Using an individual differences perspective to investigate context effects in surveys: A test of the belief-sampling model

Survey methodologists know that responses to a particular survey question can be affected by responses to previous questions; these effects are termed 'context effects'. For example, reported attitudes towards the testing of products on animals will be generally more positive if previous questions have highlighted the products as medicines rather than cosmetics. In recent years, under the banner of the Cognitive Aspects of Survey Methodology (CASM) movement, survey methodologists have borrowed concepts and techniques from cognitive psychology to develop models of attitude formation to explain how context effects occur. This study tests one model of attitude formation, the belief-sampling model developed by Tourangeau and colleagues. Much of the literature on context effects takes an aggregate approach, looking at effects across the whole population. This study takes a more disaggregated approach by investigating effects across different personality types. It is argued that the belief-sampling model predicts that individuals will differ in their susceptibility to context effects. These differences are dependent on how thoughtful the individual is and how consistent their thoughts are towards an issue. The belief-sampling model is tested by comparing context effects across four groups of individuals who differ in their propensity to think and in their propensity to evaluate. It finds general support for the model but suggests that propensity to evaluate may be the more important personality dimension in predicting context effects.