Printer Friendly

The Potential for Misinterpretation Considered More Globally: A Response to Vicente and Ethier.

The notion that one can capture some aspect of reality by making a description of it using a symbol ... seems to me a fascinating and powerful idea. But ... there is a trade-off; any particular representation makes certain information explicit at the expense of information that is pushed into the background and may be quite hard to recover.

Marr, 1982, page 21


We begin by considering the study by Bennett and Malek (this issue) in a broader context. Static mimic displays provide a domain perspective that is critical: the layout of system components, physical connections, and sensors. Existing display technology provides the ability to represent the flow of resources through a system dynamically. Not surprisingly, designers have attempted to exploit this capability (we have seen numerous examples in research laboratories as well as in the literature). We initiated a research program on animated mimic displays because we believed that these efforts had achieved only limited degrees of success and that a scientific approach would be required to produce a more effective format.

The approach taken in this study has a parallel in the work performed by Cleveland and his colleagues (e.g., Cleveland, 1985), addressing the design of static graphs. They emphasize that the design of a graph involves making choices about the specific graphical features that are used to encode quantitative information. These choices, in turn, will require observers to perform specific elementary graphical-perception tasks to decode that information. From this perspective, effective display design involves choosing graphical features that can be decoded most effectively.

The research in question has focused on a fundamental issue in the design of animated mimic displays: the evaluation of a number of graphical features (i.e., luminance and chromatic contrast, geometrical contours and borders, spatial and temporal frequency, and waveform) to assess their utility for encoding quantitative information. The graphical features responsible for more effective decoding (i.e., those producing effective apparent motion) were identified; the findings have been summarized in a set of design guidelines. Failures in visual decoding (i.e., bidirectional apparent motion) were also discovered and the graphical features responsible were identified. From the perspective of cognitive systems engineering (Rasmussen, Pjetersen, & Goodstein, 1994), these evaluations fall at the Boundary 1 level (controlled mental processes). We view this level of evaluation as necessary but insufficient. Evaluations at higher boundaries are needed as well (e.g., fault management).

Vicente and Ethier (this issue) discuss a second set of fundamental issues that involve the mapping between these visual features and the domain properties to be represented. In Bennett and Malek, the term velocity was used to refer to physical properties of the display: the degrees of visual angle that a wave form moves during a unit of time (degrees/second). We do not use this term to refer to physical properties of the domain: how fast a fluid is flowing through a pipe (feet/second). Sensors in thermodynamic systems typically measure the volume of fluid that passes a point in a pipe during a unit of time (e.g., gallons/minute). When we discuss this sensor value we use the terms flow rate, flow, and rate of flow.


Vicente and Ethier (this issue) state that animated mimic displays have the potential to be misleading when used to represent flow rate, and they describe a fault scenario (pipe leak/break downstream of the flow sensor) to illustrate this point. They also provide an alternative mimic display (Vicente & Ethier, this issue, Figure lc) that they contend will be less misleading. However, their potential for misinterpretation in this scenario applies to any representation, not simply the animated mimic that we studied.

A higher-level property of a thermodynamic system is the process of material delivery: a flow of material originates at a source, follows a path, and terminates at a sink. Flow rate, measured at some point along the flow path, constitutes one piece of evidence regarding the state of the delivery process. Vicente and Ethier are correct in pointing out that operators could misinterpret the flow rate sensed at a point as the equivalent of the state of the delivery process. They are also correct in pointing out that there is a larger class of situations when operators can mistakenly assume a single sensor value stands for the state of a higher-order process.

However, this vulnerability is a concern for the design of any representation for these kinds of processes. For example, the issues Vicente and Ethier raise have been examined in the context of designing configural and other displays for computerized nuclear power control rooms (e.g., Roth, Mumaw, & Lewis, 1994; Roth, Woods, & Pople, 1992). In the context of design, versions of this fault scenario were simulated, and it was observed that operators were vulnerable to the misinterpretation raised by Vicente and Ethier when they were using conventional control room displays. These conventional displays are designed with the same single-sensor, single-display design philosophy apparent in the computer-based mimic format they propose (Vicente & Ethier, Figure lc).

To avoid the potential misinterpretation, operators needed to consider other evidence about the state of the delivery process (e.g., the difference in pressure between the path and the sink will reveal whether or not material is reaching the desired sink in some thermodynamic systems). Thus the scenario in question raises issues about how to integrate and display multiple uncertain sources of evidence into representations of the state of a dynamic process.


How designers represent the process can affect the potential for operators to misinterpret a single sensor value as standing for the state of a higher-order process. This is simply another example of the representation effect (see the opening quote from Marr, 1982; also see Zhang & Norman, 1994). Previous design work to take this class of scenarios into account has focused on (a) assessing the varying uncertainties in the multiple sources of evidence regarding the state of the delivery process and (b) how to combine or juxtapose these multiple sources of evidence in configural or other displays as an alternative to automated assessment by intelligent systems (Woods, 1991).

Simply resorting to a single-sensor, single-display format (Vicente & Ethier, this issue, Figure lc) or any other graphic form in itself does not avoid the vulnerability. The issue is not selecting one format or another but, rather, how to combine or juxtapose multiple sources of evidence, how to develop effective representations, and how to coordinate multiple representations in the process of representation design (Woods, 1995).

Vicente and Ethier's real claim is that mapping a flow sensor value onto an animated mimic display will increase the potential for this misinterpretation. We agree that this is a possibility, because an animated mimic display may function as an integrated or a configural display. To produce the animation, the display manipulates visual elements distributed over space. When designers choose to encode flow rate as measured at a point as an animation, the larger area of visual display that results could be mistakenly decoded as a representation of the delivery process, when it actually only encodes flow rate at a point. The class of scenarios Vicente and Ethier point out may serve as an excellent model preparation for studying issues in representation design based on their differential impact on this potential misinterpretation.

Rather than providing reasons to reject the animated mimic format, Vicente and Ethier have highlighted a need for research with a broader scope that complements the Bennett and Malek research on graphical perception of animated graphics. First, the purpose of the Bennett and Malek study was to assess the potential for mixed design at the level of graphical perception (decoding the display of data), whereas Vicente and Ethier raise a broader question at a different level of analysis. Second, representation design is always a process of orchestrating representational resources to meet challenges such as the potential misinterpretation raised in their commentary. Furthermore, all representations, in making some aspects of the world more salient than others, have the potential to be misleading. Progress depends on understanding the factors and situations that can render representations misleading and developing techniques to orchestrate different kinds of representations to balance the trade-offs.


The authors thank Mitch Wolff for his discussions, as well as Deborah Mitta for her encouragement and support.

Kevin B. Bennett is an associate professor of psychology at Wright State University. He received his Ph.D. in applied experimental psychology from the Catholic University of America in 1984.

David A. Malek is a human factors analyst at Science Applications International Corp. in Dayton, Ohio. He received his M.S. in human factors psychology from Wright State University in 1996.

David D. Woods is a professor at The Institute for Ergonomics at Ohio State University. He received his Ph.D. in cognitive psychology from Purdue University in 1979.


Bennett, K. B., & Malek, D. A. (2000). Evaluation of alternative waveforms for animated mimic displays. Human Factors, 42, 432-450.

Cleveland, W. S. (1985). The elements of graphing data. Belmont, CA: Wadsworth.

Marr, D. (1982). Vision. San Francisco, CA: W. H. Freeman.

Rasmussen, J., Pejtersen, A. M., & Goodstein, L. P. (1994). Cognitive systems engineering. New York: Wiley.

Roth, E. M., Mumaw, R. J., & Lewis, P. M. (1994). An empirical investigation of operator performance in cognitively demanding simulated emergencies (NUREG/CR-6208). Washington, DC: U.S. Nuclear Regulatory Commission.

Roth, E. M., Woods, D. D., & Pople, H. E., Jr. (1992). Cognitive simulation as a tool for cognitive task analysis. Ergonomics, 35, 1163-1198

Vicente, K. J., & Ethier, C. R. (2000). Why fluid dynamics matters for display design in process control: Commentary on Bennett and Malek (this issue). Human Factors, 42, 451-454.

Woods, D. D. (1991). The cognitive engineering of problem representations. In G. R. S. Weir and J. L. Ally (Eds.), Human-computer interaction and complex systems (pp. 169-188). London: Academic.

Woods, D. D. (1995). Towards a theoretical base for representation design in the computer medium: Ecological perception and aiding human cognition. In J. Flach, P. Hancock, J. Caird, & K. Vicente (Eds.) An ecological approach to human-machine systems I: A global perspective. Mahwah, NJ: Erlbaum.

Zhang, J. & Norman, D. A. (1994). Representations in distributed cognitive tasks. Cognitive Science, 18, 87-122.
COPYRIGHT 2000 Human Factors and Ergonomics Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bennett, Kevin B.; Malek, David A.; Woods, David D.
Publication:Human Factors
Date:Sep 22, 2000
Previous Article:Why Fluid Dynamics Matters for Display Design in Process Control: Commentary on Bennett and Malek.
Next Article:Postural Instability and Motion Sickness in a Fixed-Base Flight Simulator.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |