ABSTRACT
Air traffic controllers do a demanding job under dynamic conditions. Over the years, researchers have built performance models that often blur the lines between the contributions of the operator and those of the system itself. Performance definitions and measurement continue to challenge us. In ATC we have carried on the aviation tradition using checklists and rating scales. These tools have face validity but may or may not be reliable. High fidelity, person-in-the-loop simulation created a host of objective, easily recorded measures. This generates a high volume of data. However, the variables are often intercorrelated. Some variables do focus on safety, and in aviation, safety is everything. However, controllers are good at what they do, so systems errors and a loss of separation are relatively rare events. This means that generally errors are not effective performance indicators, especially for research, where the goal is to find subtle differences induced by changes in procedure or technology. It is likely that questions related to controller performance and how it may be influenced by system changes will be around for a very long time.
