A double approach-avoidance conflict model appears adequate to describe the results of Maier paradigm studies on response fixations in rats. The model states that "go-no go" conflict must occur in the presence of response alternatives-spatial (Left or Right) and nonspatial (Bright or Dark). Evidence is presented that shows that random 50% punishment itself does not produce fixation. Reinforcement by anxiety reduction and resistance to extinction due to partial reinforcement were found wanting as explanatory concepts. Rather, altering the balance of the "go-no go" conflict with psychoactive drugs, or eliminating response alternatives during random punishment, reduces the occurrence of fixations. Introducing novel cues related to reward and punishment but not associated with previous conflict, and techniques which gradually lead Ss into the conflict situation also reduce fixation occurrence.