At first brush "mental improvement" is a massively broad topic covering development in everything from social cues to symphony composition. Existing literature normally focuses on development of specific skill, and attempts to be more general are limited by the narrow of range of lab-appropriate experiments. This projects proposes that a cross-sectional theory-driven analysis of improvement has significant predictive benefits despite the wide range of behaviours being explained. I will show that specific predictions in rate of improvement, speed, accuracy and generalizability follow naturally from considering improvement along two axes: process change and memory recall.