations between quantitative and categorical measures, respectively.Research questions regarding group comparisons were performed using MANOVA models, followed by univariate analyses and post-hoc analyses using the Bonferroni correction for multiple comparisons.The additional research question regarding the associations with COVID-19 and social support was tested using Pearson correlations.Power analysis for the main hypotheses, conducted using G*Power V3.1.9.4,ascertained that the sample size provided 95% power for detecting a medium effect size for 5% significance level.

Results
To test the study's first question concerning a comparison of the prevalence of professional identity statuses in in-person and online groups, we performed a K-means cluster analysis for the online group based on the PISQ-5d identity construction subscales.In line with Hanetz Gamliel et al.'s (2020) study, we adopted the 3-cluster solution as suitable for our data.The prevalence of professional identity statuses did not differ across supervision modalities (inperson and online) (χ 2 (249, 2) = 2.24, p = .33).Most students in both the in-person groups (52 students, 40.3%) and online groups (59 students, 49.2%) were characterized by the diffusion identity status, i.e., they had relatively high scores on the identification with commitment, affirmation, and reconsideration of commitment subscales.Next, 50 students (38.8%) from the in-person supervision groups and 37 students (30.8%) from the online supervision groups were characterized by the achievement identity status, that is, by high scores on identification with commitment and affirmation subscales and a low score on reconsideration of commitment.Finally, 27 students, (20.9%) from the in-person groups and 24 students, (20.0%) from the online groups were characterized by the moratorium status and thus scored low on identification with commitment and on affirmation and high on the reconsideration of commitment subscales.
To test the study's second question concerning a comparison between group processes (group cohesion, group climate, and working alliance with supervisor) in the in-person and online supervision groups, while accounting for the professional identity statuses, we used a twoway MANOVA model to examine the joined effect.Table 1 shows the means and SDs of group process across the various identity statuses for the two modes of supervision and the comparison by mode of supervision and identity statuses.The MANOVA resulted in a significant global effect (F (14, 204) =2.89, p =.007(for professional identity status, such that students with achievement status reported higher group cohesion, engagement, and working alliance (task, goal, and bond) than students with moratorium status.In addition, students with diffusion status reported significantly higher working alliance (task, goal, and bond) than students with moratorium status.There were no differences between online and in-person group supervision in group cohesion, group climate, and working alliance with supervisor.Likewise, no interaction effects were found between the mode of supervision and professional identity status, namely, the differences between identity statuses held for both modes of supervision.
To test the study's final question concerning the associations between students' perceptions of COVID-19 related worries and social support and group processes, Pearson coefficients were examined (see Table 2).Significant positive correlations were found between students' worries about the impact of COVID-19 on their social interactions and their reports of group engagement and group conflict.In other words, higher concerns regarding social interactions during COVID-19 were correlated with higher commitment to the group and higher conflict in the group.Group conflict was also positively correlated with worries about economic state and academic studies.Concerning social support during COVID-19, support from friends and family was positively correlated with group cohesion and working alliance with the supervisor.Social support from family was also positively correlated with group climate.To conclude, the more the students felt socially "held," especially by their family, the more they reported involvement and connection to the group.

Discussion
This study explored online group supervision by focusing on group processes and professional identity formation among graduate psychology students during the COVID-19 pandemic.It sought, in addition, to test the association between COVID-19 related worries and social support and various aspects of online group processes.
Our findings show no difference between online group supervision during COVID-19 and in-person group supervision prior to COVID-19 in graduate students' reports of group processes (group cohesion, group climate, and working alliance with the supervisor) and the prevalence of the three professional identity statuses: achievement, diffusion, and moratorium.Group processes only differed according to students' professional identity statuses regardless of the supervision format.Additionally, we found that students' worries about social interactions and their perceived social support were linked to involvement with the group and the supervisor.
First and foremost, the findings regarding supervision formats coincide with previous research demonstrating that, despite many concerns and challenges, online learning in general, and online group supervision, are comparable to in-person learning (e.g., Lowenthal et al., 2020;Pei & Wu, 2019) and supervision (Abbass et al., 2011;Elliott et al., 2016;Traube et al., 2021).Furthermore, the present findings stress the important intercorrelation between group processes and professional identity in both modes of supervision.Specifically, we found that the working alliance with the supervisor (i.e., task, goal, and bond) and with the group members (i.e., engagement and cohesion) (Elliott et al., 2016) had a key role in the distinction between committed and moratorium students, as was previously found by Hanetz Gamliel et al. (2020).
Drawing from the entitativity perspective, according to which some level of interactivity and similarity is needed in a group (Blanchard et al., 2021;Campbell, 1958), it may be assumed that, as in the in-person supervision format, supervisees' perceptions of the similarity of characteristics and goals within their online groups enabled them to belong and to experience group outcomes (Blanchard et al., 2021).It may therefore be suggested that while the format distinguished online group supervision from in-person group supervision, the function and goals of both are similar.
From a complementary perspective that focuses on the effects and consequences of COVID-19, it might be argued that the online supervision format demonstrates the importance of groups in times of crises, such as a worldwide pandemic and possibly other global crises (Marmarosh et al., 2020).Specifically, being in this social milieu created an opportunity for the participants to interact with their peers and supervisors and receive support, up-to-date information, and a more realistic approach toward their profession as psychologists (Brusadelli et al., 2020;Marmarosh et al., 2020).In turn, this interaction enabled them to experience therapeutic factors such as universality, support, and cohesion that have been empirically linked to better outcomes (Yalom & Leszez, 2005).
In an attempt to shed light on how group participation may be linked to coping with pandemic distress, we examined the associations between COVID-19 worries and support and group processes.We found that students' reports of more pandemic-related worries and less social support were linked to their reports of the group process as less constructive.Students who felt supported by family and friends were more involved in the group and experienced the group as more cohesive and the relationship with the supervisor as more constructive and positive.While supporting the assertion that groups often involve the re-enactment of the family cell (Yalom & Leszez, 2005), these findings possibly suggest that supervisees in the present study used social connections to friends, family, and colleagues to create an anxiety buffer and to enhance their self-esteem in times of distress (Barnes, 2021).Moreover, students who reported more pandemic-related worries were experiencing simultaneously more conflict yet more involvement with the group.This may be viewed in line with the group work perspective, according to which conflict within a group is valued as necessary for activating and enabling a working atmosphere in which negative emotions may be explored (Bakali et al., 2013).
Additionally, the fact that no associations were found between COVID-19's adverse consequences and the formation of professional identity implies that being a novice therapist during the hardships of COVID-19 may have imbued students with meaning and enabled them to construe themselves as valuable contributors to a meaningful universe (Pyszczynski et al., 2020).It may also reinforce their feeling of belonging to a valuable group, which is important for their future professional identity (Burlingame et al., 2011).This explanation is supported by recent research findings, which found that working individuals demonstrated less psychological distress during the COVID-19 pandemic than non-working peers (Shakil et al., 2021).
Interestingly, and as was found previously (Hanetz Gamliel et al., 2020), most students in both samples were characterized by diffusion status, which means that they neither totally committed to the profession nor engaged in further professional exploration.This finding may reflect the participants' "advanced student phase" attitude on finishing their practicum (Rønnestad & Skovholt, 2003).Accordingly, while appreciating their professional training, they realize that there is still much to learn and are likely to feel insecure and in need of actively seeking confirmation and feedback from seniors and peers (Rønnestad & Skovholt, 2003).In this sense, adhering to the diffusion status may be viewed as reflecting the fact that becoming a therapist is a long journey, characterized by ambiguity, unclarity, and struggle.

Limitations and Future Research
Our findings join an emerging and greatly needed body of research documenting the nature and practice of online group supervision in online higher education settings.However, sseveral caveats should be mentioned.First, the main concern regarding the interpretation of our findings lies in the artifact that identifies online supervision with the outbreak of COVID-19 in its earlier stages and thus prevents us from distinguishing between the effect of COVID-19 and the effect of online supervision.However, since no differences were found (no effect), we may cautiously assume that neither COVID-19 nor the mode of supervision are related to group processes.It may be interesting to simultaneously test these different modes of supervision and thus diminish the COVID-19 effect.Second, due to this study's cross-sectional design, we cannot determine directional influences or causality.We therefore suggest that future research focuses on interviewing the same students at repeated time points, which will enable the construction of a cross-lagged model of causality.Third, the relatively modest sample size prevented us from detecting additional significant effects and testing additional and more complex hypotheses.All of the measures were self-report questionnaires which may have created an informer bias and shared-method variance.Future studies might benefit from using multiple informants, for example, supervisors' perspectives on the role of group processes in online supervision.Finally, the sample recruitment method did not allow for multi-level modeling at both group and individual levels.

Conclusion and Recommendations
The current study's main finding that online supervision groups closely resemble and are as effective as in-person supervision groups provides further support for the benefits of the online format of programs with practical training.This initial finding implies that training courses can be taught using online platforms.Moreover, online group supervision can be integrated into psychology graduate programs as part of the curriculum, especially in times of crises when there are challenges of isolation and distance.However, moving from the "circle" of the in-person group to the squares of the screen requires specific knowledge, and higher education institutions should prepare their teachers to teach and supervise online (Andersen & West, 2021).For example, supervisors may be encouraged to assume a role as facilitators of the group as a whole while helping members tolerate conflict, embracing vulnerability, and discussing clear goals and tasks for the group.Such training might entail increasing supervisors' self-confidence in conducting online groups and practicing how to establish the group process and overcome the lack of in-person interaction by involving all participants online.This involvement can include encouraging all group members to express their ideas at each meeting.This may allow students to feel more confident and enable an atmosphere of safety and group belonging which will advance their effective learning of basic clinical skills and the adoption of a more committed professional identity.Finally, even though the context of the COVID-19 pandemic somewhat limits the results of the study, there can be no doubt that the pandemic has changed education and psychotherapy indefinitely.In face of ongoing worldwide uncertainties, experiencing useful online group supervision can serve as a model and increase supervisees' self-confidence in providing online therapy/consultation whenever needed.

Declarations
The authors declared no funding was provided for this research.The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Table 1
Group Process Means and SD by Supervision Mode (In-Person VS Online) and Professional Identity Statuses

Table 2
Correlations Between COVID-19 Worries and Social Support and Group Process in OnlineGroup Supervision(N= 250)