Abstract
Decision making is a complex behavior that all people engage in, both in everyday situations and in professional or academic contexts. Understanding how individuals make decisions under uncertainty has important implications for learning, training, and performance. This study examined how people learn probabilistic cue-outcome relationships using the Weather Prediction Task (WPT). This task expresses and captures the Multiple Cue Probability Learning (MCPL) paradigm that models real-world decision environments. The cognitive processes of interest were fluid intelligence (Gf) and working memory capacity (WMC), with consideration of metacognition as an additional factor. Prior research has not fully clarified how these cognitive processes contribute to MCPL success and has not included metacognition as a factor for consideration. Participants completed the WPT along with individual difference measures. Results showed that participants improved across the WPT and showed high performance in the final test block. A working memory capacity measure called symmetry span was the strongest predictor of the final test block. A working memory capacity measure called operation span and a task-specific metacognitive measure predicted performance in the initial learning block. General self-regulated learning measures were not significantly related to WPT performance. Findings suggest that learning in this environment may depend on different predictors at different phases of the task. They also suggest that task-specific metacognitive measures may be more useful than general self-regulation measures in this context.