Throughput (TP) is a global measure of input device efficiency but provides little information about user's movement behavior when interacting with a device. Psychomotor models of movement provide a framework from which to develop new “during” movement variables that can be used to explain why efficiency differences occur. Data from a previous study examining the usability of a mouse, trackball and RollerMouse™ was re-examined using TP and the kinematic variables peak velocity of the primary movement (PV) and proportion of total distance traveled in primary phase (%PMD). Partial correlation analysis found %PMD and PV to be significantly related to TP and negatively related with each other, suggesting a “speed/accuracy” tradeoff. Further analysis confirmed the variables were useful in discriminating between devices and found that the most efficient device was less constrained by the “speed/accuracy” tradeoff. It was concluded that kinematic variables offer a useful way of understanding efficiency differences between devices.