We know that the therapeutic alliance is critical to successful therapy. We all agree on that, don’t we? Beyond that, however, we seem to have a noticeable divergence in how we understand the alliance, and why it’s important.

The alliance isn’t a single phenomenon, but three interrelated ones (goals, tasks and bond) that sit at the heart of a collaborative partnership between therapist and client. Get these right and we have the potential for remarkable things to happen. Get them wrong, and we may have nothing more than an empty chair.

If you’re a therapist it’s hard to imagine that you won’t be familiar with the concept of the therapeutic alliance. If you’re anything like me you probably have a decent enough sense of it that you’ve never felt the need to question it. Maybe you too have sat in therapeutic huddles with colleagues talking over and agreeing its importance, assuming you were all talking about the same thing. But were you?

The alliance – many definitions lead to ‘creative ambiguity’

In the planning of this blog I ran an internet search for the term ‘therapeutic alliance’. I was immediately struck by the extraordinary inter-changeability of it, and similar terms. The Wikipedia page for Therapeutic Relationship, for example, starts:

‘The therapeutic relationship (also therapeutic alliance, the helping alliance, or the working alliance)….’

This inter-changeability of terms isn’t restricted to Wikipedia, it’s all over the internet, and the lack of a standard definition of the alliance has contributed to what Horvath et al  [i] describe as a state of ‘creative ambiguity’- the alliance is what I want it to mean, or understand it to be. What the terms describe is also enormously varied, encompassing trust, bond, respect, empathy, congruence, genuineness, unconditional positive regard, warmth and many others. Despite the wide range of descriptions, however, most refer broadly to the quality of the bond or relationship between therapist and client.

The working alliance is a collaboration for change for which I have identified three aspects: (1) mutual agreements and understandings regarding the goals sought in the change process; (2) the tasks of each of the partners; and (3) the bonds between the partners necessary to sustain the process.

Edward S Bordin: A Working Alliance Based Model of Supervision

That’s pretty much how I conceived the alliance, too, until I started to take a greater interest in the various elements of what makes for successful therapy. The alliance, however, according to most currently agreed definitions in the academic literature, goes well beyond the affectional bond between therapist and client, into the realm of the specific goals and tasks of the therapy process. In other words, it is not just about the ‘what we feel towards each other and how we express that’ but also the ‘what you want to achieve’ and ‘how we will get you there’ elements of our work with clients. More about these elements, and what research shows they contribute to the process of therapy, are the subject of the rest of this and the next blog.

Current definitions of the alliance

There’s nothing new about the concept of alliance. Freud referred to a positive ‘unobjectionable’ transference which bonds the client to the therapist sufficiently to collaborate in the process of unearthing and overcoming difficult or disturbing material. It wasn’t till 1956, however, that the term ‘alliance’ was coined by the American psychoanalyst Elisabeth Zetzel. [ii] Since then the concept has been further refined and expanded beyond the borders of psychoanalysis, and is now widely understood as a pan-theoretical concept. As such, it has become one of the most rigorously researched subjects in psychotherapy.

The most commonly recognised definition of the alliance today is based on that elaborated by Ed Bordin in the mid to late 1970’s. Bordin used the term ‘working alliance’, [iii]  and he conceived of this alliance being based on a collaborative stance in therapy, underpinned by three processes:

Agreement on therapeutic goals

Agreement on the tasks that make up therapy

The bond between therapist and client

Edward S. Bordin. 1913-1992

Bordin proposed that while the alliance was applicable to all therapies, that different therapies would naturally emphasise different aspects of the alliance.

Later, he would go on to propose that the strength of the alliance was not something static, but would fluctuate over the course of therapy, and that attending to stresses on the alliance would offer rich opportunities for therapeutic progress and client growth. In a way this seems like the early beginnings of what we now describe as the alliance rupture-repair cycle.

What accounts for the power of the therapeutic alliance?

I find it easier to answer this question by considering what leads clients not to engage in therapy or to dropout after a period. As a client, I need to have some realistic hope that you, the therapist, have a better idea than I do about how I can get to a better place than the one I currently occupy. If I don’t know what’s possible and have no goal or goals in mind, I have little to commit to or motivate me. Even if I can see where I want to get to, but don’t know how to achieve this, I need to hope that you might.

To have hope, I need to know where I’m headed with you, and what the journey will involve. It will also help if you appear warm and caring and genuinely want to help. Put all these together and I’m engaged. In the absence of one or more of these elements, however, that’s less likely.

By contrast, if you can help me to elaborate what I really want from therapy, that’s a good start. From there, if you can not only tell me that it’s achievable, but also how you can help me to get there, then you give me hope. Not only that, but I can start to have faith in you as someone who can help me. Then, if you have faith in me, seem to like and want the best for me, and are warm and encouraging, this helps me to trust you.

It is the goals that set the end point, the destination around which the tasks of therapy are initially organised. The tasks are the means that we will use; where we will put our focus and why; what therapeutic methods or models we’ll use; how we will monitor and review the goals and progress towards them, and revise these or our methods if we need to. A solid bond between us creates the sense of trust and safety for me, as client, to do the work.  

Echoing Carl Rogers core conditions, we can argue about whether a strong alliance is necessary and sufficient for change. It’s hard argue, however, that it’s not necessary. For me, a strong alliance is like the foundations of a house. Without it, it’s not going to be very stable.

What’s the link between alliance and client outcomes?

 Behind research into the effectiveness of therapy in general, and specific therapies in particular, the alliance is the most researched therapeutic factor in psychotherapy research literature.

The headline is that the alliance appears to contribute more to therapy outcomes than almost every other factor that has been rigourously researched save for two: empathy, and goal consensus and collaboration. The meta-analysis by Horvath et al (2011) [i] found an effect size of r = .275, that was found to be a ‘moderate but highly reliable relation between alliance and psychotherapy outcome’.

If effect sizes don’t mean much to you, this can be put another way by saying that the alliance accounts for about 7.5% of the variance in outcomes. That may not sound like much, but comparison with other factors shows its true significance. As outlined in Wampold (2015) [iv] empathy accounts for 9% of the variance, goal consensus and collaboration 11%, therapists between 3 – 7%, and specific treatment models less than 1%. Put another way, the alliance accounts for seven times more of the variance in outcomes than the therapy model.

The alliance isn’t a single phenomenon, but three interrelated ones: Agreement on therapeutic goals, agreement on the tasks that make up therapy, and the quality of the bond between therapist and client

The power of the alliance lies in the creation of hope within the client, faith in the therapist as someone who can help, commitment to the process of therapy, and expectation that the commitment will bring benefit.

The alliance has been shown to be a critical factor in therapy outcomes. In terms of its impact it accounts for seven times more of the variance in outcomes than the therapy model.

 What’s next?

 There’s so much more to say, but it must await the next blog! In brief, I’ll be exploring the following:


How early alliance strength predicts outcome and drop out


The relative contributions that client and therapist make to the alliance-outcome link and how, as therapists, we vary in our ability to form alliances.

Until the next time, and please do leave a comment below!

Read the second part of this blog series here


[i] Horvath, A. O., Del Re, A. C., Flückiger, C., & Symonds, D. 2011. Alliance in individual psychotherapy. Psychotherapy, 48, 9–16. http://dx.doi.org/10.1037/a0022186

[ii]  Zetzel E. R. (1956). Current concepts of transference. Int. J. Psychoanal. 37, 369–375

[iii] Bordin, E. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research, and Practice, 16, 252–260

[iv] B Wampold and Z Imel. 2015. The Great Psychotherapy Debate: The Evidence for What Makes Psychotherapy Work. Routledge, New York

Alliance word image attribution to Nick Youngson http://www.thebluediamondgallery.com/a/alliance.html

Share with your networks
Posted by:Barry McInnes

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.