Perceived Impact on Client Outcomes

Clinical supervision is considered to be an essential component of psychotherapy training. However, research on supervisors’ ability to affect client outcome has been mixed. This investigation aims to answer two questions: (a) What is the perceived impact of supervision on client outcome, and (b) how important is it to supervisors and supervisees that supervision affects client outcome? A mixed-methods approach was used to examine the perspectives of both supervisors and supervisees. The survey consisted of survey questions and one open-ended qualitative question. Findings suggest that both supervisors and supervisees perceive supervision as beneficial and important for impacting client outcome. Supervisees perceive supervision as more important in impacting outcome than supervisors. Existing literature suggests that supervision may not have as significant an impact on client outcome as previously believed; however, supervisors and supervisees perceive the supervision process as essential for psychotherapy training. Implications and future directions are discussed.

2009; Reese et al., 2009). Each of these studies yielded significantly different results, and despite stronger methodology than the reviews cited above, the extent to which supervision impacts client outcome is unclear. Bambling et al. (2006) used an experimental design, in which 127 clients were randomly assigned to 127 licensed therapists, half of whom received supervision for eight sessions of problemsolving treatment. Clients of therapists receiving supervision had higher scores on both measures of treatment outcome and working alliance (Bambling et al., 2006). Callahan and colleagues (2009) used a naturalistic data set to examine the outcomes of 76 adult psychotherapy clients who were randomly assigned to 40 trainee therapists, who were in supervision with nine supervisors. Supervisors were found to account for 16% of the variance in client outcome (Callahan et al., 2009). Reese et al. (2009) performed a controlled study in which the outcomes of trainees receiving supervision that included regular outcome feedback (n = 9) were compared with the outcomes of trainees receiving supervision without regular outcome feedback (n = 10). Data included 115 psychotherapy cases collected over the course of a year. Trainees in the supervision-with-feedback condition had significantly better outcomes than trainees receiving supervision without feedback. No significant differences were found between supervisors within the treatment conditions (Reese et al., 2009).
On the contrary, a recent study examined this topic with a large sample from routine practice and found null results (Author & Coauthors, 2014). In this study, a 5-year archival naturalistic data set of psychotherapy outcomes from a private, nonprofit mental health center was used to test whether client treatment outcomes (as measured by the Outcome Questionnaire-45.2 [OQ-45.2]) differed depending on who was providing the supervision (Lambert et al., 2004).The data included 6,688 clients seen by 174 trainee therapists who were supervised by 14 supervisors. Results indicated that supervisors explained less than .01% of the variance in client psychotherapy outcomes (Author, 2014). Thus, the extent to which supervisors can affect client outcome is still unclear. In light of these results, the question still remains: To what extent does supervision impact client outcome? In his summary of the research on the subject, Watkins (2011) concluded, "After a century of psychotherapy supervision and over half a century of supervision research, we still cannot empirically answer that question" (p. 252).
Given this research, it is unrealistic to state that the primary task of clinical supervision is to ensure the greatest possible treatment outcome for clients. Hence, there is a clear need to determine whether supervisors and supervisees even believe the supervision process is impactful on client outcome. If the existing literature suggests that supervision has little impact on client outcome, it would be important to determine whether supervisors and supervisees share this perception.
Our aim for this study, which has yet to be investigated, was to use a mixed-methods design to better understand supervisor and supervisee perceptions of the supervisionclient outcome relationship.
The first major goal of this investigation was to examine supervisors' and supervisees' opinions of two questions: (a) What is the extent of impact supervisors can have on client outcomes, and (b) how important is it that supervisors can affect client outcome? Drawing from treatment-outcome literature (e.g., Lambert, 2013), these two questions explored four major areas of client outcome: positive impact of treatment, client deterioration, client dropout, and speed/efficiency of treatment. In addition, we sought to examine the relation between each participant's answers to the two research questions: Research Question 1: What is the perceived impact of supervision on client outcome? Research Question 2: How important is it to supervisors and supervisees that supervision affects client outcome?
The second major goal of this study was to explore supervisors' and supervisees' perspectives on how supervisors affect client outcome. In other words, what specific activities of supervision enable the supervisor to impact client outcome? It is important to note that it is still unclear how supervision may actually impact client outcome. Given the exploratory nature of this investigation, the authors did not make any hypotheses about how supervisors and supervisees would respond to the quantitative items or their open-ended qualitative responses to this topic.
To explore whether any outstanding demographic differences existed, a series of independent-sample t tests and oneway ANOVAs were used to explore the demographic differences within supervisors' extent and importance on impact scores. These tests explored both overall differences between the constructs of extent and importance of supervision, as well as examined single-item variations for both supervisors and supervisees. Within supervisees, a series of independent-sample t tests and one-way ANOVAs were used to explore the demographic differences on extent and importance of impact scores. All analyses were conducted through SPSS.

Measures
Supervisors. Supervisors completed an online survey with a total of 15 questions. The questions on the supervisor survey were developed by the authors who are practicing supervisors and supervision researchers. The questions were then pilot-tested with six practicing supervisors and revised for clarity. The first four questions in the survey were presented as a group, with the following header: "For the specific areas below, rate the positive impact your supervision has on your supervisees' clients' outcomes." The four areas were as follows: (a) "My supervision increases the positive impact of my trainees' counseling/psychotherapy," (b) "My supervision prevents or reduces my trainees' client deterioration," (c) "My supervision prevents or reduces my trainees' client dropout," and (d) "My supervision increases the speed or efficiency of my trainees' counseling/psychotherapy." These questions were answered along a 5-point Likert-type scale, from 1 = "do not agree" to 5 = "totally agree," and a sum score was computed.
The next four questions were presented as another group, with the following header: "When you consider the diverse range of skills that someone must have to be an effective supervisor, how important is the supervisor's overall ability to impact their supervisees' clients' outcomes?" The four areas were as follows: (a) "Increase the positive impact of counseling/psychotherapy," (b) "Prevent or reduce client deterioration," (c) "Prevent or reduce client dropout," and (d) "Increase the speed or efficiency of counseling/psychotherapy." These questions were answered also along a 5-point Likert-type scale, from 1 = "no importance" to 5 = "highest importance," and a sum score was computed. Cronbach's alpha for these items in this sample was .76, indicating fair internal consistency.
Questions 9 and 10 asked whether supervisor participants have their supervisees use quantitative measures to track their client outcome, and, if so, which measure(s). For the qualitative portion of this study, Question 11 asked, "Other than what has been identified above, please list any other ways that you feel your supervision has a positive impact on your supervisee's client's outcomes." This question allowed for any length of answer. The remaining four items were demographic questions (age, gender, primary therapeutic orientation, and years of experience as a supervisor).
Supervisees. Supervisees completed an online survey with a total of 18 questions. The questions on the supervisee survey were developed by the authors who are supervisors, pilottested with six practicing supervisees, and revised for clarity. The first four questions in the survey were presented as a group, with the following header: "In your opinion, how much does your current primary individual supervisor help you to help your clients?" The four areas were as follows: (a) "My supervisor helps me increase the positive impact of my counseling/psychotherapy," (b) "My supervisor helps me prevent or reduce the risk of my clients deteriorating," (c) "My supervisor helps me prevent or reduce the risk of my clients dropping out," and (d) "My supervisor helps me increase the speed or efficiency of my counseling/psychotherapy." These questions were answered along a 5-point Likert-type scale, from 1 = "do not agree" to 5 = "totally agree," and a sum score was computed. Cronbach's alpha for these items in this sample was .87, indicating good internal consistency.
The next four questions were presented as another group, with the following header: "In your opinion, how important is it that your supervisor is able to help you help your clients?" The four areas were as follows: (a) "How important is it that your supervisor can help you increase the positive impact of your counseling/psychotherapy," (b) "How important is it that your supervisor can help you prevent or reduce the risk of client deterioration," (c) "How important is it that your supervisor can help you prevent or reduce the risk of client dropout," and (d) "How important is it that your supervisor can help you increase the speed or efficiency of your counseling/ psychotherapy." These questions were answered along a 5-point Likert-type scale, from 1 = "no importance" to 5 = "highest importance," and a sum score was computed.
Questions 9 and 10 asked whether supervisee participants have their supervisees use quantitative measures to track their client outcome, and, if so, which measure(s). Similar to supervisor inquiry, Question 11 asked, "Other than what has been identified above, please list any other ways that you feel your supervisor has a positive impact on your clients' outcomes." This question allowed any length of answer. The remaining seven items were demographic questions (age, gender, primary therapeutic orientation, supervisor's primary therapeutic orientation, months of experience working with clients in face-to-face individual counseling/psychotherapy, total number of clients worked with in face-to-face individual counseling/psychotherapy, and type of degree program).

Qualitative Methodology
A discovery-oriented-exploratory qualitative research methodology (Mahrer, 1988) was used to analyze supervisors' responses to the question, "Other than what has been identified above, please list any other ways that you feel your supervisor has a positive impact on your clients' outcomes," and supervisees' responses to the question, "Other than what has been identified above, please list any other ways that you feel your supervision has a positive impact on your supervisee's client's outcomes." Analysis was conducted by three of the authors, who attempted to follow the guidelines of discovery-oriented methodology (e.g., Yeh & Inman, 2007). Before starting data analysis, the authors discussed their own expectations, values, and biases toward the research material, including assumptions based on their self-identities, with an eye for the power differentials inherent in the roles of supervisor and supervisees.
Each judge reviewed all of the supervisors' responses, creating and sorting them into mutually exclusive categories. The judges then met as a team to discuss their different categories and create a codebook. The team coded all of the supervisors' responses, and disagreements were discussed and resolved by consensus. This process was repeated for the supervisees' responses. The three judges came to consensus on the supervisor and supervisee codes. The team then reviewed and discussed both codebooks, modifying and recoding the associated responses until there was 100% consensus. Codes that represented fewer than 1.00% of participant responses were dropped from the codebook.

Procedures
Participants for the supervisor and supervisee surveys were recruited through purposive sampling using a wide range of email listservs, including graduate programs, regional associations, and national associations, and for specific therapeutic modalities (e.g., cognitive-behavioral therapy). As an incentive to participate, an opportunity to enter a drawing for one of two US$50 Amazon.com gift certificates was offered. Participation was voluntary and anonymous. The secure online survey was hosted at http://www.psychdata.com. Both surveys were available online for 2 weeks, during which time 194 supervisors and 196 supervisees completed them. Participants with missing data or with z scores ≥ |3.5| were removed. Five supervisors and five supervisees were removed, leaving data that included 189 supervisors and 191 supervisees.

Results
The major goal of this investigation sought to explore the perspectives of both supervisors and supervisees on the extent and importance of which supervision impacts client outcome.

Demographic Differences Between Supervisor and Supervisee Responses
Through a series of independent-sample t tests and one-way ANOVAs, no significant differences were found within supervisors regarding the extent or importance scores for gender, age, therapeutic modality, years of supervising, or between supervisors who used outcome measures and those who did not. Furthermore, within supervisees, there were no significant differences found within extent and importance scores for gender, age, therapeutic modality, degree type, months of experience, number of clients, or between supervisees whose supervisors used outcome measures and those who did not.

Quantitative Findings
To compare quantitative findings across supervisors and supervisees, their responses to the four questions that made up extent of supervisor impact, and the four questions that made up importance of supervisor impact, were totaled separately and then averaged. The average score for each main question can be found in Table 1.

Extent of supervisor impact on outcomes
Overall scores. A series of independent-sample t tests were used to compare the average extent of impact scores between supervisors (n = 189) and supervisees (n = 191). Single-item analysis. A series of independent-sample t tests were used to compare differences between supervisors (n = 189) and supervisees (n = 191) in how they responded to the four items related to the extent of impact (1-4). There were no significant differences found between supervisors and supervisees within the items related to the extent of impact.

Importance of supervisors' impact on client outcomes
Overall scores. A series of independent-sample t tests were used to compare the average importance of impact scores between supervisors (n = 189) and supervisees (n = 191). On average, supervisors showed positive scores of M = 3.87, SD = 0.56, and supervisees also showed positive scores of M = 4.06, SD = 0.62, with a mean difference of M = −0.18, 95% CI = [−.30, .06], between the groups. This difference was significant, t(378) = −2.97, p < .05, d = 0.23, indicating that supervisees find the impact of supervision significantly more important for client outcomes than supervisors.
Single-item analysis. A series of independent-sample t tests were used to compare differences between supervisors (n = 189) and supervisees (n = 191) in how they responded to the four items related to the importance of impact (5-8). There were no significant differences between supervisors and supervisees on Item 7 or 8. Significance levels and t-scores for all items can be found in Table 1.
Item 5 asked about the importance of supervision in increasing the positive impact of counseling/psychotherapy.

Qualitative Data
One hundred nine (57.67%) supervisors and 120 (62.83%) supervisees responded to the qualitative questions. The qualitative methodology aimed to examine the various ways that supervisors and supervisees perceived supervision to impact client outcome.
Supervisor responses. In response to the question "ways that feel your supervision has a positive impact on your supervisees' clients' outcomes," the highest responses indicated references to personal/professional growth by supervisees. Examples such as "helping supervisee identify and address their own 'self of therapist' issues that arise" and "professional development issues for supervisees" accounted for slightly less than half of all supervisor responses. The second most frequently endorsed response specified references to the therapeutic relationship as most impactful to positive client outcomes. Supervisors reported examples such as "By drawing attention toward issues of countertransference and interpersonal impact of client on clinician and vice versa" and "challenging them to develop a collaborative relationship with their clients." Descriptive statistics and content of qualitative supervisor data can be seen in Table 2.  Supervisee responses. In response to the question "ways that you feel your supervisor has a positive impact on your clients' outcomes," the highest endorsed responses related to references to the supervisory relationship with examples such as "my supervisor helps me feel validated and supported, and me going into therapy with that sense is a benefit to clients, leading to better outcomes for them" and "by supporting me and making sure that I am being the best therapist I can be for those clients." The second most frequently endorsed responses referred to references to learning psychotherapy models and techniques. These responses consisted of examples such as "she provides great treatment and activity ideas, since she has some specialty areas that I also have clients in" and "she helps by offering techniques and strategies I can use within session with my clients." Descriptive statistics and content of qualitative supervisee data can be seen in Table 3.

Discussion
The purpose of this study was to explore the perceived extent and importance of impact that supervision has on client outcome. The first overarching finding within the study was that both supervisor and supervisee participants appeared to have a positive endorsement (i.e., agreement that supervision was important and impactful) in their responses to the research questions: Research Question 1: What is the perceived impact of supervision on client outcome? Research Question 2: How important is it to supervisors and supervisees that supervision affects client outcome?
These results are not surprising, given that current models of supervision indicate that one of the primary tasks of supervision is to ensure the best outcome for clients and would therefore suggest that supervisors and supervisees would likely assume it is both beneficial and important (Falender & Shafranske, 2004). Participant responses were consistently positive across all four major areas of client outcome that were addressed (increasing positive impact of treatment, reducing client deterioration, reducing client dropout, and increasing speed/efficiency of treatment). These findings have strong support from both the quantitative and qualitative data, suggesting that there is a stark contrast with how positively supervisors and supervisees view the impact of supervision in comparison with current outcome research, which suggests that supervision may have a smaller impact on outcome than previously believed (Ladany & Inman, 2012). Notably, in the qualitative data, only six supervisors and seven supervisees expressed hesitation or doubts about the impact of supervision on client outcome. The responses of both groups (supervision and supervisees) to the two major research questions revealed significant correlations, suggesting that participants who thought that supervisors can impact client outcome also thought that it was important that "My supervisor helps me feel validated and supported, and me going into therapy with that sense is a benefit to clients, leading to better outcomes for them" "By supporting me and making sure that I am being the best therapist I can be for those clients" Psychotherapy models and techniques 29 24.17 References to learning psychotherapy models and techniques "She provides great treatment and activity ideas, since she has some specialty areas that I also have clients in" "She helps me by offering techniques and strategies I can use within session with my clients" Supervisee growth 27 22.50 References to personal/ professional growth by supervisees "Developing my personal sense of adequacy and competency" "My supervisor positively impacts my client's outcomes by also paying attention to my own professional development" General supervision techniques 25 20.83 Topics endorsed by fewer than 2% of responses "My supervisor attends treatment team meetings" "Helps me be assertive with charging and being paid by clients" Therapeutic relationship 19 15.83 Specific references to the therapeutic relationship and countertransference "Helping me sort through countertransference" "She helps me get my own countertransference out of the way so I can really be there fully for my clients" Connecting specific aspects of supervision with specific benefits to client 17 14.17 Explicitly linking aspects of supervision to specific types of improvement in client outcome "Helps me to be confident in my work and I think that my clients can sense this confidence and then they in turn feel confident that they will benefit from the therapy" "My supervisor helps me process my own reactions to my clients, which increases my awareness and allows me to be as effective as possible with my clients" Tools used in supervision 16 13.33 References to specific tools used for supervision (e.g., books, audiotape) "Goes over audio recordings of session" "Watching videos of sessions" Conceptualizing cases 9 7.50 Specific references to case conceptualizations "Re-conceptualize cases when necessary" "Addressing diagnostic possibilities" Self-care 7 5.83 References to supervisees' self-care "Encouraging self-care" "Reminding me to pay attention to myself and maintain my health" Qualifying limits 4 3.33 Expressing doubts about, or qualifying the limits of, the impact supervisors may have on client outcome "I'm not sure if it is helpful" "At least, I hope it does!" Ethics 3 2.50 References to professional ethics "My supervisor supports my ability to keep the highest ethical standards" "Helping to clarify and promote counseling ethics in our practice" Not helpful 3 2.50 Supervision not benefiting client outcome "At this time they do not" "I enjoy meeting with her once a week for an hour but to tell the truth I am not sure I would consider her helpful" a Percentage of 120 supervisees who responded to this question.
supervisors do impact client outcome, and vice versa. Although these perceptions match the current theory proposed by most major supervision models, it begs the question of how supervision goes about impacting client outcome, both positively and negatively. Differences between how supervisors and supervisees responded to questions related to the extent of impact and importance of impact of supervision on client outcome were examined. Overall, there were no significant differences between supervisors' and supervisees' perception of the extent that supervision can impact client outcome, suggesting that both groups find supervision equally impactful on client outcomes.
Conversely, the findings revealed that supervisors and supervisees did differ in their perceptions on the importance of supervision, with significant differences between the two groups in how they responded to questions relating to the importance of supervision impacting client outcome. The data indicated that supervisees perceive supervision as more important to impacting client outcome than supervisors do. In addition, within the construct of importance to impact, supervisees perceived supervision as more important than supervisors for increasing positive impact of therapy, as well as for preventing or reducing client deterioration.
The second goal of this study was to explore participants' opinions regarding how supervisors affect client outcome. Participants responded in a variety ways, with responses ranging from a supportive supervisory relationship to general techniques and tools. Participants also reported that various aspects of supervision that address specific treatment variables benefit client outcome (e.g., "psychotherapy models and techniques," "therapeutic relationship"). Conversely, both supervisors and supervisees reported that the aspects of supervision that address these variables at the supervisee level would also benefit client outcome. For example, "supervisee growth" was the most frequently reported category for supervisors, and the third most frequently reported for supervisees. Further research into what constitutes "supervisee growth" and how it has a positive impact on client outcome is needed.
Two notable differences within the qualitative data between supervisor and supervisee responses stood out. Supervisees reported the "supervisory relationship" as more impactful to client outcome, whereas more than twice as many supervisors endorsed the "therapeutic relationship." These findings suggest a significant difference in perspective between the two groups regarding the relative importance of the supervisory and therapeutic relationship. In addition, these results suggest that further research is needed in exploring what aspects of both the supervisory and therapeutic relationships in clinical supervision are impacting client outcome, if at all.

Implications and Future Directions
Overall, the data suggest that the impact of supervision on client outcome, as perceived by the supervisors and supervisees, is perceived to be a complex process by both groups, involving a coalescence of treatment, supervisee, and supervision variables. Supervisors and supervisees both view supervision as important; however, supervisees feel supervision is more important in comparison with supervisors. These may suggest that supervisees value more support, feedback, and guidance from their supervisors, regardless of potential impact on client outcome. Despite these differences, the data suggest that both groups feel supervision is important to client outcome. These findings would seem predictable of "common sense." However, the results of this study have yet to be quantitatively or qualitatively established in the existing literature. Furthermore, this research is necessary to identify what supervisors and supervisees believe is impacted by supervision, if not client outcome (e.g., supervisee professional identity, supportive environment for supervisees, etc.).
Current research suggests that supervision, theoretically and in clinical practice, has a minimal impact on client outcome, despite the primary task of supervision being to ensure the best possible outcome for the client. Most supervisors and supervisees proceed within supervision as if it is both important and impactful to client outcome. However, if client outcome is not being impacted by supervision, further research is needed to determine what in supervision may actually be impacted. Furthermore, if we proceed as though supervision is both important and impactful to client outcome, future studies would need to examine what aspects of clinical supervision actually impact outcome and what practices would need to be changed to ensure a positive impact on client outcome, allowing for an opportunity to improve upon future models of supervision.
Given the discrepancy between study participants' beliefs that supervision is both impactful and important, and the outcome research that suggests a smaller effect, there is a clear need to uncover what it is about supervision that results in this endorsement by supervisors and supervisees. It is possible that a similar self-assessment bias that Walfish, McAlister, O'Donnell, and Lambert (2012) found among practicing psychotherapists is occurring at a supervision level, where supervisors and supervisees have a tendency to overestimate the ability of contemporary supervision practices to impact client outcome. Supervisors and supervisees may be using positive perceptions of supervision as another means of maintaining morale in the face of challenging clients within a demanding field. "Self-assessment bias may simply be the unconscious attempt to stay motivated in the face of very difficult client problems and circumstances" (Walfish et al., 2012, p. 642).

Limitations
There are multiple limitations that are often associated with a nonstandardized, self-report survey. Underreporting of negative behaviors and responding to questions in a way that portrays the responder in a positive regard are the two most common limitations associated with this method of research (Donaldson & Grant-Vallone, 2002). In addition, nonstandardized, self-report survey research may also focus too broadly on a range of topics, which may result in a lack of depth or full understanding in the topic at hand. The qualitative portion of this research was conducted with these limitations in mind, and the shortcomings of this research methodology were taken into account when examining the results.
Another limitation of this study may be attributed to the fact that the survey only asked for supervisees to comment on their experiences with their current supervisor. It is possible that supervisees may have had a different experience with supervision overall. It should be noted that supervisees may have responded differently to the study if the question inquired about overall supervision experiences.
A major limitation to this study pertains to the question used to address the qualitative aspect of this study. The qualitative question had a narrow focus, potentially limiting the range of answers that supervisees and supervisors could have provided. By specifically asking supervisees what their supervisors contributed to having a positive impact on their clients' outcome, the opportunity to explore an organic, broad-range response from supervisees about what they felt was impactful was lost. While a more open-ended question may have been beneficial in this instance, the exploratory nature of this study benefited with the specific question to use as a reference point to guide future research on this topic.
Several of the limitations within this study can be attributed to the restrictive nature of the qualitative portion, which was constrained by the format of an Internet-based survey. This medium of data collection prevented any modification to the research question, limiting the spontaneity of natural responses between participant and researcher. Future research could implement the interview approach to allow for more in-depth responses from participants and a more conductive relationship between researcher and participant.
The mixed-methods approach of this study may have impacted potential responses from participants. The quantitative portion was administered before the qualitative questions, which may have primed participant responses. In addition, the qualitative coding done by the researchers may have been unknowingly influenced by the biases of the team members. Although precautions were taken in an attempt to combat these biases, it is possible that biases shared by the three-team members may have potentially influenced the results. Further research would be helpful to capture a wider understanding of what supervisees find helpful to client outcomes from their supervisors and would be able to account for the potential limitations that exist within this study.
In summary, these results indicated that both supervisors and supervisees believe that supervision has an impact on client outcome; however, there is no clear consensus of how supervision impacts client outcome and to what degree. Further research is needed to explore what aspects of supervision, if any, impact client outcome.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article.