Executive Summary: Coaching needs robust quantitative outcome studies and meaningful coaching effectiveness measures (De Haan & Duckworth, 2012). Coaching effectiveness research needs to develop rigor and status similar to other helping professions (Passmore & Theeboom, 2015) to sustain coaching credibility (Gray, 2011). Currently, most coaching effectiveness studies present design limitations that impact upon conclusions, for example: self-reported measures; lack of random allocation; data being collected in a pre- and post-coaching design (or at three data points). The pre/post designs, which are currently a predominant approach to assessing change over time in the coaching research, can only be used to model linear change, and tend to show relatively small changes over time, as most of the before and after intervention studies.

The aim of our RCT study was to address the above issues in order to contribute to the development of a comprehensive coaching outcome effectiveness model. In our experiment participants were randomly allocated to coached (experimental) and not-coached (control) groups (over 100 participants in each group). A coached group received coaching over a period of 6 months from experienced coaches accredited by a UK-based Business School. We collected data over eight data points for all participants, enabling us to model the dose-curve type change expected over the course of, and following the coaching sessions. Responses were collected from both students and coaches, in order to overcome issues of purely self reported measures.

We examined the shape of change, and variation in the shape of change for coaching outcomes, specifically goal attainment, perceived stress, psychological wellbeing and resilience. Controlling for common factors - quality of coaching relationship, self-efficacy, outcome expectancy, hope and perceived social support — we tested a hypothesised model in which the positive effects of coaching upon change in goal attainment, perceived stress, psychological wellbeing were mediated by the increased resilience that coaching promotes. For the group who received coaching we also examined the effect of working alliance on our outcomes.

Our study confirms that coaching increases resilience, psychological well-being, goal attainment and decreases perceived stress. This increase was found to be linear whilst coaching was being undertaken, but was not extended beyond the coaching period — however there was also little loss in the benefits accrued from coaching after it ceased. The positive effects of coaching on changes in psychological well-being, goal attainment and perceived stress, were found to operate indirectly via changes in resilience.

We found perceived social support to be the only common factor that was positively related to changes in well-being. Despite the consensus in the literature about the working alliance to be the best predictor of outcomes, our results only partially support the initial level of working alliance as being a predictor of changes in outcomes.

Our findings may be of importance when demonstrating value of coaching to stakeholders. As coaching outcomes are mediated via resilience we suggest that coaches focus on developing resilience in their coaching sessions. We have also found changes in outcomes — due to coaching — to be still mostly in place three months after completing coaching. The shape of change for well-being, goal attainment and resilience suggests that there are still therapeutic benefits when the number of sessions is higher than six. Our working alliance findings indicate the focus on coachee’s goals to be of particular importance. Finally, we suggest that future coaching research pays more attention to the role of external factors in contributing to successful coaching outcomes.