Why are difficult figural matrices hard to solve? The role of selective encoding and working memory capacity

Abstract It is well documented that figural matrices tests are harder to solve when multiple rules need to be induced because multiple rules are traditionally associated with a greater demand for dynamically managed sub-goals (goal management), which requires more working memory capacity (WMC). The current research addresses the necessity to apply selective encoding as a requirement that goes beyond the ability to manage goals when solving figural matrices. In the first study ( N  = 38), we found that selective encoding demands are present in items with multiple rules in addition to goal management demands. Furthermore, eye movement data indicated that rule induction was hampered when selective encoding demands were present. The second study ( N  = 127) demonstrated that individuals' ability to filter relevant features in working memory was positively related to figural matrices items with selective encoding demands. Moreover, there was no evidence that other sources of WMC are related to goal management in figural matrices. Hence, this study provides preliminary evidence that filtering of relevant information in working memory is critical for solving figural matrices with multiple rules and challenges the view that goal management is the only driver of the relationship between WMC and performance in solving figural matrices with multiple rules.