Control of the therapeutic process by the therapist (or the patient) i
s a theoretically controversial concept, and even the empirical findin
gs are inconsistent. One reason is a difference in the way to understa
nd the construct control. A two-dimensional system to distinguish the
various variables of control is proposed. Three different operational
definitions of control, each differentiated for therapist and patient,
are compared within the framework of cognitive-behavioral treatment o
f 30 phobic patients. The three measures of control result in three co
rrespondingly different curves through the three stages of therapy. Th
ey do not correlate with each other and they have different relationsh
ips to the success of treatment. The impression of having control is a
n indicator of a global positive or negative appraisal of the therapy
and not an indicator of actual control.