Decision making deficits play an important role in the definition of pathological gambling (PG). However, only few empirical studies are available regarding decision making processes in PG. This study therefore compares decision making processes in PG and normal controls in detail using three decision making tasks examining general performance levels on these tasks as well as feedback processing using reaction time analyses. To investigate the specificity of decision making deficits in PG, a substance dependence group (alcohol dependence; AD) and an impulse control disordered group (Tourette syndrome; TS) were included. The PG group (n = 48), AD group (n = 46), TS group (n = 47), and a normal control (NC) group (n = 49) were administered (1) the Iowa Gambling Task (IGT), an ecologically valid gambling task; (2) the Card Playing Task, a task measuring perseveration for reward; and (3) a Go/No-Go discrimination task, a task measuring reward and response cost sensitivity. The PG group showed a diminished performance on all tasks and deficient feedback processing as compared to the NC group on the IGT and the Card Playing Task. In general, performance measures were not associated with levels of comorbidity or with self-reported motivational measures. For the larger part, deficiencies in decision making processes in the PG group were also present in the AD group, but not in the TS group. Subgroup analyses revealed larger decision making deficits in pathological slot machine gamblers than in pathological casino gamblers. Deficits in decision making and feedback processing in PG should be addressed in treatment and incorporated more explicitly in theoretical models of PG.