What determines the learned predictiveness effect? Separating cue-outcome correlation from choice relevance
- Jessica Lee, University of New South Wales, Sydney, NSW, Australia
- Justine Greenaway, School of Psychology , University of Sydney, Sydney, Australia
- Evan Livesey, School of Psychology, University of Sydney, Sydney, New South Wales, Australia
AbstractEvidence from a variety of learning tasks suggests that cues that are more predictive of an outcome attract greater attention and are learned about more effectively in subsequent tasks. We tested whether this learned predictiveness effect is due to the objective strength of the cue-outcome association (cue-outcome correlation), or the degree to which the cue is informative for making the correct choice on each trial (choice relevance), by manipulating the possible outcome choices available on each trial. Experiment 1 compared two sets of cues that were equally (and imperfectly) correlated with outcomes and showed learning biases in favor of the set of cues that had initially been more relevant for choices made on each trial. Experiment 2 used a more conventional learned predictiveness design in which the cue-outcome correlation was stronger for one set of cues (perfect predictors) than the other set (imperfect predictors). However, here we manipulated whether or not the imperfect predictors could be used to make a correct choice, and thus whether the imperfect predictors possessed choice relevance that was equal to or less than the perfect predictors. In this case, we found no evidence that the relevance manipulation made any difference; learning biases towards the perfect predictor were evident regardless. The results suggest that both cue-outcome correlation and choice relevance can lead to changes in associability and learning biases; both were individually sufficient but neither were necessary.