Effects of biased extrapolation on attitude extremity

Previous research has shown that merely thinking about an attitude object can make both positive and negative attitudes more extreme. The present research explored whether a specific type of thought—extrapolating from known to unknown attributes—might make attitudes and behavioural intentions more extreme than reviewing known attributes. In three experiments, extrapolating from a social group's ‘known’ personality traits made positive and negative attitudes and behavioural intentions more extreme than reviewing those traits. This pattern of results occurred whether participants self-generated extrapolations (Experiment 1) or rated the likelihood of frequently extrapolated traits (Experiment 2). Attitudes were also more extreme after extrapolating traits high versus low in cognitive relevance to known traits, regardless of trait positivity/negativity (Experiment 3). In all three experiments, the effect of biased trait extrapolation on attitude extremity was mediated by more extreme associations with the extrapolated group. The current findings are consistent with attitude theories that emphasize cognitive processes.