Increased technology reliance along with today’s global fast paced society has produced increasingly complex, dynamic operating environments in disciplines as diverse as the military, healthcare, and transportation. These complex human machine systems often place additional cognitive and metacognitive demands on the operator. Thus, there is a crucial need to develop training tools for all levels of operators in these dynamic systems. The current study was designed to empirically test the effects of four training methods on performance and mental model accuracy in a microworld simulation game. It was hypothesized that process-focused guidance targeting metacognitive level processes as well as combined process and problem focused guidance would result in better performance and mental model accuracy than problemfocused guidance alone or unguided training approaches. Additionally, it was expected that individual differences in prior decision making ability, metacognitive awareness, working memory span, and fluid intelligence would moderate the relationship between the type of instructional guidance and outcomes. Results supported the development of decision-making skills through process-focused instructional guidance, particularly for initially low performing or more novice individuals. Results highlight the importance of individual learner experience prior to training. Similarly, this research aims to expand the literature by providing support for process-focused training as a method to support non-expert decision making skills. While further research needs are outlined, the current research represents an important step forward in both the theoretical literature providing support for instruction designed to support domain general decision making skills in non-experts. iv Practical implications regarding improved guidance for future instructional and training systems design, personnel selection, operator and system performance evaluation, and safety are also discussed.