Many subfields of psychological science have recently converged on the idea that relational reasoning skills are foundational to human cognitive abilities. This critically implies that training relational responding might improve a range of cognitive skills. Indeed, a relational training tool known as SMART has recently shown substantial promise; increasing IQ by 10-15 points across several studies, as well as preliminary evidence showing far transfer to tasks such as reading ability. If these preliminary findings are accurate, this would have important practical implications and provide new theoretical insights into human cognition. As such, SMART warrants a systematic investigation. In Work Package 1, we will examine the alleged theoretical importance of relational reasoning skills in greater depth by rigorously testing SMART with schoolchildren using state-of-the-art experimental design and examining its impact along a range of effects from near- to far-transfer. In Work Package 2, we will conduct a series of experimental investigations aimed at unpacking the nature of relational responding more generally, particularly with regard to the structure of relational responding and its relation to various aspects of cognitive performance. Finally, Work Package 3 will attempt to integrate idiographic methods to relational training, and use these insights to forward both theory and practice of relational responding.