Highly Predictable Outcomes of Danger Facilitate the Interpretation Bias in Anxiety

Abstract

We examine a) whether texts about predictive inferences that contain a high level of contextual constraints, that is, a high likelihood of dangerous outcomes (89%), facilitate drawing danger-related predictive inferences in high anxiety individuals; this would be a sign of a danger inferential bias, and b) whether such context constraints speed up the time course of predictive inferences and are produced automatically. Participants with different levels of trait anxiety –high vs. low- read predictive contexts, which are danger-related and non-danger-related, and from which either a highly likely outcome or a non-predictable one can be inferred. In Experiment 1, the high-anxiety group drew more danger-related inferences than non-danger-ones, and the low-anxiety group made more non-danger-related inferences. In Experiment 2, with a 500ms SOA interval between the inducing context of the inference and the target word confirming or disconfirming the predictable event, all participants made danger- related inferences. Results demonstrate that high contextual constraints help high-anxiety individuals focus their resources on the most dangerous outcomes rather than on other harmless information and the former is then selected when answering. This shows an inferential bias linked to anxiety. However, although such constraints speed up the time course of predictive inferences, they do not produce an automatic bias.