Measuring Coordination in 2D Positioning Tasks Sriram Subramanian, Dzmitry Aliakseyeu and Jean-Bernard Martens User Centered Engineering group Faculty of Industrial Design, Eindhoven University of Technology s.subramanian,d.aliakseyeu,j.b.o.s.martens@tue.nl Abstract: Several measures for coordination have been introduced recently in the human-factors literature. These measures try to capture the quality of the path followed during a two-dimensional (2D) or three- dimensional (3D) positioning task. Most of these measures are derived in a fairly ad hoc way, rather than being based on sound theoretical concepts. After analysing some of the existing measures, we formulate a list of requirements that seem essential for any measure of coordination. Based on these requirements we propose a new measure for coordination that abides by these requirements. We subsequently describe an experiment that was performed to gather subjective impressions of coordination for a number of 2D paths and compare the existing and newly proposed coordination measures with the obtained subjective rankings. Keywords: Coordination, interaction techniques, allocation of control, evaluation methods, Human-Computer Interaction. 1 Introduction Recent years have seen the proliferation of a wide variety of input devices into the users workspace. These devices range from 2D mice to many different kinds of 3D input devices. Given this growing diversity, it becomes increasingly more difficult to distinguish between these input devices solely based on traditional performance measures such as task completion time and accuracy (error). The problem is that these measures do not tell us much about how people actually handle devices and why performance differs across devices (MacKenzie, 2001). Some people propose that a more informative measure would be one that captures the quality of the trajectories executed by the user while performing positioning tasks with the input device (Zhai 1998; Masliah, 2000; Masliah, 2001). Coordination is a term usually used for a measure that conveys this information (Zhai 1998). The term coordination is used with the same intuitive meaning in disciplines as diverse as medical diagnosis and rehabilitation and athletics (Fischer, 1997). Despite this widespread interest in coordination, a broadly accepted definition of the term is still elusive (Kondraske, 2000). While different measures for coordination have been introduced (Zhai 1998, Masliah, 2000; Fischer, 1997; Kondraske, 2000), they are mostly formulated in an ad hoc way, and are not based on any computational theory. Thus a measurement procedure for coordination that agrees with our intuitive understanding, and that is based on firm theoretical concepts, would be most useful in this context. A starting point for developing such a procedure is to analyze existing measures. 2 Existing Measures Kondraske and Vasta (Kondraske, 2000) have proposed a Neuromotor Channel Capacity measure (NMCC) that is based on Fitts’ law. Fitts’ law expresses movement time MT in terms of the ratio A/W of target distance A over target size W in a 1D selection task as The ‘index of performance’ IP=1/b (in bits/sec) is claimed to be a useful measure of the clinically more familiar term “Coordination”. In an earlier paper, Fischer and Kondraske (Fischer, 1997) have proposed an alternative quality measure that can be derived from the 3D trajectory       + = W A b a MT 2 log 2 Human-Computer Interaction -- INTERACT'03 M. Rauterberg et al. (Eds.) Published by IOS Press, (c) IFIP, 2003, pp. 295-302