Decision speed induces context effects in choice

The context in which a decision occurs can influence the decision-making process in many ways. In the laboratory, this is often evident in the effects of recent decisions. For instance, many experiments combine easy and difficult decisions, such as when word frequency is manipulated in lexical decision. The "blocking effect" describes how such decisions differ depending on whether the conditions are presented in pure blocks (comprised purely of easy or hard stimuli) or mixed blocks (also known as a "mixing cost"). We present a novel extension to these context effects, demonstrating in two experiments that they can be induced using conditions with identical difficulty, but different timing properties. This suggests that explanations of context effects based on task difficulty or error monitoring alone might be insufficient, and suggest a role for decision time. In prior work, we suggested such a hypothesis under the assumption that observers minimize their decision time, subject to an accuracy constraint. Consistent with this explanation, we find that decisions in slower conditions were based on less evidence when they were experienced in mixed compared to pure blocks.