Contextual cuing as a form of nonconscious learning: Theoretical and empirical analysis in large and very large samples

Numerous studies have demonstrated that associative learning can affect visual cognition. In one such effect, search times for a target hidden among similar distractors are faster for repeated search configurations compared with novel configurations. This contextual cuing effect is particularly interesting, because researchers routinely have failed to find evidence of recognition of the repeated configurations, concluding that the effect is a form of nonconscious learning. Vadillo, Konstantinidis, and Shanks (2016) recently criticized this conclusion on a number of methodological and conceptual grounds that suggest the area suffers from a high probability of false-negative results on awareness tests and misinterpretation of weak or absent relationships between cuing and awareness measures. We developed further predictions from theoretical models assuming that single or independent memory sources drive learning and awareness and discuss how these predictions fare in three new contextual cuing experiments involving large (n > 60) and very large samples (n > 600). The data support the absence of a positive relationship between recognition and the cuing effect both at the participant and configuration level, the probability of which being a false negative is very low in a model assuming a single memory source drives learning and awareness. This was the case using both conventional and Bayesian analyses. The combination of this theoretical and empirical analysis suggests that contextual cuing is not dependent on cue recognition and provides evidence that it reflects a genuine form of nonconscious learning.