Introduction
One of the key elements of supervision is the ability to give and receive effective feedback (Milne and James, Reference Milne and James2000, Reference Milne and James2002; Milne, Reference Milne2009; Roth, Pilling and Turner, Reference Roth, Pilling and Turner2010). Milne's empirically constructed definition of supervision places feedback as a key feature:
[Supervision is] the formal provision, by approved supervisors, of a relationship-based education and training that is work focused and which manages, supports, develops and evaluates the work of colleague/s. The main methods that supervisors use are corrective feedback on the supervisee's performance, teaching, and collaborative goal-setting. It therefore differs from related activities, such as mentoring and coaching, by incorporating an evaluative component. (Milne, Reference Milne2009).
In Roth and Pillings’ (Reference Roth and Pilling2007) map of CBT competences, supervision is viewed as a generic feature in their five-element framework, suggesting that the supervisor's role is to provide effective feedback to “allow for the identification and remediation of sub-optimal performance” (p. 23). However, they also highlight that supervision is an interactive relationship, requiring input from the supervisee. They state that being an effective supervisee is an active process, requiring a capacity to be reflective and open to criticism, willing to learn and willing to consider (and remedy) any gaps in competence that supervision reveals (p. 13).
This paper has been written for therapists and their supervisors. However, owing to the lack of relevant literature in the field, it has been necessary to draw on the wider literature from teaching, industry, and coaching as sources of information. Whenever the studies relate specifically to clinical work, this feature is highlighted.
Kluger and DeNisis’ (Reference Kluger and DeNisi1996) meta-analytic review of studies, which had included a feedback condition, found an average effect size in favour of the use of a feedback intervention of 0.41. However in numerous studies within the review the control groups outperformed the feedback groups. In explaining their overall findings, the authors proposed that feedback interventions changed the locus of attention among three levels of control: task learning, task motivation, and meta-tasks (including self-related) processes. Within the feedback conditions, effectiveness decreased as attention focused more towards “the self” and away from the task. The latter review highlights that asking the question of whether feedback is effective is not a particularly informative approach. Rather, a more sensible strategy would be to pose questions about the effects of various types of feedback in different settings. Once feedback is examined from this perspective, a Pandora's box of complexity is opened, and some of the key themes emerging are addressed in this paper. For example, Hattie and Timperley (Reference Hattie and Timperley2007) distinguished four levels of feedback: (i) feedback about the task; (ii) feedback about the processing or strategies used to do the task; (iii) feedback about self-regulation related to the task (self-evaluation and self-confidence); (iv) feedback about the trainee as a person. According to these researchers, feedback from (i) and (ii) are the most effective in producing change, while the other two (iii) and (iv) can often fail to draw the trainees’ attention to the task of learning. Feedback about the person (e.g. You are very intelligent/disorganized) is not regarded as helpful for two reasons. First, because it does not contain information that can be used to improve learning. Second, such feedback can give the impression that the abilities highlighted reflect unchangeable traits (Brookhart, Reference Brookhart2008) and thus achievement on this aspect is beyond the trainees’ control. Dweck and colleagues (Dweck, Reference Dweck2006; Mueller and Dweck, Reference Mueller and Dweck1998), from an educational perspective, have shown empirically that giving feedback about trainees’ abilities, even praise, can lead to a significant decrease in performance. Her studies show that praising individuals can lead to: less risk-taking, cheating, fear of failing, and a decrease in overall performance. Dweck's studies also reveal that when feedback is directed at levels of trainee effort, students are motivated to attempt more difficult tasks, cheat less, and their performance improves.
Butler and Nisan (Reference Butler and Nisan1986) found that the manner in which a recipient experiences feedback is important in relation to the amount of learning and motivation achieved. Their work was built on the earlier studies of Page (Reference Page1958), who found that “free comments written by a teacher” were associated with higher subsequent achievement compared to “letter graded” feedback (i.e. A, B, C, F). In the later Butler and Nisan study, three different conditions were compared: numerical grades vs comments vs no-feedback. The students were assessed in terms of performance and motivation for tasks involving both quantitative and divergent-thinking tests. In each case a particular form of feedback was provided in the early stages of the testing, and then performance assessed for the final tests. The results revealed that the provision of descriptive comments resulted in high performance and motivation, whereas those receiving no feedback performed poorly and were the least motivated. In relation to “graded” feedback the results were more complex, with good performance achieved on the quantitative tests, but poorer scores on the divergent-thinking tests. The latter participants also displayed poorer motivation compared to the descriptive comments condition.
Brookhart (Reference Brookhart2008) believes the latter programme of research is particularly helpful because it illustrates some general principles with respect to feedback. For example, it demonstrates the importance of targeting both performance and motivation as measures of effectiveness. It also demonstrates the advantage of giving descriptive over graded feedback. It is relevant to note, however, that when grades and descriptors appear together, grade tends to trump description. Thus descriptive comments have the best chance of being read as being relevant if they are not accompanied by a grade (Brookhart, Reference Brookhart2008, p. 7).
Brookhart (Reference Brookhart2008) provides a comprehensive review of the nature of how feedback should be delivered, drawing attention to the areas in which the content of feedback may vary: focus, function, comparison, valence, clarity, specificity and vocal tone. Her non-empirical overview provides recommendations on how to deliver good feedback in each of the seven areas outlined. In the field of counselling, Friedlander, Siegel and Brenock (Reference Friedlander, Siegel and Brenock1989) examined feedback using their Supervisory Feedback Rating System (SFRS). The SFRS specifies four dimensions of feedback: types (interpersonal or cognitive behavioural); specificity (global or specific); valence (positive or negative); and focus (the counselling or the supervisory relationship) (Friedlander et al., Reference Friedlander, Siegel and Brenock1989, p. 151). In a case study, Friedlander found that feedback tended to be given infrequently, and when it was given it tended to be interpersonal, global and positive.
The valence of feedback has received empirical attention (Kopala, Reference Kopala1987; Lane, Daugherty and Nyman, Reference Lane, Daugherty and Nyman1998; Daniels and Larson, Reference Daniels and Larson2001), and there is on-going debate about the value of positive versus negative feedback. For example, McCoy (Reference McCoy2005) suggests that a supervisor should try to avoid negative feedback. She thinks that the supervisor should look for occasions when to give positive feedback and plan around occasions when negative feedback is required, making the latter constructive. As above (Brookhart, Reference Brookhart2008), McCoy cautions against feedback that criticizes the trainee rather than their actions. She suggests that criticism should be factual, impersonal and timely. Using her principles, an example of what she would advocate in a clinical setting might be:
This week I’ve noticed you did not set homework tasks for three out of five clients because you ran over time. As you know homework is a major component in helping the person to get better, and I know you really want your clients to improve. So what can we do to ensure your clients get the best chance to get the most from your therapy?
Daniels and Larson (Reference Daniels and Larson2001) have undertaken a direct comparison of feedback valence in relation to self-efficacy and anxiety in the context of counselling. Forty-five volunteers from clinical trainee backgrounds were recruited and assessed over a 10-minute mock counselling session. After the session the volunteers were given a rating out of a 100, and standard statement about their performance (either positive or negative). Results from scores taken pre and post the session revealed that those receiving positive feedback scored higher on self-efficacy and became less anxious, whereas negative feedback resulted in poorer self-efficacy and an increase in anxiety. The authors concluded that supervisors should ideally attempt to increase trainees’ sense of mastery via positive feedback, and thereby improve confidence, building on existing skills and developing new ones through positive nurturing. It is relevant to note that anxiety within supervision is not always a problem. For example, James, Allen and Collerton (Reference James, Allen and Collerton2004) have shown that new learning occurring at the edge of a trainee's current level of understanding (i.e. within the zone of optimal development, ZPD, Vygotsky, Reference Vygotsky1978), although causing anxiety, can promote knowledge and pride. However, when working within the ZPD it is important that learning is supported appropriately by supervisory techniques, such as scaffolding, in order for the levels of anxiety not to be debilitating (James, Milne, Morse and Blackburn, Reference James, Milne, Morse and Blackburn2007).
Let us now look at various methods for giving feedback in clinical supervision, determining which are better at being able to motivate, improve practice, within a positive and collaborative relationship.
Feedback methods
Research suggests that supervisors provide insufficient feedback (Larson, Reference Larson1998) and frequently avoid critical feedback (Hoffman, Hill, Holmes and Freitas, Reference Hoffman, Hill, Holmes and Freitas2005). We have learned that supervisors have a role to play in increasing self-efficacy, and reducing dysfunctional forms of anxiety, by using constructive techniques that build on existing skills. According to Daniels and Larson (Reference Daniels and Larson2001), supervisors should:
. . . actively increase the trainee's counselling self-efficacy by focusing on techniques that have been effective for the trainee during training sessions. Feedback that enhances the positives and is accompanied by specific suggestions about ways to improve should lead to higher counselling self-efficacy, lower anxiety, and more confidence. (p. 128)
The next section will examine a small number of methods taught to structure supervisors’ feedback, to increase its use, and to achieve the above goals. I will start with potentially the most critical of the techniques, leading to some more constructive and collaborative approaches.
One of the techniques developed to make negative feedback more palatable was the Sh*t Sandwich (SS). It suggests that critical feedback should be prefixed and suffixed by positive statements that serve to cushion the criticism. For example, one might say:
Overall, it was a good session today, BUT I think you set too many items in the agenda. Despite the pressure you put yourself under, I think you handled the situation really well by incorporating the agenda items not discussed within the homework task . . . that was great!
To work well, the SS must be delivered with skill in order not to appear patronising. For example, the three elements (opening, middle and closing statements) need to be both relevant and matching. A poor match might be: “I like your voice, BUT you have a tendency to out-pace your client. . . however, the therapy room was arranged really well.” In this example, the opening statement is not specific enough (i.e. ideally, one needs to say what one liked about the therapist's voice). And the final statement is not related to the first two, perhaps indicating that one is struggling to say anything else positive about the trainee's mode of delivery.
I used this method for many years in both my clinical and supervisory practices, until I began to recognize a further problem relating to the fact that some trainees found the critical element of the SS feedback too subtle, resulting in them failing to reflect on what needed to change. Struggles with this particular feature led me to the work of Daniels and Larson (Reference Daniels and Larson2001), who – as outlined above – advocate more positive and constructive approaches. One of the first approaches I found of this type was the work of Pendleton and colleagues (Pendleton, Schofield, Tate and Havelock, Reference Pendleton, Schofield, Tate and Havelock1984). This approach has been used for many years in education, and is composed of the following seven steps:
1. Ensuring the learners are aware of the purpose of the feedback.
2. Learners commenting about the goals they were trying to achieve during their task.
3. Learners stating what features of the task they thought they’d done well.
4. Supervisor stating what features were done well.
5. Learners stating what could be improved.
6. Supervisor stating what should be improved.
7. Agreeing action plans for improvement.
(adapted from the Londondeaneryeportfolio, 2013)
The critics of this model suggest that it can be overly rigid, and may be prone to becoming bogged down in detail, particularly when used in a group format. Nevertheless, its supporters state that learners quickly become socialized to the model and are able to reflect on their performance prior to questions being asked within their supervision sessions. Thus, even if a supervisor does not ask about a certain topic, the learners may improve in areas not covered in supervision because the supervisees have reflected on potential questions in anticipation of them being raised in supervision.
A similar approach was developed by Vassilas and Ho (Reference Vassilas and Ho2000). Their problem-based perspective recommends:
1. Setting an agenda.
2. Examining the function of the session.
3. Encouraging self-assessment and problem-solving.
4. Using descriptive feedback.
5. Using balanced feedback (what worked and what could be done differently).
6. Suggesting alternatives, modelled by the teacher and then rehearsed via role play.
7. At the end, giving a structured summary of the learning points.
Both of these feedback models place emphasis on the trainee as an active participant in the learning process, and thus are often referred to as “interactive approaches”. Hill (Reference Hill2007) believes that the interactive methods avoid a number of pitfalls. For example, the approaches prevent the build-up to the “. . . BUT – juncture”, which those socialized to the SS model begin to expect whenever they hear a positive statement being made about them. Also, by setting an agenda, the supervisor obtains a good understanding about the level of self-awareness of the learner. Therefore an “over” or “under” critical learner can be guided to making more realistic assessments. Further, the structured summary and future planning aspects help to obtain agreement on the areas of greatest need. The interactive approaches also have the advantage in relation to the issue of valency, because by asking the trainees to comment on their own performance in a specific area (e.g. setting agendas, change techniques), the supervisor is provided with opportunities for collaborative analysis and constructive comments to be made in relation to the trainee's own reflection on his/her performance.
According to Hill, the importance of the various versions of the interactive approach cannot be over-stressed. She believes that these methods allow the supervisor to facilitate a trainee's reflection and self-feedback, helping the trainee to move through stages of learning – from “unconscious incompetence” to “unconscious competence”. In the former state the trainee is unaware of their failings, but gradually grows in awareness via self-observation and reflection. The next step is to get the person to become “consciously competent” via the use of effective feedback. At this stage, the clinician can perform a skill at will, but the skill still needs conscious support and is not yet second nature nor automatic. Eventually, through practice, the trainee will become an “unconsciously competent” clinician – competent, but unaware of the detailed processes involved in his/her activities and no longer needing regular feedback.
Further, it is relevant to note that unlike the rather unidimensional SS methodology, the interactive methods are consistent with a number of learning theories, such as Kolb's experiential learning cycle (Reference Kolb1984). In Kolb's cycle, effective feedback facilitates the movement around the cycle consolidating the acts of reflection, experimenting and conceptualizing.
All of the feedback processes outlined above can be further enhanced via the “chronological account” (CA); in this time-sampling method the supervisor keeps detailed records of the learner's performance in a session via either direct- or media-observation. With such detailed time-series observations, the supervisor is able to give precise feedback across time. While undoubtedly extremely helpful, it is a time consuming method and therefore has limited use in many clinical settings. Nevertheless, the CA is useful when employed in conjunction with the interactive methodologies (Hill, Reference Hill2007). For example, a supervisor might ask a trainee to record his/her session and review the recorded therapy prior to supervision, choosing excerpts that demonstrate good and poor skills. These CA features can then be incorporated within the interactive approach. For example, the recorded excerpts could be used at step 5 in Vassilas and Ho's (Reference Vassilas and Ho2000) feedback list.
This short review has attempted to highlight some areas in relation to this important, but poorly researched, topic. There are many aspects of feedback that could be investigated, with Brookhart (Reference Brookhart2008) conservatively highlighting seven potential areas that would be worthy of study (see above). Daniels and Larson (Reference Daniels and Larson2001) have also demonstrated the value of mock analogue studies, and Milne's (Milne and James, Reference Milne and James2002) single-case studies highlight the relevance of time-series analyses. Our knowledge with respect to feedback could also be extended by analyses of the numerous competence rating scales (CTS, Young and Beck, Reference Young and Beck1980; CTS-R, Blackburn, James, Milne and Reichelt, Reference Blackburn, James, Milne and Reichelt2001; CTS for Psychosis, Haddock et al., Reference Haddock, Devane, Bradshaw, McGovern, Tarrier and Kinderman2001) used in the normative and formative assessment of therapy. All of these scales contain items on feedback, and it would be interesting to examine “feedback scores” in relation to both therapy outcome and against the profile of scores on other items of the scale.
Conclusion
This paper describes the relevance of feedback in improving clinicians’ skills and client outcomes. It has been argued that the SS method of feedback has limitations compared to the more integrative approaches, and the latter methods are more consistent with the “learning-theory” literature. The value of providing positive and constructive information has been discussed, and using descriptive feedback in favour of ratings has been argued. The paper has also highlighted the relevance of praising effort over ability, because telling someone they are gifted or a “natural” may lead to a reduction in performance.
Comments
No Comments have been published for this article.