Back to Wayne Gray's home page / To Research Interests Index / To Downloadable Pubs / The CogWorks Lab / Cognitive Science Department
Gray, W. D. (2002). Simulated task environments: The role of high-fidelity simulations, scaled worlds, synthetic environments, and microworlds in basic and applied cognitive research. Cognitive Science Quarterly, 2(2), 205-227.
In field research, there is often too much [complexity] to allow for any more definite conclusions, and in laboratory research, there is usually too little complexity to allow for any interesting conclusions (Brehmer & Dûrner, 1993 p. 172).
Studying interactive behavior in an ETA framework is difficult to do. It requires tasks of some complexity else the interactions of embodied cognition with task and artifact are stilted. Yet it requires a situation in which action protocols can be collected and timed stamped to at least 16.67 msec accuracy. Such work requires working in a real task environment (as in the Gray and Anderson, 1987 work) or a simulated task environment (Gray, manuscript submitted). Simulated task environments attempt to address the researcher's dilemma raised by Brehmer and Dûrner by balancing complexity against tractability, correspondence, and engagement.
A simulated task environment is tractable only in relationship to a given set of research issues. For data collection a tractable simulated task environment must allow the researcher to collect the right data, at the right grain size, with the right timestamp. For example, in Argus Prime the right data and grain-size include every mouse click made by the subject, every system response, every mouse movement, and point-of-gaze. For mouse clicks, point-of-gaze, and mouse movements the timestamp is 16.667 msec (60 hz).
Few, if any, simulated task environments are built as an end unto themselves. Most are intended to produce results that will generalize to some real task environment. High correspondence simulated task environments simulate many aspects of one task environment. Low correspondence simulated task environments simulate one aspect of many task environments.
Finally, a simulated task environment is engaging to the degree to which it involves and affects the subjects; that is, the degree to which they take it seriously. Engagement describes something about the subjectÍs motivation. Subjects may be engaged because we are paying them money to do well. They may be engaged because they view the simulated task environment as an interesting game that they like to play. Or they may be engaged because they have deep knowledge of the real-world task, believe that it is interesting and important, and are able to fill in the blanks that are missing from the simulation.
Much of my work has involved building simulated task environment. The early work with the COBOL tutor required that we simulate a structured editor. This was a painful process as rather than having a structured editor with which the tutoring system interacted, we had to essentially build our own structure editor into the code for the tutoring system. The first thing I did as an academic was to build a simulated task environment of a VCR with which both my subjects and ACT-R models could interact. I subsequently built the Ned simulated task environment for Project Nemo (Ehret et al., 2000) that we used to collect data from 30 U. S. Navy submarine officers (and are now using to develop ACT-R models of this task). My lab has also built a VCR construction kit in which we can construct a range of interfaces from perverse to optimal. We have also constructed a blocks world environment (Fu & Gray, 2000). Of course, the most complex and complete simulated task environment we have is the Argus suite -- Argus Prime and Team Argus (Schoelles & Gray, 2001 & Schoelles & Gray, 2000).
I find the building of simulated task environment to be a productive methodology. However, when the research question demands it, my lab is willing and able to fall back to more traditional experimental psychology paradigms (as illustrated by my work on serial attention, see Altmann & Gray, 2000).
Back to Wayne Gray's home page / To Research Interests Index / To Downloadable Pubs / The CogWorks Lab / Cognitive Science Department
Last changed: 2004-03-01 wdg