Humans have limited capacities in understanding complex problems with at least one variable developing over time. For instance, adult participants perform poorly in predicting exponential growth (e.g., Wagenaar & Sagaria, 1975). And more recently, students from highly ranked technical colleges in the US demonstrated difficulties to estimate the temporal development of one variable determined by two given variables (e.g., Cronin, Gonzalez, & Sterman, 2009). In this research we will investigate medical students’ performance for solving temporal complex tasks that are central to medical education. Managing temporal complexity is relevant whenever treatments affect patients over a period of time. For instance for patients in intensive care, fluids might be administered and urination monitored for keeping the level of body fluids in the patient constant. Small deviations from that level might not be critical for one point in time. But accumulating over time they could have fatal consequences. Right now, there is little evidence on medical students’ understanding of temporal complexity: In a pilot study (Brunstein, Gonzalez, & Kanter, submitted), medical students performed better than undergraduates for estimating temporal developments for the fluid scenario above and some related tasks. However, the evidence from that study was limited in several ways: First, medical students excelled for some, but not all scenarios used. Second, the range of medical experience of students within medical school was very limited. Third, filling in a questionnaire on physiological phenomena is associated, but not central to medical education. In this research, we will build on the Brunstein et al.’s pilot to fix methodical issues and to systematically extend our understanding of medical students’ performance in temporally complex scenarios. We will recruit pre-medical and medical students from Weill Cornell and present them with a task central to their medical education. They will estimate and manage temporal complexity for this task. First, we will conduct a web survey on the estimation task with a new medical scenario and the original non-medical scenario inviting pre-medical and medical students to participate. We will need to recruit about 60 Weill Cornell (30 pre-medicine and 30 last year medical) students for about 5 min of their time. Half of each group will perform the original non-medical estimation task. The other half will perform on the newly developed medical estimation task. Based on our experience, this survey will take about 14 days to be conducted with the majority of responses within the first hours after sending out an email invitation. As dependent variable, we will collect the number of correct responses for each group for a Chi Square analysis. The whole study should take about 1.5 months to be conducted. Second, we will expand on solving temporally complex problem by asking students to manage temporal complexity for a simulated scenario. Meanwhile, there exist several great simulation tools that allow to train and test students under both, realistic as well as laboratory accurate and controlled conditions. In our Pittsburgh lab, we collected experience with MEDIC, a microworld in the context of medical diagnosis, to study probability learning. If medical students perform better in the first study for estimating temporal developments for the medical scenario than pre-medical students, we would expect them also to perform better when actually managing temporal complexity. If they don’t excel on the artificial estimation task, they still might perform better in a realistic task providing us with guidance on how to interpret the related system dynamics literature. We will invite about 60 participants (30 pre-medical students and 30 last year medical students) into the lab for about one hour to perform the task in the simulation. Half of each group will perform the relevant and realistic medical task and the other half will perform a parallel version of a non-medical task, for instance managing the number of patrons in a shopping mall during the holiday season. We will measure their performance for both tasks in terms of deviations from the desired optimum level over time and we will analyze those scores using a MANOVA for repeated measures. Including the development of the simulation, this study will take about 6 months to be conducted. We expect this research to answer for the first time the question whether medical students can rely on their medical education for managing temporal complex medical scenarios based on empirical evidence. This will contribute to expertise research and inform about needs in medical education.