Moderation Effects of Supervisee Levels on the Relationship Between Supervisory Styles and the Supervisory Working Alliance

Aug 10, 2022 | Author Videos, Volume 12 - Issue 2

Dan Li

Supervisee development is integral to counselor training. Despite the general acknowledgement that supervisors adopt different styles when supervising counselor trainees at varying levels, there is a paucity of studies that (a) measure supervisee levels using reliable and valid psychometric instruments, other than a broad categorization of supervisees based on their training progression (e.g., master’s level vs. doctoral level, practicum vs. internship, counselor trainee vs. postgraduate); and (b) empirically document how the matching of supervisory styles and supervisee levels relates to supervision processes and/or outcomes. The supervisory working alliance is key to the supervision process and outcome. To test the hypothesized moderation effects of supervisee levels on the relationship between supervisory styles and the supervisory working alliance, the author performed a series (n = 16) of moderation analyses with a sample (N = 113) of master’s- and doctoral-level counseling trainees and practitioners. Results suggested that supervisee levels and their three indicators (self and other awareness, motivation, and autonomy) were statistically significant moderators under different contexts. These findings (a) revealed extra intricacies of the relationships among the study variables, (b) shed light on future research directions concerning supervisee development, and (c) encouraged supervisors to adopt a composite of styles to varying degrees to better foster supervisee growth.

Keywords: supervisee development, supervisory styles, supervisory working alliance, supervisee levels, moderation analyses

Clinical supervision is integral to promoting counseling supervisees’ learning (Goodyear, 2014), safeguarding the quality of professional services offered to supervisees’ clients, and gatekeeping the counseling profession (Bernard & Goodyear, 2019). Because supervisors and supervisees are two parties of the tripartite entity of supervision, literature has extensively documented supervisor characteristics (e.g., supervisory styles, self-disclosure, cultural humility), supervisee characteristics (e.g., professional development levels), and the relationship between the two (e.g., supervisory working alliance) as related to supervision processes and outcomes (King et al., 2020; Ladany, Walker, & Melincoff, 2001; Stoltenberg & McNeill, 2010).

Of these relationships, research has consistently revealed a positive correlation between supervisory styles and the supervisory working alliance (Efstation et al., 1990; Heppner & Handley, 1981; Ladany & Lehrman-Waterman, 1999; Ladany, Walker, & Melincoff, 2001). Although such direct positive correlation is theoretically appealing and statistically compelling, there is limited research that further investigates the intricacy of this association, if at all (e.g., whether the direction or strength of this relationship may alter in different contexts). Particularly, abundant supervision literature (Friedlander & Ward, 1984; Li et al., 2018; Li et al., 2019; Li, Duys, & Granello, 2020; Li, Duys, & Vispoel, 2020; Stoltenberg & McNeill, 2010) suggested the adoption of different supervision approaches when working with supervisees at various levels of professional development. Therefore, supervisee levels present as a potential context to examine how supervisory styles relate to the supervisory working alliance.

However, supervisee levels are frequently conceptualized based on supervisees’ training progression (e.g., master’s level vs. doctoral level, practicum vs. internship, counselor trainee vs. postgraduate), which may not accurately approximate where supervisees are. As such, I adopted the Supervisee Levels Questionnaire-Revised (SLQ-R; McNeill et al., 1992), a reliable and valid psychometric instrument, to measure supervisee levels (collectively as an overall assessment and separately with their three indicators) in this study.

Supervisory Styles
     Supervisory styles embody a constellation of behavior patterns that supervisors exhibit in establishing a working relationship with supervisees (Hunt, 1971) and are related to the interactional pattern that is fostered by supervisors in a direct or indirect manner (Munson, 1993). Specifically, supervisory styles encompass supervisors’ consistent focus in supervision, the manner in articulating their theoretical orientation, as well as the philosophy of practice and supervision and how it is communicated to supervisees (Munson, 1993). Friedlander and Ward (1984) identified three distinctive factors that correspond to three supervisory styles—attractive, interpersonally sensitive, and task-oriented—as measured by the Supervisory Styles Inventory (SSI) used in the present study. Attractive style supervisors appear to be warm, supportive, friendly, open, and flexible, denoting the collegial dimension of supervision; the interpersonally sensitive style is a relationship-oriented approach, and supervisors of this style tend to be invested, committed, therapeutic, and perceptive; and task-oriented supervisors are content-focused, goal-oriented, thorough, focused, practical, and structured (Friedlander & Ward, 1984). These styles resonate with the consultant, counselor, and teacher roles of the supervisor, respectively, in Bernard’s (1997) discrimination model.

Of the three styles, the interpersonally sensitive and task-oriented styles appear to be empirically distinct from one another and distinct from the attractive style (Shaffer & Friedlander, 2017). For instance, Li, Duys, and Vispoel (2020) studied 34 supervisory dyads and found the interpersonally sensitive style was the only discriminant variable, based on which supervisory dyads exhibited statistically different state-transitional patterns (i.e., movement patterns across six common supervision states). Earlier, Fernando and Hulse-Killacky (2005) also found this same style was the only predictor that uniquely and significantly explained supervisees’ satisfaction with supervision, but the task-oriented style was the only significant predictor in explaining supervisees’ perceived self-efficacy.

Supervisory Working Alliance
     Park et al.’s (2019) meta-analysis indicated that the supervisory working alliance was positively related to supervision outcome variables. Bordin (1983) first coined the concept of the supervisory working alliance as a parallel concept to the therapeutic working alliance and introduced the three aspects of the therapeutic working alliance to the alliance in supervision—mutual agreements on the goals, tasks, and bond—which laid the foundation for the adapted Working Alliance Inventory (WAI; Bahrick, 1989) for both supervisors and supervisees. Efstation et al. (1990) instead used three supervisor factors (client focus, rapport, and identification) and two supervisee factors (rapport and client focus) to conceptualize the supervisory working alliance in their Supervisory Working Alliance Inventory (SWAI). In view of the collinearity issue for the goal and task dimensions in the WAI (Hatcher et al., 2020), I adopted the SWAI in the present study.

The working alliance is one of the most robust predictors of outcome in psychotherapy (Norcross, 2011). Although such robust prediction cannot be directly replicated in supervision between the supervisory working alliance and supervision outcome (Goodyear, 2014), scholars (DePue et al., 2016; DePue et al., 2022) have found the supervisory working alliance to be related to the therapeutic working alliance. Specifically, supervisees’ perception of the supervisory working alliance was positively related to their perception of the therapeutic alliance (DePue et al., 2016). However, supervisees’ perception of the supervisory working alliance did not significantly contribute to clients’ perception of the therapeutic working alliance (DePue et al., 2016).

Supervisory Styles and the Supervisory Working Alliance
     Extensive research has documented a close relationship between supervisory styles and the supervisory working alliance (Efstation et al., 1990; Heppner & Handley, 1981; Ladany, Walker, & Melincoff, 2001; Shaffer & Friedlander, 2017). Broadly, as supervisees perceived a greater mixture of supervisory styles in their supervisors (i.e., higher ratings on all three styles; Ladany, Marotta, & Muse-Burke, 2001), supervisees were more likely to report a stronger supervisory working alliance (Li et al., 2021). Despite this global positive correlation, when scholars examined each style independently in relation to each dimension of the supervisory working alliance, such statistical significance was not consistent (Ladany, Walker, & Melincoff, 2001). For instance, in Ladany, Walker, and Melincoff’s (2001) study, participants’ perceptions of an attractive style uniquely and significantly accounted for their perceptions of the bond dimension in alliance, whereas both the interpersonally sensitive and task-oriented styles had this unique and significant association with the task dimension in alliance.

The Moderating Role of Supervisee Levels
     It is not uncommon for a counselor supervisor to start supervision with an expectation of a supervisory style to use (Hart & Nance, 2003). But supervisors have to decide what to address with the supervisee and adopt the most functional style (Bernard, 1997), which could be subject to a myriad of factors, such as contextual factors (Holloway, 1995), cultural considerations (Li et al., 2018), and supervisees’ developmental levels and needs (Friedlander & Ward, 1984; Stoltenberg & McNeill, 2010), among others. Particularly, in Friedlander and Ward’s (1984) study, supervisory styles were differentially related to supervisees’ experience levels. For example, supervisors reported that they were more task-oriented with practicum students but more attractive and interpersonally sensitive with internship students. This interaction effect was also echoed by practicum students’ higher ratings on the task-oriented style but lower ratings on the interpersonally sensitive style, compared to their internship counterparts (Friedlander & Ward, 1984). Similarly, in the study conducted by Li, Duys, and Granello (2020), supervisory dyads with less experienced supervisees tended to be more preoccupied with foundational competencies (e.g., counseling skills and theories, maintenance of standards of service) than dyads with more experienced supervisees. Consistently, more experienced supervisees in Li et al.’s (2019) study were more likely to display positive social emotional behaviors (e.g., self-disclosure, empathy, reflection of feelings, expanding on supervisors’ ideas, praise) in response to supervisors’ opinions, which in turn were more likely to elicit supervisors’ opinions that helped facilitate supervisees’ growth.

However, supervisees’ developmental levels were not always significantly associated with supervision processes or outcomes. For instance, in Bucky et al.’s (2010) study, doctoral-level supervisees did not rate their supervisor characteristics as related to the supervisory working alliance differently based on their developmental levels. Nevertheless, researchers in that study (Bucky et al., 2010) gauged supervisees’ developmental levels based on supervisees’ training progression (i.e., the current level or year level) as commonly practiced (e.g., practicum vs. internship), which may not accurately capture the actual developmental levels of supervisees. Or supervisee levels may not be strikingly distinct in doctoral programs, at least in that sample. In this study, supervisee levels were conceptualized not only as an overall assessment of where supervisees are but with three dimensions (self and other awareness, motivation, and autonomy) aligned with Stoltenberg and McNeill’s (2010) integrative developmental model (IDM) using the Supervisee Levels Questionnaire-Revised (SLQ-R; McNeill et al., 1992).

Statement of Purpose
     Although literature evidenced the overall positive correlation between supervisory styles and the supervisory working alliance, the direction and strength of such a relationship in different contexts warrants additional attention. Particularly, supervisees’ developmental progression entails a flexible mixture of different supervisory styles as suggested theoretically and empirically, but whether and how the relationship between supervisory styles and the supervisory working alliance may vary across different supervisee levels calls for further investigation. To this end, the purpose of the current study was to test the potential moderation effects of supervisee levels on the relationship between supervisory styles and the supervisory working alliance.

Given that supervisees at earlier stages of professional development may need more guidance and support from supervisors, which necessitates a variety of supervision styles that are critical to their perception of the working alliance with their supervisors, I hypothesized that the positive relationship between supervisory styles and the supervisory working alliance would be more sensitive for supervisees at earlier stages of development, compared to their more experienced counterparts. In other words, the positive relationship would be stronger for supervisees at lower levels of professional development and weaker for supervisees at higher levels of professional development.

Method

Participants
     The data set of this study is part of a larger national quantitative study with a cross-sectional sample (Li et al., 2021). Yet, researchers have not examined supervisee levels that are crucial to measuring supervisee development using a robust psychometric instrument. The current sample comprised 113 participants (see Table 1), with the majority as master’s-level (n = 54, 47.79%) or doctoral-level students (n = 46, 40.71%). Approximately 17% of participants (n = 19) identified themselves as post-master’s or post-doctoral practitioners or other. Some participants reported both their training and practicing levels (e.g., both as a doctoral student and a post-master’s practitioner), which caused the sample size to be larger than 113 if simply adding the frequencies across the three categories together. Most participants reported their specialty areas in clinical mental health counseling (n = 53, 46.90%), school counseling (n = 43, 38.05%), and counselor education and supervision (n = 27, 23.89%). Because some participants indicated more than one specialty area, the total percentage did not add up to 100.

In this sample, approximately 80% were female (n = 90) and 23 were male (20.35%). At the time of filling out the questionnaire, most of them fell in the 21–30 age range (n = 72, 63.72%), with 19 in the 31–40 range (16.81%), 13 in the 41–50 range (11.50%), and nine beyond 50 years old (7.96%). Participants in this sample predominantly identified themselves as White (n = 97; 85.84%), with eight as Asian (7.08%), five as Black or African American (4.42%), one as American Indian and Alaska Native (0.88%), one as biracial or multiracial (0.88%), and one indicating other (0.88%). Most participants reported their counseling experience as 1 year or less (n = 44, 38.94%) or longer than 3 years (n = 37; 32.74%), with the rest reporting in between (n = 32, 28.31%). See Table 1 for more detailed demographic information.

Procedure
     Upon receiving IRB approval, I started collecting data online through Qualtrics in 2017–2018. The recruitment criteria included (a) one is at least 18 years of age by the time of filling out the survey; and (b) one is a student or a practitioner who had supervision experience in the counseling field. I disseminated the recruitment post through several professional networks, including the Counselor Education and Supervision Network-Listserv (CESNET-L) and American Counseling Association (ACA) Connect. In addition to this convenience sampling, I also used snowball sampling because participants were encouraged to share the recruitment post with anyone who they thought might be eligible to participate in the study. The recruitment post contained a survey link that directed potential participants to the informed consent webpage and then a compiled questionnaire webpage.

Instruments
Demographic Questionnaire
     The purpose of including this self-constructed Demographic Questionnaire was to report the basic demographic information of participants. Specifically, the questionnaire included the gender, age, race/ethnicity, length of counseling-related work experience, training/practicing level, and training or specialty area of participants.

Supervisory Styles Inventory
     The SSI (Friedlander & Ward, 1984) is a 33-item instrument used to measure the degree to which one endorses descriptors representative of each of the three dimensions of supervisory style: Attractive (7 items), Interpersonally Sensitive (8 items), and Task-Oriented (10 items), with the remainder as the filler items (8 items). Participants rate each item along a 7-point Likert scale from 1 (not very) to 7 (very). Higher scores in each dimension mean that one endorses descriptors of a certain supervisory style to a larger extent. Sample items for the Attractive, Interpersonally Sensitive, and Task-Oriented subscales are “supportive,” “perceptive,” and “didactic,” respectively.

Friedlander and Ward (1984) reported the Cronbach’s alphas of the three subscales separately and combined ranged from .76 to .93 (Ns ranging from 105 to 202). Additionally, the item–scale correlations ranged from .70 to .88 for the Attractive subscale, from .51 to .82 for the Interpersonally Sensitive style, and from .38 to .76 for the Task-Oriented scale (N1 = 202, N2 = 183; Friedlander & Ward, 1984). The test-retest reliability (N = 32) for the combined scale was .92; they were .94, .91, and .78 for the Attractive, Interpersonally Sensitive, and Task-Oriented subscales, respectively (Friedlander & Ward, 1984). They also reported the convergent validity based on moderate to high positive relationships (ps < .001) between the SSI and Stenack and Dye’s (1982) measure of supervisor roles (i.e., consultant, counselor, and teacher; N = 90). In the present study, the Cronbach’s alpha was .96 for the Attractive style, .94 for the Interpersonally Sensitive style, .92 for the Task-Oriented style, and .96 for the entire measure.

Supervisory Working Alliance Inventory
     The SWAI (Efstation et al., 1990) is used to measure the relationship in counselor supervision. It has both the supervisor and supervisee forms. The supervisee form applied to the current study includes two scales: Rapport (12 items) and Client Focus (7 items). Supervisees indicate the extent to which the behavior described in each item seems characteristic of their work with their supervisors on a 7-point Likert scale, with 1 being almost never and 7 being almost always. Higher scores in the Rapport scale indicate a stronger perceived rapport with their supervisor, and higher scores in the Client Focus scale suggest more attention to issues related to the client in supervision. A sample item for the Rapport scale is “I feel free to mention to my supervisor any troublesome feelings I might have about him/her.” A sample item for the Client Focus scale is “I work with my supervisor on specific goals in the supervisory session.”

Efstation et al. (1990) reported that the alpha coefficient was .90 for Rapport and .77 for Client Focus (N = 178) for the supervisee form. Moreover, the item–scale correlations ranged from .44 to .77 for Rapport, and from .37 to .53 for Client Focus. They used the SSI to obtain initial estimates of convergent and divergent validity for the SWAI (Efstation et al., 1990). As expected, the Client Focus dimension of the SWAI showed moderate correlation (r = .52) with the Task-Oriented style in the SSI supervisee’s form, but low correlation (r = .04) with the Attractive style and low correlation (r = .21) with the Interpersonally Sensitive style. The Rapport dimension from the SWAI had low correlation (r < .00) with the Task-Oriented style of the SSI. In the present study, the Cronbach’s alpha was .95 for Rapport, .90 for Client Focus, and .96 for the entire scale.

Supervisee Levels Questionnaire-Revised
     The Supervisee Levels Questionnaire-Revised (SLQ-R; McNeill et al., 1992) is used to measure supervisees’ developmental levels (Stoltenberg & Delworth, 1987). It has 30 items developed around three dimensions: Self and Other Awareness (12 items), Motivation (8 items), and Dependency-Autonomy (10 items). Supervisees can indicate their current behavior along a 7-point Likert scale, with 1 representing never, 2 rarely, 3 sometimes, 4 half the time, 5 often, 6 most of the time, and 7 always. Higher scores (after reverse-scoring for some of the items) in these dimensions reflect higher levels of supervisee development in Self and Other Awareness, Motivation, and Autonomy, respectively. A sample item for the Self and Other Awareness dimension is “I feel genuinely relaxed and comfortable in my counseling/therapy sessions”; a sample item (reverse-scoring) for the Motivation dimension is “The overall quality of my work fluctuates; on some days I do well, on other days, I do poorly”; and a sample item for the Dependency-Autonomy dimension is “I am able to critique counseling tapes and gain insights with minimum help from my supervisor.”

McNeill et al. (1992) reported that the Cronbach alpha coefficients of the SLQ-R (N = 105) were .83, .74, and .64 for the three subscales, respectively, and .88 for the total scores. To assess the construct validity of the SLQ-R, they examined the differences in subscale and total scores across the beginning, intermediate, and advanced groups. Hotelling’s test of significance indicated that the three groups differed significantly both on the total SLQ-R scores, F(2, 102) = 7.37, p < .001, and on a linear combination of SLQ-R subscale scores, F(6, 198) = 2.45, p < .026. In the present study, the Cronbach’s alpha was .89 for Self and Other Awareness, .85 for Motivation, .57 for Autonomy, and .91 for the entire measure.

Data Analysis
     To thoroughly test the potential moderation effects of supervisee levels on the relationship between supervisory styles and the supervisory working alliance, I carried out three rounds of moderation analysis in which the supervisory working alliance was always the outcome variable. In the first round (n = 1), supervisory styles as a whole were the predictor, and supervisee levels as a whole were the moderator. The second round (n = 6) involved two series of analyses. In the first series (n = 3), each supervisory style was the predictor, and supervisee levels as a whole were the moderator. In the second series (n = 3), supervisory styles as a whole were the predictor, and each indicator of supervisee levels was the moderator. In the third round (n = 9), each supervisory style was the predictor, and each indicator of supervisee levels was the moderator. Figure 1 presents path diagrams of three rounds of tests and Table 2 lists all tested models (n = 16).

I followed up each significant moderation effect (n = 5) with a simple slopes analysis (Aiken & West, 1991) to interpret the nature of the interaction effect. The PROCESS v4.0 tool in SPSS was employed to perform all these analyses. A total of 166 potential participants accessed the survey, but only 113 of them completed all the study instruments (SSI, SWAI, and SLQ-R) in the present study. To alleviate the impact of significantly incomplete responses, I removed the 53 respondents who left at least one instrument unanswered. The a priori power analysis via G*Power 3.1.9.7 indicated that the minimum sample size would be 55 to detect an interaction effect with a medium effect size (f 2 = .15), given the desired statistical power level of .80 and type I error rate of .05. As such, the ultimate sample size of 113 meets this requirement.

I made the linearity and homoscedasticity assumptions using the zpred vs. zresid plot, which did not show a systematic relationship between the predicted values and the errors in the model (Field, 2017). Provided that participants independently filled out the study survey, I held the assumption of independence that the errors in the model were not dependent on each other. Further screening detected 12 missing values scattered across the three scales, which accounted for 0.13% of the entire 9,266 possible values. To determine the nature of these missing values, I performed the Little’s test (1988), and the results signified that these values were missing completely at random (MCAR; χ2 = 884.185, df = 890, p = .549). Because multiple imputation (MI; Schafer, 1999) can provide unbiased and valid estimates of associations based on information from the available data and can handle MCAR (Pedersen et al., 2017), I adopted MI to replace the missing values before performing further analyses in this study.

Results

Results of this study in part supported my broad hypothesis that the positive relationship between supervisory styles and the supervisory working alliance would be more sensitive for supervisees at earlier stages of development, compared to their more experienced counterparts. Examining each supervisory style and each indicator of supervisee levels independently revealed the intricacy of the relationship between the two constructs.

There were two groups of major findings. First, supervisee levels as a whole were a significant moderator between the interpersonally sensitive style and the supervisory working alliance according to supervisees’ perceptions, ΔR2 = .0272, F(1, 109) = 7.8551, p = .006, with a small to medium effect size
(f 2 = .07; Lorah & Wong, 2018). Specifically, the strength of the relationship between the interpersonally sensitive style and the supervisory working alliance differed based on supervisee levels (see Table 3).

In view of this significant moderation effect, I conducted a simple slopes analysis as a follow-up, which indicated that the simple slopes for 1 standard deviation (SD) below the mean, at the mean, and 1 SD above the mean of supervisee levels were 1.6185, 1.4019, and 1.1853, respectively (see Figure 2). In other words, the interpersonally sensitive style and the supervisory working alliance were positively associated (B = 1.4019, p < .001), but the strength of this correlation decreased as supervisees reported higher levels of professional development. It is worth noting that supervisees at higher developmental levels tended to report a stronger supervisory working alliance in general, compared to those at lower levels. The linear model of the interpersonally sensitive style, supervisee levels, and the product of the two (interpersonally sensitive style × supervisee levels) explained 62.31% (p < .001) of the variance in the supervisory working alliance. A further look into the moderation effect of supervisee levels indicated that statistical significance consistently persisted as each indicator of supervisee levels (self and other awareness, motivation, and autonomy) was independently tested as a moderator between the interpersonally sensitive style and the supervisory working alliance (see Round 3 in Table 2).

Figure 2
Moderation Effect of Supervisee Levels With the Interpersonally Sensitive Style on the Supervisory Working Alliance

Note. N = 113. Predictor = Interpersonally Sensitive Style; Moderator = Supervisee Levels; Outcome = Supervisory Working Alliance. The three lines of color represent three regressions with the interpersonally sensitive style as predictor and the supervisory working alliance as outcome at different supervisee levels. The blue regression line denotes the group in which supervisee levels were one standard deviation (SD) below the mean, the green denotes the group in which supervisee levels were at the mean, and the pink denotes the group in which supervisee levels were one SD above the mean.

The second major finding was about the task-oriented supervisory style. When the three indicators of supervisee levels were independently examined as moderators, it was found that self and other awareness moderated the relationship between the task-oriented style and the supervisory working alliance, ΔR2 = .0311, F(1, 109) = 5.0639, p = .0264, with a small to medium effect size (f 2 = .05; Lorah & Wong, 2018). Similar to the first group of findings, the strength of the relationship between the task-oriented style and the supervisory working alliance varied based on the level of supervisee self and other awareness (one indicator of supervisee levels; see Table 4). A simple slopes analysis signified a consistent pattern—the task-oriented style and the supervisory working alliance were positively correlated, but the strength of this relationship decreased as supervisees rated higher on self and other awareness (see Figure 3). Specifically, the simple slopes for one SD below the mean, at the mean, and one SD above the mean of supervisee self and other awareness were 1.2620, 0.9540, and 0.6460, respectively. The area below the moderator (self and other awareness) value of 13.3857 constituted a region of significance in which the relationship between the task-oriented style and the supervisory working alliance was significant (p < .05; Johnson & Neyman, 1936). The linear model of the task-oriented style, supervisee self and other awareness, and the product of the two (task-oriented style × self and other awareness) accounted for 33.13% (p < .001) of the variance in the supervisory working alliance.

Discussion

Findings of the present study corroborated the positive correlation between supervisory styles and the supervisory working alliance that has been consistently identified in the existing literature (Efstation et al., 1990; Heppner & Handley, 1981; Ladany & Lehrman-Waterman, 1999; Ladany, Walker, & Melincoff, 2001). The intricacy of this relationship was further explored, and the current study confirmed that the strength of such correlation varied across different contexts. Supervisee levels and their three indicators turned out to be significant moderators in five models out of the 16 tested. Explicitly, the positive correlation between the interpersonally sensitive style and the supervisory working alliance was stronger for supervisees at lower levels of professional development but weaker for supervisees at higher levels. Furthermore, this significant moderation effect existed not only when supervisee levels were viewed as an overarching construct but when each indicator of supervisee levels was independently examined. Moreover, this moderation pattern was echoed by the positive association between the task-oriented style and the supervisory working alliance, wherein the correlation was stronger for supervisees at lower levels of self and other awareness (one indicator of supervisee levels) but weaker for those at higher levels of self and other awareness. Notably, supervisees at higher developmental levels (including indicators of supervisee levels) in all models with significant moderation effects reported a stronger supervisory working alliance than did their counterparts at lower levels.

According to developmental theories of supervision, supervisees broadly progress through a series of qualitatively different levels in the process of becoming effective counselors, despite myriad individual idiosyncrasies (Chagnon & Russell, 1995; Stoltenberg & McNeill, 2010). Entry-level supervisees typically focus on their own anxiety, their lack of skills and knowledge, and the likelihood that they are being regularly evaluated (Stoltenberg & McNeill, 2010). Accordingly, beginning supervisees identified supervisor care and concern as one of the most important supervisor variables to allow supervisees to take risks and grow (Jordan, 2007). As such, interpersonally sensitive supervisors who are invested, committed, therapeutic, and perceptive (Friedlander & Ward, 1984) would be easily perceived as relationship-oriented and helpful in rapport building (one indicator of the supervisory working alliance) for supervisees early on in their training. Similarly, task-oriented supervisors are content-focused, goal-oriented, thorough, focused, practical, and structured (Friedlander & Ward, 1984).


Figure 3
Moderation Effect of Self and Other Awareness With the Task-Oriented Style on the Supervisory Working Alliance

Note. N = 113. Predictor = Task-Oriented Style; Moderator = Self and Other Awareness; Outcome = Supervisory Working Alliance. The three lines of color represent three regressions with the task-oriented style as predictor and the supervisory working alliance as outcome at different levels of self and other awareness (one indicator of supervisee levels). The blue regression line denotes the group in which supervisee self and other awareness was one standard deviation (SD) below the mean, the green denotes the group in which supervisee self and other awareness was at the mean, and the pink denotes the group in which supervisee self and other awareness was one SD above the mean.

Task-oriented supervisors can be perceived as particularly helpful and informative with client focus (a second indicator of the supervisory working alliance) for beginning supervisees (as indicated by their lower self and other awareness) who commonly experience substantial anxiety or fear pertaining to their lack of confidence in knowing what to do, being able to do it, and being evaluated by their clients or supervisors (Stoltenberg & McNeill, 2010).

Therefore, supervisees at lower levels of professional development were more likely to report a stronger supervisory working alliance as they perceived more interpersonally sensitive or task-oriented supervisor characteristics. As they progress to higher levels of development with accumulated knowledge, skills, and competencies, supervisees become more aware of clients and themselves, intrinsically and consistently motivated, and autonomous as practitioners (Stoltenberg & McNeill, 2010), which may in part explain why their ratings of the supervisory working alliance were less related to their perceptions of supervisor characteristics but generally higher than supervisees at lower levels of development.

In the present study, the moderator of supervisee levels as a composite score was only significant when the interpersonally sensitive style was the predictor; the moderator of self and other awareness (one indicator of supervisee levels) was also significant when the task-oriented style was the predictor. These findings resonated with the existing literature in that compared to the attractive style, the interpersonally sensitive and task-oriented styles tend to have stronger discriminating effects (Friedlander & Ward, 1984). For instance, practicum and internship students differed significantly in rating the task-oriented and interpersonally sensitive styles of their supervisors, but their perceptions about the attractive style were similar at both levels (Friedlander & Ward, 1984). Li, Duys, and Vispoel (2020) also found that supervisory state–transitional patterns differed significantly only based on the interpersonally sensitive style but not the other two styles.

Implications for Clinical Supervision
     The supervisory working alliance is inextricably intertwined with supervisees’ willingness to disclose (Ladany et al., 1996), supervisee satisfaction with clinical supervision (Cheon et al., 2009; Ladany, Ellis, & Friedlander, 1999), supervisee work satisfaction and work-related stress (Sterner, 2009), and therapeutic working alliance (DePue et al., 2016; DePue et al., 2022), among others. Nelson et al. (2001) proposed that a key task in early supervision is to build a strong supervisory working alliance that serves as a foundation to manage future potential dilemmas in supervision, and the ongoing maintenance of this working alliance should be the supervisor’s responsibility throughout the supervisory relationship. Although the three supervisory styles appear to be clear-cut with distinguishable characteristics and roles (Friedlander & Ward, 1984), supervisors are encouraged to adopt a composite of different styles to varying degrees to better serve supervisees’ needs. As revealed by the present study, and also the extant literature (Efstation et al., 1990; Ladany, Walker, & Melincoff, 2001; Li et al., 2021), supervisees were more likely to report a stronger supervisory working alliance as they perceived their supervisors to adopt a mixture of three supervisory styles (i.e., higher overall ratings of supervisory styles).

Particularly, beginning supervisees are characteristic of a strong focus on self, extrinsic motivation, and high dependency on supervisors (Stoltenberg & McNeill, 2010). Supervisors’ emphases on relationship-building (interpersonally sensitive style) and task focus (task-oriented style) would help build a safe, predictable supervision environment and enhance the working alliance with supervisees. Notably, although the strengths of the correlation between the interpersonally sensitive or task-oriented style and the supervisory working alliance were stronger for beginning supervisees, they did not suggest that these styles would not be effective in augmenting the alliance for supervisees at higher levels of professional development. The positive correlations still existed, albeit smoother, for more advanced supervisees, and they reported higher levels of supervisory working alliance in general, which may imply that these styles help maintain the working alliance that has been established early on in supervision.

Another point that is worth noting is that although no significant moderator was detected between the attractive style and the supervisory working alliance in the present study, the attractive style explained the most variance (68.1%, p < .001) in the supervisory working alliance, compared to the interpersonally sensitive (55.9%, p < .001) and task-oriented styles (24.1%, p < .001). This finding made it clear that the warm, supportive, friendly, open, and flexible features of attractive style supervisors are foundational to building and maintaining the supervisory working alliance, which does not differentiate across different levels of supervisees. As such, supervisors are encouraged to bring these qualities to their supervision and make them perceived by supervisees.

Limitations and Future Research
     This study is not exempt from limitations that may be addressed in future research. Although two moderators (supervisee levels, self and other awareness) were found to be significant in the present study, the effect sizes of both were small to medium (f 21 = .07 and f 22 = .05), which were lower than the speculated medium effect size (f 2 = .15) during the a priori power analysis. Provided the effect sizes of .07 and .05 for the moderation effect, to achieve the statistical power of .80 with the α error probability of .05, the required sample size would be 115 and 159, respectively. Researchers need to be more mindful when recruiting participants to ensure the sufficient sample size. Additionally, although supervisees were asked to respond to the questionnaires consistently based on their perceptions of one supervisor, a constellation of factors could have affected their perceptions—for example, the timing of a participant’s supervisee status (e.g., currently receiving supervision vs. received supervision in the past), the potential dual role that a participant may be in (e.g., a doctoral student who is both a supervisee and a supervisor), the level of supervision (e.g., practicum, internship), and the length of the supervisory relationship (e.g., 2 months vs. 2 years). Researchers in future studies could also collect more information about participants (e.g., geographic distribution) to help readers better contextualize study results. Also, the current data set was collected in 2017–2018, which would not be able to capture more recent societal, cultural, political, and economic changes (e.g., the COVID-19 pandemic) that could have affected supervisee perceptions.

In the present study, the association between supervisory styles and the supervisory working alliance was examined in the context of different supervisee levels. Indeed, this alliance could be subject to many other factors, such as discussions of cultural variables in supervision (Gatmon et al., 2001), supervisor adherence to ethical guidelines (Ladany & Lehrman-Waterman, 1999), and relational supervision strategies (Shaffer & Friedlander, 2017), among others. Scholars may include more related variables to expand the current model so as to further disentangle the complex relationships among predictors of the supervisory working alliance.

Last, although multiple moderation effects identified in the present study were statistically significant and theoretically coherent, exactly how supervisees experience the supervisory working alliance in relation to different supervisory styles as they proceed along the professional development is less known. A longitudinal track of the same sample using repeated measures or a qualitative inquiry into participants’ lived experiences of the targeted phenomenon could enrich our understanding of the study variables in this research.

Conclusion

Although the positive correlation between supervisory styles and the supervisory working alliance is well documented in the existing literature, the present study examined such relationships specifically in the context of supervisee levels. Both supervisee levels (as a whole) and self and other awareness (one indicator of supervisee levels) appeared to be significant moderators under different contexts. These findings further revealed the intricacies embedded in the broad relationship between supervisory styles and the supervisory working alliance, pointed out future research directions concerning supervisee development, and encouraged supervisors to adopt a composite of styles to varying degrees to better support supervisee growth.

Conflict of Interest and Funding Disclosure
The authors reported no conflict of interest
or funding contributions for the development
of this manuscript.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. SAGE.

Bahrick, A. S. (1989). Role induction for counselor trainees: Effects on the supervisory working alliance (Order No. 9014392) [Doctoral dissertation, The Ohio State University]. ProQuest Dissertations and Theses Global.

Bernard, J. M. (1997). The discrimination model. In C. E. Watkins, Jr. (Ed.), Handbook of psychotherapy supervision (pp. 310–327). Wiley.

Bernard, J. M., & Goodyear, R. K. (2019). Fundamentals of clinical supervision (6th ed.). Pearson.

Bordin, E. S. (1983). A working alliance based model of supervision. The Counseling Psychologist, 11(1), 35–42. https://doi.org/10.1177/0011000083111007

Bucky, S. F., Marques, S., Daly, J., Alley, J., & Karp, A. (2010). Supervision characteristics related to the supervisory working alliance as rated by doctoral-level supervisees. The Clinical Supervisor, 29(2), 149–163. https://doi.org/10.1080/07325223.2010.519270

Chagnon, J., & Russell, R. K. (1995). Assessment of supervisee developmental level and supervision environment across supervisor experience. Journal of Counseling & Development, 73(5), 553–558.
https://doi.org/10.1002/j.1556-6676.1995.tb01793.x

Cheon, H.-S., Blumer, M. L. C., Shih, A.-T., Murphy, M. J., & Sato, M. (2009). The influence of supervisor and supervisee matching, role conflict, and supervisory relationship on supervisee satisfaction. Contemporary Family Therapy, 31(1), 52–67. https://doi.org/10.1007/s10591-008-9078-y

DePue, M. K., Lambie, G. W., Liu, R., & Gonzalez, J. (2016). Investigating supervisory relationships and therapeutic alliances using structural equation modeling. Counselor Education and Supervision, 55(4), 263–277. https://doi.org/10.1002/ceas.12053

DePue, M. K., Liu, R., Lambie, G. W., & Gonzalez, J. (2022). Examining the effects of the supervisory relationship and therapeutic alliance on client outcomes in novice therapists. Training and Education in Professional Psychology, 16(3), 253–262. https://doi.org/10.1037/tep0000320

Efstation, J. F., Patton, M. J., & Kardash, C. M. (1990). Measuring the working alliance in counselor supervision. Journal of Counseling Psychology, 37(3), 322–329. https://doi.org/10.1037/0022-0167.37.3.322

Fernando, D. M., & Hulse-Killacky, D. (2005). The relationship of supervisory styles to satisfaction with supervision and the perceived self-efficacy of master’s-level counseling students. Counselor Education and Supervision, 44(4), 293–304. https://doi.org/10.1002/j.1556-6978.2005.tb01757.x

Field, A. (2017). Discovering statistics using IBM SPSS statistics (5th ed.). SAGE.

Friedlander, M. L., & Ward, L. G. (1984). Development and validation of the Supervisory Styles Inventory. Journal of Counseling Psychology, 31(4), 541–557. https://doi.org/10.1037/0022-0167.31.4.541

Gatmon, D., Jackson, D., Koshkarian, L., Martos-Perry, N., Molina, A., Patel, N., & Rodolfa, E. (2001). Exploring ethnic, gender, and sexual orientation variables in supervision: Do they really matter? Journal of Multicultural Counseling and Development, 29(2), 102–113. https://doi.org/10.1002/j.2161-1912.2001.tb00508.x

Goodyear, R. K. (2014). Supervision as pedagogy: Attending to its essential instructional and learning processes. The Clinical Supervisor, 33(1), 82–99. https://doi.org/10.1080/07325223.2014.918914

Hart, G. M., & Nance, D. (2003). Styles of counselor supervision as perceived by supervisors and supervisees. Counselor Education and Supervision, 43(2), 146–158. https://doi.org/10.1002/j.1556-6978.2003.tb01838.x

Hatcher, R. L., Lindqvist, K., & Falkenström, F. (2020). Psychometric evaluation of the Working Alliance Inventory—Therapist version: Current and new short forms. Psychotherapy Research, 30(6), 706–717. https://doi.org/10.1080/10503307.2019.1677964

Heppner, P. P., & Handley, P. G. (1981). A study of the interpersonal influence process in supervision. Journal of Counseling Psychology, 28(5), 437–444. https://doi.org/10.1037/0022-0167.28.5.437

Holloway, E. (1995). Clinical supervision: A systems approach. SAGE.

Hunt, D. E. (1971). Matching models in education: The coordination of teaching methods with student characteristics. Ontario Institute for Studies in Education, Monograph, 10, 87.

Johnson, P. O., & Neyman, J. (1936). Tests of certain linear hypotheses and their application to some educational problems. Statistical Research Memoirs, 1, 57–93.

Jordan, K. (2007). Beginning supervisees’ identity: The importance of relationship variables and experience versus gender matches in the supervisee/supervisor interplay. The Clinical Supervisor, 25(1–2), 43–51.
https://doi.org/10.1300/J001v25n01_04

King, K. M., Borders, L. D., & Jones, C. T. (2020). Multicultural orientation in clinical supervision: Examining impact through dyadic data. The Clinical Supervisor, 39(2), 248–271.
https://doi.org/10.1080/07325223.2020.1763223

Ladany, N., Ellis, M. V., & Friedlander, M. L. (1999). The supervisory working alliance, trainee self-efficacy, and satisfaction. Journal of Counseling & Development, 77(4), 447–455.
https://doi.org/10.1002/j.1556-6676.1999.tb02472.x

Ladany, N., Hill, C. E., Corbett, M. M., & Nutt, E. A. (1996). Nature, extent, and importance of what psychotherapy trainees do not disclose to their supervisors. Journal of Counseling Psychology, 43(1), 10–24. https://doi.org/10.1037/0022-0167.43.1.10

Ladany, N., & Lehrman-Waterman, D. E. (1999). The content and frequency of supervisor self-disclosures and their relationship to supervisor style and the supervisory working alliance. Counselor Education and Supervision, 38(3), 143–160. https://doi.org/10.1002/j.1556-6978.1999.tb00567.x

Ladany, N., Marotta, S., & Muse-Burke, J. L. (2001). Counselor experience related to complexity of case conceptualization and supervision preference. Counselor Education and Supervision, 40(3), 203–219.
https://doi.org/10.1002/j.1556-6978.2001.tb01253.x

Ladany, N., Walker, J. A., & Melincoff, D. S. (2001). Supervisory style: Its relation to the supervisory working alliance and supervisor self-disclosure. Counselor Education and Supervision, 40(4), 263–275.
https://doi.org/10.1002/j.1556-6978.2001.tb01259.x

Li, D., Duys, D. K., & Granello, D. H. (2019). Interactional patterns of clinical supervision: Using sequential analysis. Asia Pacific Journal of Counselling and Psychotherapy, 10(1), 70–92.
https://doi.org/10.1080/21507686.2018.1553791

Li, D., Duys, D. K., & Granello, D. H. (2020). Applying Markov chain analysis to supervisory interactions. The Journal of Counselor Preparation and Supervision, 13(1), Article 6. https://digitalcommons.sacredheart.edu/cgi/viewcontent.cgi?article=1325&context=jcps

Li, D., Duys, D. K., & Liu, Y. (2021). Working alliance as a mediator between supervisory styles and supervisee satisfaction. Teaching and Supervision in Counseling, 3(3), Article 5. https://doi.org/10.7290/tsc030305

Li, D., Duys, D. K., & Vispoel, W. P. (2020). Transitional dynamics of three supervisory styles using Markov chain analysis. Journal of Counseling & Development, 98(4), 363–375. https://doi.org/10.1002/jcad.12339

Li, D., Liu, Y., & Lee, I. (2018). Supervising Asian international counseling students: Using the integrative developmental model. Journal of International Students, 8(2), 1129–1151. https://doi.org/10.32674/jis.v8i2.137

Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83(404), 1198–1202. https://doi.org/10.1080/01621459.1988.10478722

Lorah, J. A., & Wong, Y. J. (2018). Contemporary applications of moderation analysis in counseling psychology. Journal of Counseling Psychology, 65(5), 629–640. https://doi.org/10.1037/cou0000290

McNeill, B. W., Stoltenberg, C. D., & Romans, J. S. (1992). The Integrated Developmental Model of supervision: Scale development and validation procedures. Professional Psychology: Research and Practice, 23(6), 504–508. https://doi.org/10.1037/0735-7028.23.6.504

Munson, C. E. (1993). Clinical social work supervision (2nd ed.). Haworth Press.

Nelson, M. L., Gray, L. A., Friedlander, M. L., Ladany, N., & Walker, J. A. (2001). Toward relationship-centered supervision: Reply to Veach (2001) and Ellis (2001). Journal of Counseling Psychology, 48(4), 407–409.
https://doi.org/10.1037/0022-0167.48.4.407

Norcross, J. C. (Ed.). (2011). Psychotherapy relationships that work: Evidence-based responsiveness (2nd ed.). Oxford University Press.

Park, E. H., Ha, G., Lee, S., Lee, Y. Y., & Lee, S. M. (2019). Relationship between the supervisory working alliance and outcomes: A meta-analysis. Journal of Counseling & Development, 97(4), 437–446.
https://doi.org/10.1002/jcad.12292

Pedersen, A. B., Mikkelsen, E. M., Cronin-Fenton, D., Kristensen, N. R., Pham, T. M., Pedersen, L., & Petersen, I. (2017). Missing data and multiple imputation in clinical epidemiological research. Clinical Epidemiology, 9, 157–166. https://doi.org/10.2147/CLEP.S129785

Schafer, J. L. (1999). Multiple imputation: A primer. Statistical Methods in Medical Research, 8(1), 3–15.
https://doi.org/10.1177/096228029900800102

Shaffer, K. S., & Friedlander, M. L. (2017). What do “interpersonally sensitive” supervisors do and how do supervisees experience a relational approach to supervision? Psychotherapy Research, 27(2), 167–178. https://doi.org/10.1080/10503307.2015.1080877

Stenack, R. J., & Dye, H. A. (1982). Behavioral descriptions of counseling supervision roles. Counselor Education and Supervision, 21(4), 295–304. https://doi.org/10.1002/j.1556-6978.1982.tb01692.x

Sterner, W. (2009). Influence of the supervisory working alliance on supervisee work satisfaction and work-related stress. Journal of Mental Health Counseling, 31(3), 249–263.
https://doi.org/10.17744/mehc.31.3.f3544l502401831g

Stoltenberg, C. D., & Delworth, U. (1987). Supervising counselors and therapists: A developmental approach. Jossey-Bass.

Stoltenberg, C. D., & McNeill, B. W. (2010). IDM supervision: An integrative developmental model for supervising counselors and therapists (3rd ed.). Routledge.

Dan Li, PhD, NCC, LSC (NC, K–12), is an assistant professor of counseling at the University of North Texas. Correspondence may be addressed to Dan Li, Welch Street Complex 2-112, 425 S. Welch St., Denton, TX 76201, Dan.Li@unt.edu.

Individual and Relational Predictors of Compassion Fatigue Among Counselors-in-Training

Nov 28, 2019 | Volume 9 - Issue 4

Nesime Can, Joshua C. Watson

 

Scholars have described compassion fatigue as the result of chronic exposure to clients’ suffering and traumatic stories. Counselors can struggle when they experience compassion fatigue because of various reasons. As such, an exploration of factors predictive of compassion fatigue may help counselors and supervisors buffer adverse effects. Utilizing a hierarchical linear regression analysis, we examined the association between wellness, resilience, supervisory working alliance, empathy, and compassion fatigue among 86 counselors-in-training (CITs). The research findings revealed that resilience and wellness were significant predictors of compassion fatigue among CITs, whereas empathy and supervisory working alliance were not. Based on our findings, counselor educators might consider enhancing their current training programs by including discussion topics about wellness and resilience, while supervisors consider practicing wellness and resilience strategies in supervision and developing interventions designed to prevent compassion fatigue.

 

Keywords: compassion fatigue, counselors-in-training, wellness, resilience, supervisory working alliance

 

 

Balancing self-care and client care can be a challenge for many counselors. When counselors neglect self-care, they can become vulnerable to several issues, including increased anxiety, distress, burnout, and compassion fatigue (Ray, Wong, White, & Heaslip, 2013). Counselors might be especially prone to experiencing compassion fatigue because they repeatedly hear traumatic stories and clients’ suffering in sessions (Skovholt & Trotter-Mathison, 2016). This phenomenon is likely pronounced among counselors-in-training (CITs), as lack of experience, skillset, knowledge, and support can lead to struggles when working with clients (Skovholt & Trotter-Mathison, 2016). Coupled with the increased anxiety, distress, and disappointment, CITs can experience compassion fatigue early in their career development, which can lead to exhaustion, disengagement, and a decline in therapeutic effectiveness (Rønnestad & Skovholt, 2013). At this developmental stage, negative experiences can lead to feelings of doubt and a lack of confidence among CITs and potentially lead to career dissatisfaction. Therefore, it is essential and necessary to better understand the predictive factors of compassion fatigue among CITs to prevent its early onset.

 

Compassion Fatigue in Counseling

 

Counselors listening to their clients’ fear, pain, and suffering can feel similar emotions. Figley (1995) defined this experience as compassion fatigue; it also can be defined as the cost of caring (Figley, 2002). Whether working in mental health agencies, schools, or hospital settings, counselors experience compassion fatigue because of exposure to large caseloads, painful stories, and lack of support and resources (Skovholt & Trotter-Mathison, 2016). Despite this exposure, counselors are expected to place their personal feelings aside and provide the best treatment possible in response to the presenting issues and needs of their clients (Figley, 2002; Ray et al., 2013; Turgoose, Glover, Barker, & Maddox, 2017). Maintaining this sense of detached professionalism has its costs, as a number of counselors find themselves at risk for experiencing physical, mental, and emotional exhaustion, as well as feelings of helplessness, isolation, and confusion—a situation collectively referred to as compassion fatigue (Eastwood & Ecklund, 2008; Thompson, Amatea, & Thompson, 2014).

 

Merriman (2015b) stated that ongoing compassion fatigue negatively impacts counselors’ health as well as their relationships with others. Additionally, compassion fatigue can lead to a lack of empathy toward clients, decrease in motivation, and performance drop in effectiveness, making even the smallest tasks seem overwhelming (Merriman, 2015b). When this occurs, counselors can project their anger on others, develop trust issues, and experience feelings of loneliness (Harr, 2013). Therefore, the demands of the counseling profession can affect many counselors’ wellness and potentially could hurt the quality of client care provided (Lawson, Venart, Hazler, & Kottler, 2007; Merriman, 2015a). Further, counselors experiencing compassion fatigue might have difficulties making effective clinical decisions and potentially be at risk for harming clients (Eastwood & Ecklund, 2008). Consequently, scholars appear to agree that compassion fatigue is an occupational hazard that mental health care professionals need to address (Figley, 2002; Merriman, 2015a).

 

Factors Associated With Compassion Fatigue

 

Many researchers have studied the relationships between compassion fatigue and various constructs, such as empathy, gender, mindfulness, support, and wellness (e.g., Beaumont, Durkin, Martin, & Carson, 2016; Caringi et al., 2016; Ray et al., 2013; Sprang, Clark, & Whitt-Woosley, 2007; Turgoose et al., 2017). Researchers conducted most of these studies among novice and veteran mental health professionals. Scant research among CITs exists. Our research attempts to fill this gap by exploring factors affecting CITs given their unique position as both students and emerging professionals. The following review of the literature supports the inclusion of predictor variables used in this study.

 

Empathy and Compassion Fatigue

One of the most widely studied concepts across various cultures is empathy, as it has been determined to be one of the major precipitants of compassion fatigue (Figley, 1995). However, findings in the literature regarding the association between compassion fatigue and empathy remain mixed (e.g., MacRitchie & Leibowitz, 2010; O’Brien & Haaga, 2015; Wagaman, Geiger, Shockley, & Segal, 2015). For instance, O’Brien and Haaga (2015) compared trait empathy and empathic accuracy with compassion fatigue after showing a videotaped trauma self-disclosure among therapist trainees (a combined group of advanced and novice graduate students) and non-therapists. The results indicated that there was no significant association between participants’ levels of compassion fatigue and empathy scores. However, MacRitchie and Leibowitz (2010) found a significant relationship between compassion fatigue and empathy after exploring the relation of these variables on trauma workers whose clients were survivors of violent crimes. The mixed results of these previous studies suggest further research is needed to understand better the relationship between empathy and compassion fatigue and how this relationship impacts counseling practice.

 

Supervisory Working Alliance and Compassion Fatigue

Although reviewed literature addressed studies suggesting supervision and support are related factors to compassion fatigue, research on this relationship is still insufficient. Kapoulitsas and Corcoran (2015) conducted a study and found that a positive supervisory relationship has a significant role in developing resilience and reducing compassion fatigue among counselors. Knight (2010) also found that students uncomfortable talking with their supervisor reported a higher risk for developing compassion fatigue. Additionally, organizational support appears to reduce compassion fatigue, whereas an absence of support increases practitioners’ and interns’ risk of developing compassion fatigue symptoms (Bride, Jones, & MacMaster, 2007). Given the intense need for support and guidance CITs need during their initial work with clients, it is expected that those students who do not actively work with their supervisors can struggle and be more vulnerable for compassion fatigue.

 

Wellness, Resilience, and Compassion Fatigue

Although counselors are encouraged to practice self-care activities to continue to enhance personal well-being (American Counseling Association [ACA], 2014; Coaston, 2017; H. L. Smith, Robinson, & Young, 2008), not all CITs can balance caring for self and others. When CITs do not receive training in the protective factors for compassion fatigue, they risk becoming more vulnerable to violating the ACA code of ethics (Merriman, 2015a; Merriman, 2015b). Kapoulitsas and Corcoran (2015) and Skovholt and Trotter-Mathison (2016) highlighted the importance of resilience and self-care activities as protective factors for compassion fatigue. Wood et al. (2017) evaluated the effectiveness of a mobile application called Provider Resilience to reduce compassion fatigue scores of mental health professionals. After a month of utilization, the results indicated that the application was effective in reducing compassion fatigue. Additionally, Lawson and Myers (2011) conducted a study with professional counselors to examine counselor wellness about compassion fatigue and found a negative correlation between total wellness scores and compassion fatigue scores. As CITs balance academic, family, and work demands, the probability of decreased wellness and a corresponding increase in compassion fatigue exists.

 

Compassion Fatigue Among CITs

 

Most CITs are often unable to master all counselor competencies (Rønnestad & Skovholt, 2013), and therefore they might not know how to deal with possible stressors and the emotional burden of their work (Star, 2013). Although they are learning counseling skills to provide the best care possible to clients, CITs may find themselves working with seriously troubled or traumatized clients without obtaining quality supervision and support (Skovholt & Trotter-Mathison, 2016). Lack of skills and resources increases the likelihood of CITs developing compassion fatigue (Merriman, 2015b). However, there is a lack of focus in compassion fatigue education on preparing CITs to manage compassion fatigue symptoms (Merriman, 2015a). Although scholars have examined compassion fatigue among counselors, there is still a dearth of studies investigating the level of compassion fatigue among CITs and addressing its protective factors within this population (Beaumont et al., 2016; Blount, Bjornsen, & Moore, 2018; Thompson et al., 2014). Subsequently, further research is needed to understand better potential protective factors that can be enhanced to offset the negative impact of compassion fatigue on CITs and the counseling process. Thus, with this study, we aimed at assessing the relationship between resilience, wellness, supervisory working alliance, empathy, and compassion fatigue among CITs in the United States. To accomplish this goal, we sought to answer the following research questions: (1) What is the prevalence of compassion fatigue among CITs? and (2) Do empathy, supervisory working alliance, resilience, and wellness significantly predict levels of compassion fatigue among CITs?

 

Method

 

Participants

Participants recruited for this study consisted of master’s-level counseling students who are at least 18 years of age and enrolled in an internship course in the United States through mostly professional listservs (e.g., Counselor Education and Supervision Network Listserv [CESNET-L], Texas Association for Counselor Education and Supervision Network Listserv [TACESNET-L], Counseling Graduate Student Network [COUNSGRADS]). Because of the impossibility of knowing how many individuals received the email invitation, we were unable to calculate and determine a response rate. Accordingly, a total of 114 CITs initially agreed to participate in this study. Before data analysis, we inspected the data set for possible entry errors and missing data. After the inspection, we excluded 28 participants from all subsequent data analyses resulting in the reduced sample of 86 CITs used to address our research questions. Overall, the sample consisted of 78 female (90.7%) and eight male (9.3%) participants, and the mean age of the participants was 32.89 years (SD = 9.72) with participants’ ages ranging between 21 and 62 years. Participants were from diverse ethnic and racial backgrounds, with the sample consisting of White (n = 48, 55.8%), Hispanic/Latino (n = 18, 20.9%), Black/African American (n = 12, 14.0%), and Asian (n = 5, 5.8%) CITs. Three participants (3.5%) listed their ethnicities as “other” when providing demographic information. Participants reported their program enrollment as follows: clinical mental health counseling program (n = 47, 54.7%); school counseling program (n = 23, 26.7%); marriage, couple, and family counseling program (n = 4, 4.7%); college counseling and student affairs program (n = 3, 3.5%); addiction counseling program (n = 2, 2.3%); and other programs (n = 7, 8.1%). Additionally, most of the participants (n = 73, 84.9%) reported enrollment in a CACREP-accredited program with the remaining participants (n = 13, 15.1%) enrolled in a non–CACREP-accredited program.

 

Procedure

Upon receiving institutional review board approval, we recruited participants from different institutions with the primary researcher contacting professional colleagues at various departments to disseminate the online survey link to potential participants during the 2017 summer and fall semesters. We also recruited participants through professional listservs (e.g., CESNET-L, TACESNET-L, COUNSGRAD), with listserv participants being provided the same informed consent and survey link through a secure website. The survey completion process took approximately 15–20 minutes.

 

Measures

We used the following self-administered survey questionnaires and a separate demographic data sheet in our data collection.

 

     Professional Quality of Life Scale (ProQOL). This scale is designed to measure the mental and emotional consequences of working with individuals who experienced trauma or painful events (Stamm, 2010). The ProQOL includes two main traits, Compassion Satisfaction (positive) and Compassion Fatigue (negative). Compassion Satisfaction is related to the joy individuals develop when they do their work well (Stamm, 2010). Compassion Fatigue consists of two subscales: Secondary Traumatic Stress (STS) and Burnout. Scholars have defined STS as an emotional state that occurs when an individual becomes upset or traumatized as a result of their exposure to victim experiences (Figley, 2002). The second part of Compassion Fatigue is Burnout, which is a multidimensional syndrome related to the social work environment. Burnout could be related to work overload, lack of control, insufficient rewards, unfairness, and value conflict at a workplace (Skovholt & Trotter-Mathison, 2016).

 

The ProQOL is a 30-item Likert-type self-report assessment with responses of never, rarely, sometimes, often, and very often for each item. A sample item is “I feel depressed because of the traumatic experiences of the people I [help].” This assessment has 10 questions per each of three main scales measuring separate constructs. However, the Compassion Fatigue scale includes two of these constructs, which are the Burnout and the STS scales. According to Stamm (2010), the ProQOL has good construct validity, as researchers have noted its efficacy in over 200 published articles. Finally, alpha coefficient values for the Burnout and STS scales were .75 and .81, respectively (Stamm, 2010), and are similar (.72 and .79) to the Cronbach’s alpha values from the current study presented in Table 1.

 

Table 1

Descriptive Statistics of the Study Variables (N = 86)

Range
Variable M SD Min Max Skew α
Compassion Fatigue 41.48 8.03 22 60    .19
BO 21.34 4.38 12 32 – .72
STS 20.14 4.96 10 38 – .79
Empathy 21.86 4.12   9 28   -.51 .80
Supervisory Working Alliance   5.82   .97   2.16   7  1.26
CF   6.65 1.30   2.17   8.17 – .90
R   5.80   .96   2.33   7 – .93
Resilience   3.43   .79   1   4.67   -.74 .89
Wellness 47.58 6.23 27 56 -1.39 .86


Note. BO = Burnout; STS = Secondary Traumatic Stress; CF = Client Focus; R = Rapport

 

 

     Interpersonal Reactivity Index (IRI). Davis (1983) developed the IRI to measure the reactions of a person to other individuals’ observed experiences. The 28-item instrument has four subscales: Empathic Concern, Perspective Taking, Fantasy, and Personal Distress (Davis, 1983). Researchers report separate subscale scores, as a total score for the instrument has not been recommended (Davis, 1983). In this study, we only used the Empathic Concern subscale to collect data regarding empathy scores of CITs.

 

Davis (1983) described empathic concern as an emotional response, such as compassion and sympathy, to someone else in need. The 7-item subscale is a self-report assessment with a 5-point Likert-type scale, ranging from Does not describe me well to Describes me very well. A sample item is “I am often quite touched by things that I see happen.” An alpha coefficient of .77 has been reported for the Empathic Concern subscale (Péloquin & Lafontaine, 2010), while the Cronbach’s alpha value of the IRI in the current study was .80.

 

     Supervisory Working Alliance Inventory: Trainee Form (SWAI-T). Efstation, Patton, and Kardash (1990) developed this inventory to measure supervisees’ perceptions about the effectiveness of the working relationship with their supervisors, and we used the SWAI-T to measure the construct of the supervisory working alliance. With a total of 19 items, the self-report assessment includes a 7-point Likert-type scale with responses ranging from almost never to almost always. A sample item is “When correcting my errors with the client, my supervisor offers alternative ways of intervening with the client.” The SWAI-T has two subscales—Client Focus and Rapport—and the Cronbach alpha coefficients of these scales were .77 and .90, respectively (Efstation et al., 1990). For the current study, we calculated Cronbach alpha values of .90 for the Client Focus subscale and .93 for the Rapport subscale. Because some researchers have found high correlations between these two subscales, they decided to combine them in their studies (e.g., Ganske, 2007; White & Queener, 2003). Therefore, in this study, after conducting a correlation analysis with the subscale scores, we also chose to combine subscales as the results of subscale scores were highly correlated.

 

     Brief Resilience Scale (BRS). The BRS was developed to measure a person’s ability to recover from stress and cope with challenging situations (B. W. Smith et al., 2008). The BRS is used to measure the construct of resilience. As a 6-item self-report assessment, the BRS includes a 5-point Likert-type scale with responses ranging from strongly disagree to strongly agree. A sample item is “I usually come through difficult times with little trouble.” B. W. Smith and colleagues (2008) reported that the Cronbach’s alpha values of the BRS range from .80 to .91, and we calculated a Cronbach alpha of .89 for the current study.

 

     Flourishing Scale (FS). The FS was designed to measure individuals’ self-perceived success in areas like optimism and relationships (Diener et al., 2010) and used to measure the construct of wellness in this study. The FS is an 8-item self-report assessment with a 7-point Likert-type scale with responses ranging from strongly disagree to strongly agree (Diener et al., 2010). A sample item is “I lead a purposeful and meaningful life.” Diener and colleagues (2010) reported moderately high reliability with a .87 Cronbach’s alpha coefficient, and in the current study, the FS had a Cronbach alpha of .86.

 

Data Analysis

     Statistical power analysis. We used an a priori type of the G*Power to set the minimum number of participants needed to detect statistical power for this research design. Based on an alpha of .05, a power level of .90, and four predictors (Faul, Erdfelder, Buchner, & Lang, 2009), the computation results suggested that a minimum of 73 participants was required to detect statistical significance with at least a moderate size effect (.15). We had 86 participants, suggesting adequate power.

 

     Preliminary analyses. We analyzed all data using the Statistical Package for the Social Sciences, Version 20 (SPSS; IBM Corporation, 2011). Before addressing our stated research questions, we cleaned the dataset and addressed missing data. We did not observe any pattern between missing data points. Therefore, the type of missing data was completely random, which was addressed using the series of mean function within the SPSS. Next, we calculated descriptive statistics and alpha coefficients for each scale used in the study (see Table 1). Before performing hierarchical regression analyses, we tested all associated model assumptions. First, we examined study variables based on their types and concluded each utilized a continuous scale. We then assessed normality with the Shapiro-Wilk test of normality (W > .05), indicating data was normally distributed for the dependent variable. To identify outliers, we examined boxplots. Although there were a few mild outliers, no extreme scores were detected. We assessed linearity and homoscedasticity through inspection of standardized residual plots. To assess for the assumption of multicollinearity, we examined the correlation matrix of study variables to determine if any correlated highly. According to Field (2013), correlations above .80 are considered high and may indicate the presence of multicollinearity. In the present study, none of the correlation coefficients were above .50 (see Table 2). Collectively, these findings indicated no evidence suggesting any of the model assumptions had been violated. As a result, the dataset was deemed appropriate for analysis using a hierarchical regression design.

 

     Primary analysis. Descriptive statistics were calculated to organize the data by producing means, mode, median, standard deviations, and minimum and maximum scores for the study variables (Field, 2013). Individually, we reviewed descriptive statistics for the compassion fatigue variable, and results were reported to address the first research question. Next, we performed a three-step hierarchical linear regression to address the second research question.

 

Table 2

 

Intercorrelations for Scores on the Study Variables

Variable 1 2 3 4 5
1. ProQOL-CF –
2. SWAIT-T   .04 –
3. IRI-EC  -.06  .04 –
4. BRS    -.47** -.09 -.11 –
5. FS    -.45**  .12    .25* .35** –


Note. N = 86; ProQOL = Professional Quality of Life (Compassion Fatigue [CF] subscale score is presented); IRI = Interpersonal Reactivity Index (Empathic Concern [EC] subscale score is presented); SWAI-T = Supervisory Working Alliance Inventory: Trainee Form; BRS = Brief Resilience Scale; FS = Flourishing Scale.

*p < .05.  **p < .01.

 

 

 

Results

 

Compassion fatigue scores of CITs represent the sum of scores of all items on the STS and Burnout subscales. According to the ProQOL administration manual (Stamm, 2010), individuals scoring below 22 may indicate little or no issues with Burnout and STS, while scores between 23 and 41 indicate moderate levels of Burnout and STS, and scores above 42 indicate higher levels of Burnout and STS. For this sample, participants’ Burnout scores ranged from 12 to 32 with a mean of 21.34 (SD = 4.38), and STS scores ranged from 10 to 38 with a mean of 20.14 (SD = 4.96). These results indicated a low risk of both Burnout and STS among CITs.

 

To address the second research question, we performed a three-step hierarchical linear regression analysis. With this analysis, we aimed to assess the association between resilience, wellness, supervisory working alliance, empathy, and compassion fatigue. We chose to implement a hierarchical multiple regression analysis because scholars previously have highlighted the essential relationship between empathy, supervision, and compassion fatigue (Figley, 2002; MacRitchie & Leibowitz, 2010). In the first step, empathy scores entered the model as a predictor variable, as Figley (1995) stated that empathy is one of the main factors contributing to compassion fatigue. However, among this sample, we found that empathy was not a significant predictor of compassion fatigue: F(1, 84) = .2, p = .66 , R2 = .002 (adjusted R2 = -.01). Then, we added supervisory working alliance scores to the model in the second step, as both Knight (2010) and Miller and Sprang (2017) emphasized the importance of supervisory support for mental health practitioners. Results revealed that the supervisory working alliance variable also was not a significant predictor of compassion fatigue: F(2, 83) = .16, p = .85, R2 = .004 (adjusted R2 = -.02). In the third step, resilience and wellness scores were entered into the model to determine whether these variables significantly improved the amount of explained variance in compassion fatigue. Results showed that this combination of variables significantly predicted 26% of the variance in compassion fatigue: F(4, 81) = 8.57, p < .001, R2 = .30. Therefore, it was concluded that CITs with greater wellness and resilience reported developing less compassion fatigue (see Table 3).

 

Table 3

Hierarchical Regression Analysis Results for Variables Predicting Compassion Fatigue

Variables B SEB β R2 ΔR2
Step 1 .002 -.01
Empathy   -.09 .21 -.05
Step 2 .004 -.02
Empathy   -.10 .21 -.05
SWA    .33 .91  .04
Step 3 .30*  .26
Empathy   -.03 .19 -.02
SWA    .36 .78  .04
Wellness   -.39 .14  -.30*
Resilience  -3.66     1.05  -.36*


Note. SWA = Supervisory Working Alliance

*p < .05.

 

 

Discussion

 

In this study, CITs reported having a low risk of compassion fatigue. When we examined the Burnout and STS scores separately, the main contributors of compassion fatigue (Stamm, 2010), both subscale scores indicated participants having a low risk for STS and Burnout. This finding is similar to results found by Beaumont and colleagues (2016) in their study of compassion fatigue, burnout, self-compassion, and well-being relationships among student counselors and student cognitive behavioral psychotherapists. According to their research findings, a total of 54 student participants reported high scores on self-compassion and well-being and reported less compassion fatigue and burnout (Beaumont et al., 2016).

 

One of the goals of this study was to seek understanding of whether wellness and resilience explain a statistically significant amount of variance in compassion fatigue among CITs after accounting for empathy and supervisory working alliance. The results indicated that empathy and supervisory working alliance were not significant predictors of compassion fatigue. Regarding empathy and compassion fatigue relation results, the findings of this study did not support Figley’s (1995) assumption of empathy as one of the main contributors to compassion fatigue. This result also is inconsistent with Wagaman and colleagues’ (2015) results indicating a significant association between empathy and compassion fatigue among social workers. However, current results aligned with those studies that found no correlation between empathy and compassion fatigue (e.g., O’Brien & Haaga, 2015; Thomas & Otis, 2010). An explanation of the variability between this inquiry and previous studies might lie with the difference between participants’ field of study and measurement differences. Also, none of the previous studies used CITs solely as their sample, nor used a similar way to measure the construct of empathy. Additionally, CITs would have less experience working with clients compared to experienced counselors, and thus less time for feelings of compassion fatigue to build.

 

Although scholars addressed the importance of supervision and supervisory working alliance to help prevent compassion fatigue (Kapoulitsas & Corcoran, 2015; Merriman, 2015a), this study’s results indicated supervisory working alliance was not a significant predictor of compassion fatigue among CITs. Like current results, Ivicic and Motta (2017) and Williams, Helm, and Clemens (2012) found no statistically significant association between supervisory working alliance and compassion fatigue among mental health practitioners. It is noteworthy that these studies highlighting the importance of supervision and the supervisory relationship are qualitative in design, and participants did not consist solely of CITs. Additionally, their results emphasized the importance of supervision as support to counter the negative impact of trauma exposure (Kapoulitsas & Corcoran, 2015; Ling, Hunter, & Maple, 2014). According to the current study results, CITs did not report experiencing a high level of compassion fatigue. This finding could be interpreted as CITs not yet feeling the need for supervisory support to help with compassion fatigue.

 

Results also indicated that resilience and wellness were significant predictors of compassion fatigue among CITs. In other words, when reflecting on both the regression and correlation results, CITs with greater resilience and wellness reported lower scores of compassion fatigue and these results were consistent with Tosone, Minami, Bettmann, and Jasperson’s (2010) research findings. Regarding a wellness and compassion fatigue relationship, Beaumont and colleagues (2016) conducted a study with student counselors and student cognitive behavioral psychotherapists. The results of Beaumont et al.’s study revealed that individuals with high scores of self-compassion and well-being reported having less compassion fatigue and burnout. Thomas and Morris (2017) also highlighted the significance of self-care and well-being not only for preventing and helping to manage the potentially damaging impact of practice, but also for facilitating the counselor’s personal and professional growth.

 

Implications for Counselor Educators and Supervisors

 

The research findings provide data-driven results regarding compassion fatigue among CITs that have meaningful implications for counselor educators and supervisors. Present study results revealed that CITs indicated experiencing a low risk of compassion fatigue. However, raising awareness on this issue may still help CITs as a preventative measure to cope with possible compassion fatigue experience in the future. To address this issue, counselor educators may consider raising awareness on this topic by reviewing current counseling program curricula to add discussion questions related to compassion fatigue and its empirically predictive factors—wellness and resilience. Roach and Young (2007) stated that students in counseling programs reported group counseling, counseling techniques, legal and ethical issues, practicum, and wellness courses as contributing most to their knowledge and skills regarding wellness. Therefore, counselor educators might use different assignments, including group discussions, projects, and role-playing exercises, to open a discussion about the compassion fatigue phenomenon and the relation with its predictive factors and these courses. Counselor educators may also use the ProQOL scale as an assignment in an assessment and testing course to inform CITs about how to use this instrument as a self-monitoring aid. For example, professional counselors may feel overwhelmed because of working with trauma survivors after graduation and start noticing compassion fatigue symptoms in themselves. These individuals may self-administer the ProQOL scale to determine whether they have developed compassion fatigue. Additionally, in a practicum or an internship course, CITs may fill out the ProQOL as part of their continuing personal wellness plan by comparing personal results over time and sharing their thoughts and reflections about the results.

 

Supervisors need to find ways to raise awareness of compassion fatigue and its protective factors with CITs. For instance, during internship experience, supervisors may develop a site training including compassion fatigue awareness for CITs, as CITs should be prepared for the possible emotional and psychological consequences in working with trauma survivors. Student counselors also should be encouraged to advocate for themselves when they notice symptoms of compassion fatigue. Supervisors might consider the administration of the ProQOL scale regularly to assess both organizational and individual risks (Newell & MacNeil, 2010). Additionally, supervisors can use the ProQOL scale with their supervisees to start a conversation about compassion fatigue. Although the ProQOL is not a diagnostic test, the 30-item self-report scale can be utilized readily as a conversation starter in supervision sessions.

 

The results suggested that empathy and supervisory working alliance did not predict CITs’ compassion fatigue level. However, wellness and resilience are significantly related to contributing to it. Therefore, both counselor educators and supervisors might consider enhancing CITs’ resilience and wellness a worthwhile endeavor. For example, Miller and Sprang (2017) developed a component-based practice and supervision model to reduce compassion fatigue for use in training, supervision, and clinical practice. A tool like this one can be added to existing training curricula and supervision practice to improve CITs’ resilience and wellness.

 

Limitations

The results of this study aim to provide greater clarity regarding the predictive factors of compassion fatigue among CITs. However, interpretation of results should take into consideration the limitations that emerged because of uncontrollable influences and choices we made. The study was limited in its ability to represent all CITs throughout the United States, as we utilized a convenience sampling approach. Additionally, we gathered data through self-report questionnaires, which introduce the possibility of response bias in the findings. Although we assumed participants answered each question honestly, they might not have been honest in their responses because of the fear of being perceived as weak or less competent. It is important to note that being in an internship class might also increase participants’ interest in the profession as they currently are engaged in the practice of counseling. Therefore, participants might have had a higher level of enthusiasm and reported less compassion fatigue. Also, individuals who suffer from compassion fatigue might have preferred not to respond to these items. Finally, although participants were enrolled in an internship class, each participant may have different numbers of hours of client experience.

 

Future Directions for Research

Additional research should be conducted to expand and clarify the current research findings of compassion fatigue among CITs. A phenomenological study using a qualitative research method is recommended to expand the findings of this current study. Future researchers may use the ProQOL scale to assess CITs’ level of compassion fatigue and then conduct interviews with the volunteer participants reporting a higher level of compassion fatigue to better understand CITs’ experience with compassion fatigue and its contributing factors. The data collected through a qualitative study may provide greater insight into the phenomenon of compassion fatigue among CITs. Additionally, researchers can replicate the present study with early-career counselors who have recently graduated, because of the noted intensity of those first years after graduation (Skovholt & Trotter-Mathison, 2016). Therefore, future researchers exploring novice counselors’ experiences with compassion fatigue will help counselor educators and supervisors better understand when counselors may start developing compassion fatigue symptoms, as well as how they cope with the symptoms.

 

Conclusion

 

CITs may struggle when they continuously hear painful stories of clients because of a lack of experience, skillset, or support (Skovholt & Trotter-Mathison, 2016). Researchers have described this experience as compassion fatigue. With this study, we aimed to provide a better understanding of the predictive factors of compassion fatigue among CITs. Using data-driven research results to determine ways to work with CITs on compassion fatigue and its predictive factors can be beneficial in preventing compassion fatigue symptoms from an early onset. CITs may take precautionary measures to ensure they remain enthusiastic and energized by the work they do. Further, implications of the current study may help CITs start their professional careers better prepared to provide their clients with the optimal care needed throughout the counseling relationship by minimizing compassion fatigue.

 

 

Conflict of Interest and Funding Disclosure

The authors reported no conflict of interest

or funding contributions for the development

of this manuscript.

 

 

 

References

 

American Counseling Association. (2014). ACA code of ethics. Alexandria, VA: Author.

Beaumont, E., Durkin, M., Martin, C. J. H., & Carson, J. (2016). Measuring relationships between self-compassion, compassion fatigue, burnout and well-being in student counsellors and student cognitive behavioural psychotherapists: A quantitative survey. Counselling & Psychotherapy Research, 16, 15–23. doi:10.1002/capr.12054

Blount, A. J., Bjornsen, A. L., & Moore, M. M. (2018). Work values, occupational engagement, and professional quality of life in counselors-in-training: Assessments in a constructivist-based career counseling course. The Professional Counselor, 8, 60–72. doi:10.15241/ajb.8.1.60

Bride, B. E., Jones, J. L., & MacMaster, S.A. (2007). Correlates of secondary traumatic stress in child protective services workers. Journal of Evidence-Based Social Work, 4(3/4), 69–80. doi:10.1300/J394v04n03_05

Caringi, J. C., Hardiman, E. R., Weldon, P., Fletcher, S., Devlin, M., & Stanick, C. (2017). Secondary traumatic stress and licensed clinical social workers. Traumatology, 23(2), 186–195. doi:10.1037/trm0000061

Coaston, S. C. (2017). Self-care through self-compassion: A balm for burnout. The Professional Counselor, 7, 285–297. doi:10.15241/scc.7.3.285

Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113–126. doi:10.1037/0022-3514.44.1.113

Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D., Oishi, S., & Biswas-Diener, R. (2010). New measures of well-being: Flourishing and positive and negative feelings. Social Indicators Research, 39, 247–266. doi:10.1007/978-90-481-2354-4_12

Eastwood, C. D., & Ecklund, K. (2008). Compassion fatigue risk and self-care practices among residential treatment center childcare workers. Residential Treatment for Children & Youth, 25, 103–122. doi:10.1080/08865710802309972

Efstation, J. F., Patton, M. J., & Kardash, C. M. (1990). Measuring the working alliance in counselor supervision. Journal of Counseling Psychology, 37, 322–329. doi:10.1037/0022-0167.37.3.322

Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149–1160. doi:10.3758/brm.41.4.1149

Field, A. (2013). Discovering statistics using IBM SPSS statistics (4th ed.). Thousand Oaks, CA: SAGE.

Figley, C. R. (Ed.). (1995). Compassion fatigue: Coping with secondary traumatic stress disorder in those who treat the traumatized. New York, NY: Routledge.

Figley, C. R. (Ed.). (2002). Treating compassion fatigue. New York, NY: Brunner-Routledge.

Ganske, K. H. (2007). The relationship between counselor trainee perfectionism and working alliance with supervisor and client. (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses database. (UMI No. 3377933)

Harr, C. (2013). Promoting workplace health by diminishing the negative impact of compassion fatigue and increasing compassion satisfaction. Social Work & Christianity, 40, 71–88.

IBM Corporation. (2011). IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: Author.

Ivicic, R., & Motta, R. (2017). Variables associated with secondary traumatic stress among mental health professionals. Traumatology, 23, 196–204. doi:10.1037/trm0000065

Kapoulitsas, M., & Corcoran, T. (2015). Compassion fatigue and resilience: A qualitative analysis of social work practice. Qualitative Social Work: Research and Practice, 14, 86–101. doi:10.1177/1473325014528526

Knight, C. (2010). Indirect trauma in the field practicum: Secondary traumatic stress, vicarious trauma, and compassion fatigue among social work students and their field instructors. The Journal of Baccalaureate Social Work, 15, 31–52.

Lawson, G., & Myers, J. E. (2011). Wellness, professional quality of life, and career-sustaining behaviors: What keeps us well? Journal of Counseling & Development, 89, 163–171. doi:10.1002/j.1556-6678.2011.tb00074.x

Lawson, G., Venart, E., Hazler, R. J., & Kottler, J. A. (2007). Toward a culture of counselor wellness. The Journal of Humanistic Counseling, Education and Development, 46, 5–19. doi:10.1002/j.2161-1939.2007.tb00022.x

Ling, J., Hunter, S. V., & Maple, M. (2014). Navigating the challenges of trauma counselling: How counsellors thrive and sustain their engagement. Australian Social Work, 67, 297–310.
doi:10.1080/0312407X.2013.837188

MacRitchie, V., & Leibowitz, S. (2010). Secondary traumatic stress, level of exposure, empathy and social
support in trauma workers. South African Journal of Psychology, 40, 149–158. doi:10.1177/008124631004000204

Merriman, J. (2015a). Enhancing counselor supervision through compassion fatigue education. Journal of Counseling & Development, 93, 370–378. doi:10.1002/jcad.12035

Merriman, J. (2015b). Prevention-based training for licensed professional counselor interns. Journal of Professional Counseling: Practice, Theory & Research, 42, 40–53.

Miller, B., & Sprang, G. (2017). A components-based practice and supervision model for reducing compassion fatigue by affecting clinician experience. Traumatology, 23, 153–164. doi:10.1037/trm0000058

Newell, J. M., & MacNeil, G. A. (2010). Professional burnout, vicarious trauma, secondary traumatic stress, and compassion fatigue: A review of theoretical terms, risk factors, and preventive methods for clinicians and researchers. Best Practices in Mental Health: An International Journal, 6, 57–68.

O’Brien, J. L., & Haaga, D. A. F. (2015). Empathic accuracy and compassion fatigue among therapist trainees. Professional Psychology: Research and Practice, 46, 414–420. doi:10.1037/pro0000037

Péloquin, K., & Lafontaine, M. (2010). Measuring empathy in couples: Validity and reliability of the
Interpersonal Reactivity Index for couples. Journal of Personality Assessment, 92, 146–157. doi:10.1080/00223890903510399

Ray, S. L., Wong, C., White, D., & Heaslip, K. (2013). Compassion satisfaction, compassion fatigue, work life conditions, and burnout among frontline mental health care professionals. Traumatology, 19, 255–267. doi:10.1177/1534765612471144

Roach, L. F., & Young, M. E. (2007). Do counselor education programs promote wellness in their students? Counselor Education & Supervision, 47, 29–45. doi:10.1002/j.1556-6978.2007.tb00036.x

Rønnestad, M. H., & Skovholt, T. M. (2013). The developing practitioner: Growth and stagnation of therapists and counselors. New York, NY: Routledge.

Skovholt, T. M., & Trotter-Mathison, M. (2016). The resilient practitioner: Burnout and compassion fatigue prevention and self-care strategies for the helping professions (3rd ed.). New York, NY: Routledge.

Smith, B. W., Dalen, J., Wiggins, K., Tooley, E., Christopher, P., & Bernard, J. (2008). The Brief Resilience Scale: Assessing the ability to bounce back. International Journal of Behavioral Medicine, 15, 194–200. doi:10.1080/10705500802222972

Smith, H. L., Robinson, E. H. M., III, & Young, M. E. (2008). The relationship among wellness, psychological distress, and social desirability of entering master’s-level counselor trainees. Counselor Education & Supervision, 47, 96–109. doi:10.1002/j.1556-6978.2007.tb00041.x

Sprang, G., Clark, J. J., & Whitt-Woosley, A. (2007). Compassion fatigue, compassion satisfaction, and burnout: Factors impacting a professional’s quality of life. Journal of Loss and Trauma, 12, 259–280. doi:10.1080/15325020701238093

Stamm, B. H. (2010). The concise ProQOL manual (2nd ed.). Pocatello, ID: ProQOL.org.

Star, K. L. (2013). The relationship between self-care practices, burnout, compassion fatigue, and compassion satisfaction among professional counselors and counselors-in-training. (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses database. (UMI No. 3618924)

Thomas, D. A., & Morris, M. H. (2017). Creative counselor self-care. In Ideas and research you can use: VISTAS 2017. Retrieved from https://www.counseling.org/docs/default-source/vistas/creative-counselor-self-care.pdf?sfvrsn=4

Thomas, J. T., & Otis, M. D. (2010). Intrapsychic correlates of professional quality of life: Mindfulness, empathy, and emotional separation. Journal of the Society for Social Work and Research, 1(2), 83–98. doi:10.5243/jsswr.2010.7

Thompson, I. A., Amatea, E. S., & Thompson, E. S. (2014). Personal and contextual predictors of mental health counselors’ compassion fatigue and burnout. Journal of Mental Health Counseling, 36, 58–77. doi:10.17744/mehc.36.1.p61m73373m4617r3

Tosone, C., Bettmann, J. E., Minami, T., & Jasperson, R. A. (2010). New York City social workers after 9/11: Their attachment, resiliency, and compassion fatigue. International Journal of Emergency Mental Health, 12(2), 103–116.

Turgoose, D., Glover, N., Barker, C., & Maddox, L. (2017). Empathy, compassion fatigue, and burnout in police officers working with rape victims. Traumatology, 23, 205–213. doi:10.1037/trm0000118

Wagaman, M. A., Geiger, J. M., Shockley, C., & Segal, E. A. (2015). The role of empathy in burnout, compassion satisfaction, and secondary traumatic stress among social workers. Social Work, 60, 201–209.
doi:10.1093/sw/swv014

White, V. E., & Queener, J. (2003). Supervisor and supervisee attachments and social provisions related to the
supervisory working alliance. Counselor Education and Supervision, 42, 203–218.
doi:10.1002/j.1556-6978.2003.tb01812.x

Williams, A., Helm, H., & Clemens, E. V. (2012). The effect of childhood trauma, personal wellness, supervisory working alliance, and organizational factors on vicarious traumatization. Journal of Mental Health Counseling, 34, 133–153. doi:10.17744/mehc.34.2.j3l62k872325h583

Wood, A. E., Prins, A., Bush, N. E., Hsia, J. F., Bourn, L. E., Earley, M. D., . . . Ruzek, J. (2017). Reduction of
burnout in mental health care providers using the Provider Resilience mobile application. Community
Mental Health Journal, 53, 452–459. doi:10.1007/s10597-016-0076-5

 

Nesime Can is an instructor at Ankara University in Turkey. Joshua C. Watson, NCC, is a professor and department chair at Texas A&M University–Corpus Christi. Correspondence can be addressed to Nesime Can, Ankara University Faculty of Educational Sciences, Department of Educational Sciences, Program of Counseling and Guidance, Office 3111, Çankaya, Ankara, Turkey 06590, nesime.can@ankara.edu.tr.