The therapeutic relationship was first outlined in Cognitive Behaviour Therapy (CBT) as a primary source of information or ‘raw data’ for the case formulation in A.T. Beck, Rush, Shaw, and Emery's (Reference Beck, Rush, Shaw and Emery1979) Cognitive Therapy for Depression (c.f. Dattilio & Padesky, Reference Dattilio and Padesky1990; Dattilio & Hanna, Reference Dattilio and Hanna2012; Scott & Freeman, Reference Scott, Freeman, Kazantzis, Reinecke and Freeman2010). Embedded in this description is the concept of empiricism — the process of contextualising therapy within the patient's unique experience. Empiricism is a basic strategy or form of interaction between patient and therapist that pervades both structural processes (agenda, homework, pacing), relationship (J. S. Beck, Reference Beck2005), and the techniques of CBT (Hollon & A.T. Beck, Reference Hollon, Beck, Kendall and Hollon1979). One of the key objectives for the therapeutic process in CBT is to enable experimentation and empirical evaluation of beliefs and behaviours outside therapy via homework assignments (A.T. Beck, Reference Beck1970; J.S. Beck, Reference Beck2011; Kazantzis, Deane, Ronan, & L'Abate, Reference Kazantzis, Deane, Ronan and L'Abate2005).
An overlapping but distinct characteristic of the ideal CBT patient-therapist relationship is collaboration. The definition of collaboration within CBT requires an active role on behalf of both the patient and therapist (A.T. Beck et al., Reference Beck, Rush, Shaw and Emery1979). This active role reflects an intention to have a genuine sharing in the identification of problem situations, and in the design, implementation, and evaluation of empirical tests of the patient's beliefs. Thus, collaboration is separate from, but also pervades, the structural processes (agenda, homework, pacing), relationship (J. Beck, Reference Beck2005) and techniques of CBT. Although collaboration and empiricism have been heralded as distinguishing features of the therapeutic relationship in CBT (Hollon & A.T. Beck, Reference Hollon, Beck, Kendall and Hollon1979; Madsen, Reference Madsen2009; Persons, Reference Persons2008; Tryon & Misurell, Reference Tryon and Misurell2008), there exists relatively little guidance for the practising clinician about how to develop a relationship that is strong in these elements.
The Australian Association of Cognitive and Behaviour Therapy (AACBT) held its 33rd national conference in 2010. In alphabetical order, invited keynote speakers to the AACBT national conference — Drs Arthur Freeman, Alan Fruzzetti, Jacqueline Persons, and Mervin Smucker kindly agreed to join in a panel discussion to respond to questions regarding the definition of collaborative empiricism, the mechanism by which collaborative empiricism facilitates cognitive change, and highlighted the importance of collaborative empiricism for effective therapy (Dr Nikolaos Kazantzis was Chair). The session was well attended and the panel enjoyed the opportunity to discuss questions and comments from those in attendance. This article presents those questions and their responses so that the broader AACBT membership may benefit from the ideas and comments generated.
Definition of Collaborative Empiricism
There has been considerable variation in the way in which collaborative empiricism has been discussed in research and practice contributions to the CBT literature (Tee & Kazantzis, Reference Tee and Kazantzis2011). Collaborative empiricism is sometimes referred to as the major therapeutic method used by A.T. Beck (Abramson, Alloy, & Dykman, Reference Abramson, Alloy and Dykman1990), a method of reality testing (Bell, Grech, Maiden, Halligan, & Ellis, Reference Bell, Grech, Maiden, Halligan and Ellis2005), and as a cognitive strategy (Dattilio, 2000). In other contexts, rather than a method or specific technique, collaborative empiricism has been labelled a ‘philosophy’ or ‘foundation of CBT’, and a ‘technique’ and ‘atmosphere’ that pervades the client-therapist interaction (Wilkes, Reference Wilkes1994). Collaborative empiricism is described as the ideal CBT alliance and as the basis of the alliance. Perhaps collaborative empiricism is all of these things and there is no contradiction, or perhaps it is none of these things. Thus, a first unresolved issue concerning collaborative empiricism in CBT relates to its definition. The following questions were asked of the panel regarding the definition:
1. What do we mean by collaborative empiricism?
2. And if we were a fly on the wall, a third-party observer, a colleague, a supervisor — what does collaborative empiricism look like? How do we know when/if we are facilitative of collaborative empiricism?
Jacqueline Persons
To me, ‘collaborative empiricism’ is working together to collect data to test hypotheses. And it could be a lot of different types of hypotheses about causes of problems, hypotheses about what is going to be helpful in therapy. A long list of different possibilities.
Art Freeman
I agree. I think there are two pieces. Beck has always talked about cognitive therapy being the scientific method, and I think Jackie exemplifies that so beautifully in her work. The idea of hypothesis statement, the idea of collecting data, and the evaluation or analysis of that data exemplify the empiricism we see as the basis of cognitive therapy. But there's another piece that we are somehow skipping over and that's collaboration. Collaboration is a whole separate issue.
Collaboration is not ‘50–50’, to some patients collaboration is ‘20–80’. I have students who will say to me: ‘Well, my professor said that the therapist shouldn't work harder than the patient.’ This is stated by a therapist who doesn't do therapy. This is stated by someone who is teaching it, but doesn't do it. If you're working with an adolescent and you think you can get them to do more than half the work, I want to study at your knees, because you've got a skill that I want to learn.
So, I think we have two pieces here. One is: collaboration has to be assessed. How much can the patient bring to the scientific endeavour? That's the collaborative piece. And then, if they can produce 3%, we've got to produce 97%, and then we have to work, given that context, on helping them develop an empirical view, a scientific view of dealing with their issues. So collaborative empiricism is a very, very complex term. It's deceptively simple just to see it on the screen.
Alan Fruzzetti
I agree, and I agree. As a person representing Dialectical Behavior Therapy [DBT], if it's okay I would like to push this a little bit further than within cognitive therapy, to think about collaborative empiricism. Because I think that more generally when we aren't necessarily looking for the utility or the veracity of the thought, we end up with a different conundrum I think, which is: Are we trying to establish a collaboratively, empirically descriptive reality, or a useful hypothesis about that reality? At the risk of making this too philosophical, when we talk about emotion, we can say that there is a kind of descriptive reality to emotion, the experience of it, the phenomenology of it, but there are also the concepts we attach to that, what it is. As soon as we move away from the phenomenon of an emotion, or any phenomenon like [taps table] this is a table, it's a real phenomenon, but once we hollow the table we apply a concept to it, it's a very useful concept because we all have the word ‘table’, we have the utility of ‘table’ in our brain.
With patients of course we are dealing with people who have experiences which are distorted, so even getting to descriptive reality can be very complicated. So I think, in DBT anyway, the process of collaborative empiricism is exactly as Jackie and Art said, but it's not only around thinking, it is also around descriptive reality of experience. In DBT we have other words for it — we use ‘mindfulness’, describing, we also talk about primary emotions. Primary emotions are like the table: it's here, it's very close to the real phenomenon. As opposed to secondary emotions, which are emotions about emotions, or reconditioned emotions. So in a situation where 99% of people might feel sad, this person's very angry. You might say that's not entirely descriptive reality, that's interpreted reality in a sense. So in fact, much of DBT, and I think much of psychotherapy, when it works well, doesn't clue this process of collaborative empiricism, ‘figuring’ that out. I think that's a bit philosophical but I think that's worth paying attention to.
Mervin Smucker
I would like to pick up on this notion of collaboration that Art mentioned a moment ago. When Art and I were at Beck's clinic in the 80s and early 90s, one of the things that Beck continuously would emphasise is, in addition to everything else which is emphasised in cognitive therapy, we need to never forget that what we are trying to do is help the patients to become their own cognitive therapists. And in doing so, we need to work with them collaboratively and teach them this concept of collaborative empiricism by modeling for them the questions that we ask in the cognitive therapy that we do. Because the ultimate goal is for patients to become their own cognitive therapists so they can learn how to think, and problem solve, and ask themselves the right questions. So if we are giving them the answers as we are going through and telling them how to think rationally, we're not teaching them how to think. That was one thing I wanted to emphasise and make it a bit more specific, even when it comes to homework. You don't say as the cognitive therapist: ‘Now I want you to do “such-and-such” and “such-and-such” for next time for homework.’ Going in this direction with the collaborative empiricism:
What might be a good homework assignment for us to consider, for you to do for next time?
What might be the pros and cons for that?
Well, would you be willing to test that out?
Then of course at the beginning of the next session, what you do is you spend the first 5 minutes:
How was the homework?
What went well?
What didn't go well?
How might we conceptualise or how might we . . . for next time?
What might be a good homework assignment?
So in the way that we ask questions to our clients we're actually modelling for them and teaching them how to think, and this is what Beck was totally emphasising 25 to 30 years ago.
Art Freeman
You just did something Merv that was so beautifully done, I'm not even sure you knew you did it. You started by saying that we don't just want to tell the patient what to do. You say to the patient: ‘Well, what do you think would be good homework?’, thereby inviting the patient to contribute to the construction of the homework. Let us suppose the patient looks at you and goes: ‘Well, I don't know.’ What would you do? You said: ‘Well, what do you think would happen if you tried “X”?’ So in point of fact, you didn't just wait for them to come up with it; if they are having trouble coming up with it, you jumped right in collaboratively and said: ‘Well, what do you think would happen . . .’. And offered a suggestion, a very clear direction. And the patient then can say: ‘I don't think so.’ And then I would expect you would help them modify it. What you did was, I thought, ‘Boy, he's a really good therapist,’ because I know how well you were trained.
Mervin Smucker
He [Art] was my first supervisor 30 years ago.
Art Freeman
I hardly even remember. That is exactly what I mean by ‘collaboration’. The patient says: ‘I have given you about 6% of the collaboration, Merv. I can't figure this out. It's too much for me.’ So you don't say: ‘Well, here's what I want you to do.’ You give them a lead, and you get them to now take on 7.6% of the collaboration. And then teaching them to be their own therapist, maybe eventually you'll get them to be 10% collaborative, and then maybe 18% collaborative. It may never be 50–50. When I work with couples, one of my favorite — maybe more correctly, least favorite — comments is: ‘Why do I have to be the one that always does the discipline? Why isn't it 50–50?’ And the therapeutic answer is: ‘Because.’ That's it. Because that's the way things are. Life is not always 50–50, so we need to think of how we are going to help the patient grow into the role of taking the lion's share of the collaboration.
Mechanism of Effect
In studies of the effectiveness of CBT, the therapeutic relationship has been most often measured as the ‘therapeutic collaboration’ or ‘working alliance’ (Horvath & Greenberg, Reference Horvath and Greenberg1989, Reference Horvath and Greenberg1994). These relationship constructs capture only part of what was intended for the therapeutic relationship in CBT. The working alliance inventory, for instance, conceives collaboration in the working alliance as an ‘agreed-upon contract’ between patient and therapist regarding the tasks and goals of therapy (Bordin, Reference Bordin, Horvath and Greenberg1994).
Bordin's (Reference Bordin, Horvath and Greenberg1994) conception of the alliance drew influence from different psychological paradigms, and centred on a strong ‘bond’ between client and therapist, and mutual ‘agreement’ on therapy goals and tasks (Horvath & Greenberg, Reference Horvath and Greenberg1994). However, collaboration in CBT is not simply willing participation by the patient, nor agreement on tasks or goals. The therapist aims to engage the patient in an active therapeutic role, progressively encouraging the patient to take the lead role in developing goals and therapeutic tasks (A.T. Beck et al., Reference Beck, Emery and Greenberg1985; A.T. Beck et al., Reference Beck, Rush, Shaw and Emery1979; J.S. Beck, Reference Beck1995). The aim of this partnership is to actively share the work of empirically evaluating the patient's beliefs, not simply to secure patient agreement with the therapist.
Self-determination theory states that people have a fundamental propensity towards growth, self-determination, and the resolution of psychological discrepancy. The theory seeks to explain how individuals self-regulate behaviour and internalise extrinsic motivation, and is concerned with the effect of environmental factors, such as directives or rewards, on intrinsic motivation to engage in behaviours (Ryan & Deci, Reference Ryan and Deci2008; Ryan, Lynch, Vansteenkiste, & Deci, Reference Ryan, Lynch, Vansteenkiste and Deci2011). Applied to the therapeutic relationship in CBT, the theory suggests that a non-collaborative attempt to persuade a client to engage with a therapeutic intervention would represent an example of an externally regulated motivation (Tee & Kazantzis, Reference Tee and Kazantzis2011). Conversely, an experience of a technique selected because of its relevance to the client's unique situation, and evaluated based on client experience (i.e., empiricism) would yield internally sourced motivation supported by self-approval. Thus, a second unresolved issue concerns the theoretical basis for collaborative empiricism, and how it supports the mechanism of cognitive change in CBT. The following questions were asked of the panel regarding the role of collaborative empiricism in facilitating change in CBT:
1. To what extent does social determination theory make sense in explaining collaborative empiricism?
2. How does collaborative empiricism support cognitive change processes in CBT?
Alan Fruzzetti
I think that this issue, how you framed it, you framed it very nicely. About trying to understand what the relationship is between collaboration as a mechanism of change, and the alliance as you call it. The data are actually pretty clear in DBT that what we call the alliance is not a predictor of outcome in DBT. It's not a mediator. Clients in DBT report very high working alliance scores, it's not that they don't have good working alliance, it's that it doesn't predict outcome at all. I think we have to be careful not to reify alliance. The Holy Grail is not to have a good relationship with the client. That's not why people come to treatment. People come into treatment because they want help; we have to have a good enough relationship to help them. ‘Good enough’ I would argue is establishing this collaboration. Some minimal amount of collaboration, which as Art so nicely put, could be in the beginning when the client is doing objectively very little, because that's all the client can do. But by the end of treatment the client is doing a lot, is actually takes over his or her own life. That's the part I actually have some data on.
Art Freeman
I think about a comment that Al [Albert] Ellis made on more than one occasion when I asked him about the relationship issue. Because people who watched Al Ellis work, would often comment that he didn't slow down enough to establish a relationship. And Al said: ‘I don't care if they like me, as long as they listen to me.’ And this was his idea of collaboration. He was saying” ‘I'm a pretty smart fellow, I really know what I'm talking about, I'm really good at what I do.’ For our colleagues here in the room, they say: ‘I come to AACBT and I go to all these workshops and I really learn a lot of stuff, and if you listen to me, you'll be successful.’ When I asked Al: ‘What percent of the people stay and listen to you?’ And he said: ‘Enough to keep me in business.’ I have no question that there were a lot of patients who after one session said: ‘I'm out of here. He's even crazier than I am.’ But those that stuck with him were very successful, he was a superb therapist. There is no ‘one size fits all’ on this relationship business, on this alliance business. For Al an alliance was: ‘If you listen to me, I'll get you well.’ That was it. And he didn't bother with all this other stuff. For Al, he believed that: ‘I know what I'm talking about, just listen.’ It worked for a lot of people.
Jacqueline Persons
It doesn't sound very collaborative.
Art Freeman
Why not? Why is that not collaborative?
Jacqueline Persons
Well, because he's saying: ‘Do what I say.’ It's not asking questions to find out, you know, what you were happy with Merv doing about the homework assignments: ‘What homework assignment do you think would be helpful to you?’
Art Freeman
Al says: ‘Here's what you have to do.’
Jacqueline Persons
Yes, I know.
Art Freeman
But the collaboration was you came to therapy, you did the homework . . .
Jacqueline Persons
Or not.
Art Freeman
. . . Or not. You listened to what he had to say. You might debate with him. You might question him. So his view of collaboration was different than what we learned at the feet of Beck. But I think I can make a case that Al was collaborative in his way.
Mervin Smucker
I'm going to take this in a slightly different direction, and pick up on what you said Nikolaos, about motivation, which I think is a real key here, and I think that interfaces very nicely with this. You mentioned about intrinsic motivation versus externally regulated motivation, or extrinsic motivation. I think this is also a key, and I think in the beginning of therapy, we need to try to establish how motivated is this client to actually make the changes we think the client needs to make. I know that early on in my work as a therapist, I would work very hard trying to get them to change, and they would resist. It was like I was trying to pull teeth, and getting them to do something that they were fighting me on, and yet they were coming to me because I thought they wanted to change. So there was at some point a dramatic shift that occurred in my work where I would spend a lot of time asking questions about:
Well, why would you want to change this behaviour?
Hasn't it been working for you?
It's not working?
Well how has it been working, and how has it not been working?
And is it that you, deep within yourself, feel you have to change this, or are you going to have to change your behavior or you're going to lose your partner?
So you really need to establish though asking questions, and I think that's part of this collaboration or collaborative empiricism, why they want to change. Having them listen to themselves and as they start to talk about why they do or do not want to change, it then becomes clearer to them what they're dealing with. Then if they say they want to change, and you say:
Well, OK, if I understand you correctly, here's some things you want to change, and here's why. Now, how might be the best way that we get there?
And how hard are you willing to work, and how many risks are you willing to take in the process?
And anytime you meet with some resistance along the way, you can say:
I'm sorry, perhaps I misunderstood, I thought this was one of the areas you said you wanted to make some changes in. Maybe it's a little more difficult than what you thought?’
Should we go back and take a look at that again, and perhaps re-examine the pros and cons, the costs and benefits, of making these changes?
I think working the motivation into this whole process right in the beginning as your establishing therapeutic goals is critical. Its going in a bit of a different direction than saying: ‘Here's what I think you need to do. ABC. Etc. etc. etc.’ This I think is extremely important. And then as a therapist, you can feel like no matter what happens, you've done OK, because if they're not motivated enough to change, OK, the motivation isn't there. And then you can simply say:
That's fair enough if you want to keep on with the behaviours that you have and really not change because it's too scary. Perhaps you're not ready now.
That is a collaboration, I think, that comes out then in the arena of motivation, goal setting, and then following through with what sort of CBT interventions you follow through with. I find it much easier as a therapist to go that route than to not have the patient ‘hooked’ in and connected in this collaborative way.
Jacqueline Persons
I'm listening to you, Mervin. If we hear what you are saying, it seems like you're describing a lot of beautiful ideas from motivational interviewing, which is so interesting (developed by Miller and Rollnick), as to how therapists and patients think about how to manage motivational difficulties in treatment. And what's really interesting is that a key piece is listening to the patient. And their idea that it's the patient that is the expert. So the therapist is opening up his or her mind, asking questions, listening, letting go of pushing, and in that way that's the collaboration. Therapists listening, asking questions, collecting data, as a way of solving the problem in contrast to pushing the patient to do what the therapist thinks the patient should do.
Mervin Smucker
I would contend that 30 years ago at Beck's clinic we were doing this, that motivational interviewing grew out of that and they borrowed it from Beck. We were doing this naturally 30 years ago as part of cognitive therapy.
Alan Fruzzetti
I agree that the motivational piece of this is very important. I think it's a tricky slope though. A slippery slope, a little bit, because in DBT we talk about the dialectic, that's the core metaprocess in DBT. I would say this is a good example of a dialectic because on the one hand we say that the patient is his or her own expert, and on the other hand in DBT we also say the patient doesn't have a clue. At the beginning of treatment. They can both be true.
That the patient is his or her own expert on their experience, what they're wanting in the moment. So in DBT we go back and forth with that, so motivation as a psychotherapeutic process is one where we have to on the one hand, I think, provide a lot of the motivation early in treatment when things are difficult, and over time if we are too good at providing that, we disempower the client over time, and we end up with a client who only does well with us. It might feel very nice to be that important, but it's actually not so good for the client. So over time we have to make sure that that psychotherapeutic process has some impact on some psychological process; some mechanism in the patient. There has to be some mechanism there. And that's a good question: what that mechanism is, what they are. I would say there's quite a few possibilities there. But the ability to, whatever the therapy is, whatever version of CBT it might be, I think we could agree, correct me if I'm wrong, that ultimately the patient has to adopt some aspects of that and do that for him or herself. Ellis called that the ‘grand change’, not the little change. I could change this thought and that's better, but now I can observe my own thinking and change my own thinking. Beck does the same thing. Skinner talked about people being able to. We're free when we understand the contingencies that govern our behaviour, and we control our own contingencies. This is the same thing; whether you're talking about the ‘B’ or the ‘C’. So I think that that's a good question. Because I think there are lots of different mechanisms that collaborative empiricism can affect. Ultimately if we don't perfect those mechanisms inside the client's repertoire, it's kind of a false outcome.
Art Freeman
Is it ‘okay’ for the therapist to be this collaborative, empirically based, therapist who uses the Socratic dialogue as a major technique, using well-placed questions? Does collaborative empiricism also include ‘didactic teaching’? Saying: ‘Here's what I think you need to do.’ Not saying: ‘What do you think you need to do?’ Saying: ‘You know, it seems you've been trying this a lot. Do you think it would really help you if you . . .’ There's this whole idea: follow the patient, they know where they're going. If they knew where they were going, they wouldn't be going to therapy. I really think that when I'm in my office with a patient, one of us should have an idea where we're going, and I don't think it's always the patient, so I just want to make sure I'm doing the right thing after all these years: That it's OK for me to be more directive. To say: ‘You got to stop that. That's crazy.’
Mervin Smucker
Well, I want to make a point. I think your work with DBT [referring to Alan's comments], let's not forget who DBT was developed for. I think Marsha Linehan has been a major contribution to the field by developing a treatment for these, pretty much low-functioning borderline patients who weren't doing well with traditional CBT. But that's just one piece of what we're discussing here today. We're also talking about other clinical disorders as well. I think when you're looking at working with low-functioning borderline patients, of course you're going to be more directive.
Well, they come in, and they want something to change in their lives, and they're not sure what. So you're going to help them through your questions, and through your didactic piece, you're going to help them understand what it is. A lot of it is going to be this emotional dysregulation kind of thing. They want to have perhaps less conflict. You help them through your questioning, through closed-ended questions perhaps, rather than open-ended questions, to help them define more what their goals are. But they do want some things to change, they want less conflict in their lives, they want things to calm down some. It's that population that you're probably predominately working with and I think that we need to understand that with DBT it might look a little bit different with some of the other disorders.
I want to pick up on something you said, Art. Again, 25 years ago, Beck said, and I quote: ‘Cognitive therapy is 10% psychoeducational, where you're given the patients information, and 90% asking questions’ in the Socratic form that he advocated. So yes, there is a certain amount of information dissemination we're giving, but we are constantly asking: ‘Well, now what do you think of what I just said? How might that fit, or not fit in your situation?’ So you're right back into the collaborative piece even when you're giving information.
Nikolaos Kazantzis
May I just check that I've understood a theme here, and please let me know if I have mistaken this — are we talking about that the collaborative component is clearly different from agreement, and that collaboration is something that might be tailored for the individual client? Are we also saying that the empirical component to the relationship is something that might remain constant, that it's something that is always present. Have I heard that correctly?
Alan Fruzzetti
I would think so, yes.
Nikolaos Kazantzis
Thank you.
Concluding Comments
It has been an interesting learning experience to participate in the panel discussion and producing this article on the topic of collaborative empiricism. When we first considered the idea for the panel we mainly had questions, and were not sure where the conversation might lead. In reaching the conclusion, it is clear that the expert panel has offered some novel and interesting ideas, and have connected the discussion about this central feature of the therapeutic relationship with broader relational processes in CBT. In addition, the panel offered guidance for clinical practice. Art, Alan, Jackie, and Merv are thanked for kindly contributing their expertise for this purpose. Of course, the article would not exist without their ideas and participation, and it is hoped that it proves useful to the readership to have their ideas on collaborative empiricism presented here.
