Selective attention to predictive cues is often considered an efficient way to address the exploration-exploitation dilemma in decision-making. Yet in some circumstances, it can also lead to sub-optimal decision-making due to false beliefs about the environment acquired early in learning - a learning trap. In this study, we examined the relationship between attention selectivity and the emergence of a one-dimensional learning trap in a multidimensional categorization learning task. Combining empirical work (N=75) and computational modeling, we find that more selective attention, especially in the early phase of learning, increases the likelihood that an individual will fall into a learning trap. This finding sheds light on the causal role of attentional biases in the way that individuals explore and learn about choice-options.

Authors:

Yanjun Liu: University of New South Wales; Ben Newell: University of New South Wales; Jaimie E Lee: University of New South Wales ; Brett Hayes: University of New South Wales