This article presents the main results of the First Experimental Study of Transference (FEST). The background of the study, as well as the aims, main hypotheses, and methods are described. The participants were 100 patients who were randomized to psychodynamic psychotherapy of one year's duration, with transference work or without transference work. The analyses of the effects of psychodynamic psychotherapy with transference work versus psychodynamic psychotherapy without transference work are presented. The two treatments were equally effective, but analyses of moderators revealed differential effects. Patients with low quality of object relations (QOR) and/or presence of personality disorder showed specific positive effects of transference work. Female patients responded better than men. Further analyses included mechanisms of change and three-way interactions among factors, with an impact on outcome. We also found that (1) insight was a mediator of change, and (2) that the specific effects of transference work were influenced by interaction of object relations and alliance, but in the direct opposite direction of what is generally maintained in mainstream clinical theory. For patients with more mature object relations and high alliance, a negative effect of transference work was observed.