Approach-avoidance conflict in rats was investigated as a function of physical confinement, number of approach-training trials, similarity of cues and number of shocks. Physical confinement had no effect upon rate of acquisition of approach behavior but did influence penetration to the goal on the conflict trials. Similarity of cues between the training and testing situation was important in determining the extent to which the animal approached the goal during conflict trials; the greater the dissimilarity of cues, the nearer the animals would approach the goal. Approach was also found to be a decreasing function of number of shock trials.