The relationship between blocking and inference in causal learning

The blocking effect in causal learning, once taken as a hallmark of associative learning, has recently been explained in terms of an explicit deductive reasoning process. Yet when the conditions necessary for deduction are removed, a small blocking effect is often still present. We examined the relationship between blocking and participants’ performance on analytical thinking and probabilistic reasoning measures. Inferential processes predict blocking or an absence of blocking in this situation, depending on the observer’s consideration of conditional probabilities. Although Bayesian inference predicts blocking, most individuals are not inclined to use this form of probabilistic reasoning explicitly, an observation we confirmed using a logical problem with similar properties to the relationships present in the blocking effect. Furthermore, participants who showed the greatest capacity for analytical reflection were less likely to show a blocking effect, suggesting that blocking in causal learning is the product of an intuitive and unreflective thought process.