The coaching relationship as ‘best predictor’? How does the working alliance help to achieve outcomes?

The coaching relationship in the form of the ‘working alliance’ between coach and coachee has been proposed as the best predictor of coaching outcomes and the foremost active ingredient that coach and coachee can influence and optimise during their work together. This chapter investigates that claim. Even though working alliance is indeed a very good predictor of overall outcome from coaching, we find little evidence that sessional changes are predicted by the working alliance from session to session. From two recent studies, which are the only ones to have measured the working alliance over time, from both a coach and a client perspective, we have to conclude that there is no correlation between ‘working alliance’ and ‘outcome increase per session.’ The same has been observed in several psychotherapy studies. If this is generally true then the working alliance is only related to overall outcome and may have much more to do with a response bias (through general trust or optimism) than with a proper driver or active ingredient of coaching outcomes.

A: some controversies

As we have seen, there is now a lot of evidence for the coaching relationship as the ‘container’ or a ‘vehicle’ for the coachee to achieve his or her goals and other results with coaching. This goes back to the original definition of coaching as offering a ‘vehicle’ (literally, a ‘coach’ - a word originating from the Hungarian village Kocs where very good coaches/carriages were once made) for learning and development. The coaching environment or relationship is then the direct surrounding of the coachee, where space is offered, conditions are peaceful and reflective and safe, so that the coachee can do his or her own work to forge ahead with the goals. The relational field is increasingly seen as the core active ingredient for coaching, particularly in the way that it is experienced by the client. To measure this ingredient we need to ask, is the relationship clear and helpful and conducive, is it pleasant, is it agreeable, wholesome, secure?

As mentioned before, the most widely used tool to measure the coaching relationship in coaching is the Working Alliance Inventory (WAI; Horvath & Greenberg, 1986) which consists of three subscales: Task, Goal, and Bond. The term Task refers to what coach and coachee agree needs to be done for the coachee to reach his or her goals for coaching. A typical item is, ‘I am clear as to what my coach wants me to do in these sessions.’ The term Goal refers to the outcomes that the coachee hopes to gain from coaching. A typical item is, ‘The goals of these sessions are important to me.’ The term Bond refers to what extent the coachee trusts, respects, and feels confidence in the other person. A typical item is, ‘I believe my coach is genuinely concerned for my welfare.’

A lot of the discussion around the coaching relationship has been driven by earlier and much more extensive work in psychotherapy outcome (McKenna & Davis, 2009). There is now a very strong evidence base in psychotherapy showing that working alliance as perceived by the patient is a very good predictor for outcome (5 = 0.6; p < 0.001; Fliickiger et al., 2020).

All of this started with a broad summary of the psychotherapy outcome literature that Michael Lambert made in 1992. In his summary which took the form of what became an influential pie chart (see Figure 3.1), he suggested that 40% of change through therapy would be ‘extra-therapeutic’ (i.e. due to support and life’s circumstances outside the therapy), 30% would be due to ‘common factors’ to all therapies, including the therapeutic relationship, 15% would be due to techniques, and 15% would be due to expectancy (the ‘placebo’ effect). An updated version of this pie chart, based on much more research since 1992, is displayed in Figure 3.2. It shows that the extra-therapeutic factors have now rightly been left out of the pie chart (although they are still considered to be the strongest influences on therapy even if they come from the outside) and the expectancy effect is not made explicit anymore, since it cannot be distinguished from other (personality) factors in the client, something we also saw in Chapter 2 (we cannot determine if expectancy, motivation, hope, and preparedness factors are part of the personality of the client but we can attribute them clearly to the client). In the new figure we can see that there is now broad consensus in psychotherapy research that the patient contribution is the largest, around 30%, that the relationship is the next largest single factor (12%), and that the therapist contribution (7%) is around the same size on average as the technique contribution (8%).

The Lambert pie chart from 1992 estimating the different active ingredients in therapy, with relationship ‘common factors’ on 30%

Figure 3.1 The Lambert pie chart from 1992 estimating the different active ingredients in therapy, with relationship ‘common factors’ on 30%.

Unexpected variance, 40%


Figure 3.2 The updated Lambert pie chart (Norcross, 2011) with the therapy relationship on 12%, which is the largest contribution to do with the therapy itself.

In this chapter we will take a closer look at the coaching relationship and what we now know about its importance for coaching outcome. But first let us look at some broader controversies amongst scientists which pertain to the coaching relationship measurement.

< Prev   CONTENTS   Source   Next >