This study aimed to better understand how metacognitive skills develop in young children aged 5 to 7 years. In particular, we addressed whether developmental changes reflect quantitative or qualitative improvements, and how metacognitive skills change with age and task-specific ability. Previous research into the development of metacognitive skills has been somewhat limited by methodology--often there is an over-reliance on language skills and it is assumed that children are fully conscious of the skills they use. In this study, a new observational method was developed which aimed to better represent young children's (n = 66) metacognitive skills by coding their verbalizations and non-verbal behavior during a problem-solving task. This method proved to be developmentally sensitive and illustrated both a quantitative increase in metacognitive skills, and qualitative changes in the types of monitoring and planning used throughout early development. Further, the results indicated that monitoring processes improve with age, control processes improve with both age and task-specific ability, and "failures of metacognitive skills" are primarily affected by task-specific ability rather than age.