The Evidence
The research behind this site.
Each finding is stated in a line, with the study it comes from.
The therapist and the method
Differences between legitimate treatment methods account for about 1 percent of client outcomes (Wampold and Imel, 2015).
The individual therapist accounts for about 5 percent on average: about 3 percent in controlled trials and 7 percent in everyday practice (Baldwin and Imel, 2013).
Across 119 therapists and 10,786 patients, recovery rates ranged from 23.5 to 95.6 percent, and the gap widened with more distressed clients (Saxon and Barkham, 2012).
What does not explain it
Following a treatment protocol more closely does not predict better outcomes (Webb, DeRubeis and Barber, 2010).
Over as long as eighteen years, the outcomes of 170 therapists did not improve with clinical experience, and on average slipped slightly (Goldberg and colleagues, 2016).
What does
Therapists who form strong alliances across their whole caseload get better outcomes. Differences between one therapist's clients do not predict outcome (Baldwin, Wampold and Imel, 2007).
Empathy, positive regard, genuineness, and the repair of ruptures each predict outcome across many studies (Norcross and Lambert, 2018).
Therapists who manage their own reactions to clients get better results (Hayes, Gelso and colleagues, 2018).
Clients did best with therapists who doubted their work while holding warmth toward themselves as people (Nissen-Lie and colleagues, 2017).
Ratings of how therapists respond to video of hostile, withdrawn, or self-blaming clients predict their clients' outcomes (Anderson and colleagues, 2009, 2016).
Therapists whom observers rated as more skilled with other people in a group discussion went on to get better client outcomes. Ratings from a one-to-one interview showed no such link (Schöttke and colleagues, 2017).
Why we cannot see it in ourselves
Of 129 clinicians, the average rated their own skill at the 80th percentile, a quarter at the 90th or above, and none below average (Walfish and colleagues, 2012).
Of 550 clients, 40 got worse during therapy. Their therapists predicted three, and were right about one (Hannan and colleagues, 2005).
How a group is led
In a study of 209 participants in 18 encounter groups, the best and safest outcomes came from leaders who were caring and who helped members make sense of what happened. Harm was concentrated under aggressive, charismatic leaders (Lieberman, Yalom and Miles, 1973).
Full references for these studies →
How a group builds these capacities →