Two experiments are reported which examine the relationship between a person's estimate of the likelihood of a future change in his environment and his assessment of its desirability. The first experiment showed a general tendency for probability and desirability ratings to be positively correlated. This correlation was higher when the desirability of a predicted change was seen, on average, as matching its probability of occurrence, e.g., if the change was seen as both desirable and probable rather than probable but undesirable. In the second experiment subjects wrote an essay arguing either that a predicted change was probable, improbable, desirable or undesirable. Arguing for desirability of the predicted change had as much effect on subjects' probability estimates as arguing for its probability; similarly, arguing for its probability had as much effect on desirability ratings as arguing for its desirability. These results are taken to imply that individuals may seek to achieve greater cognitive simplicity by treating probability and desirability as a single dimension.