Traditional economic models state that riskier investments should have a higher expected return. Psychological models of choice showed that people is influenced by the kind of information they are provided with and by the context of the choice, since they do not have a stable order of preferences but, rather, they construct it while giving a judgment. Experiment 1 showed that people consider the expected return of an equity only when it is explicitly stated, rather than computing it on the basis of the expected outcomes and probabilities. Experiment 2 and 3 showed that people judgments about the expected return of a stock are influenced by the way they are required to provide their estimates. In Experiment 2 participants' judgments were inconsistent with the risk-return trade off, whereas in Experiment 3 people asked a higher return from a riskier stock than from a safer one. Implications of the results for investors' behavior are discussed.