Several decision-making models predict that it should be possible to affect real binary choices by manipulating the relative amount of visual attention that decision-makers pay to the two alternatives. We present the results of three behavioral experiments testing this prediction. Visual attention is controlled by manipulating the amount of time subjects fixate on the two items. The manipulation has a differential impact on appetitive and aversive items. Appetitive items are 6 to 11% more likely to be chosen in the long fixation condition. In contrast, aversive items are 7% less likely to be chosen in the long fixation condition. The effect is present for primary goods, such as foods, and for higher-order durable goods, such as posters.