| Abstract: | The Pollyanna hypothesis is extended into the field of intertrait inference to predict that inferential thresholds for positively evaluated characteristics will be lower than those for negatively evaluated characteristics. This prediction is confirmed, and is shown to have important implications for models of inference rules. Data from several studies are analysed to reveal that the Pollyanna threshold effect is reliably greater for women than for men, and it is shown that this effect is unrelated to sex differences in extremity of responding. |
