My current research interests are listed below.
GENERALIZATION IN ASSOCIATIVE LEARNING
I am currently working on a project investigating the role of inductive reasoning processes in how people generalize simple learned associations. I use computer-based causal learning tasks, fear conditioning tasks with electric shock, and computational modelling in my research. Within this project, there are several streams:
- Peak shift and rules – in this stream we explore whether the peak shift effect (traditionally explained in terms of associative excitation and inhibition) can be explained by individual differences in generalization rules.
- Negative evidence – in this stream we examine the degree to which people treat a CS- (stimulus that does not lead to an outcome) as “negative evidence” in highlighting relevant category boundaries or dimensions.
- Diversity effects – in this stream we test whether more diverse evidence (i.e., range of stimuli which lead to an outcome) increases generalization to novel members of the same category, and whether these diversity effects also occur for negative evidence (stimuli which do not lead to an outcome).
One of the most contentious issues in learning research is whether it is possible to learn without awareness; in other words, whether implicit learning is possible. I am interested in the relationship between learning and awareness, and how the conditions under which we learn (e.g., incidental vs. intentional) affect measures that are thought to be largely implicit in nature.
CUE COMPETITION EFFECTS
Cue competition describes a class of effects in associative learning where multiple stimuli compete for associative strength with an outcome. Humans show many of the same cue competition effects (e.g., blocking, conditioned inhibition) displayed in animal studies, but there are often large effects of manipulations such as time pressure and causal assumptions. I am interested in exploring the factors that influence the degree to which cue competition effects occur, and whether these findings can be accommodated by purely associative models or whether more complex reasoning processes are involved.