The information gain of explicitly provided over self-generated contextual knowledge for behavioral control

Abstract

In time-pressured decisions, humans exploit contextual knowledge to reduce uncertainty about the unfolding situation and to improve behavioral control. However, in complex real-world settings, it remains unclear whether the explicit provision of contextual information is beneficial or not. We thus examined the information gain of explicitly provided information as a function of expertise, information uncertainty and acquisition phase. To this end, we measured the positioning of female handball players (N = 36 experts + 36 near-experts) in a virtual-reality defensive task as a function of their teammates’ defensive-strength patterns, which was either explicitly instructed or had to be self-generated. Furthermore, the certainty of provided information was experimentally varied (67% vs. 83% consistent information). All eight groups–expertise (2) x acquisition condition (2) x information certainty (2)–improved performance in terms of the positional difference in their defense movements, meaning that they either moved more sideways to support their neighboring teammate or remained more often in their position when no support was required. However, an explicit-knowledge test showed no differences regarding pattern detection between the acquisition conditions, implying that the performance enhancement of the self-generated groups was not due to explicit-knowledge accumulation. Most notably, experts generally benefitted from explicit instructions whereas for near-experts, an information gain could only be revealed for comparably certain information. This interaction implies that future research on explicit provision vs. self-generation of contextual knowledge should pursue a more differential approach, thereby also considering gender and age as well as personality factors.