Psychophysics is a field of experimental psychology that uses specific behavioral methods to determine the relationship between the physical world and people's subjective experience of that world. Psychophysicists conduct scientific experiments that are carefully designed to let them figure out which physical parameter(s) actually determine a subjective perceptual dimension. We begin by trying to think of all possible parameters; then we design experiments to quantitatively evaluate the most likely contenders under carefully controlled conditions.
Psychophysical methods allow us to ask of the haptic system: how, and how well, do people sense and resolve intensive, spatial and temporal variations in mechanical (and thermal) inputs to the skin (the "cutaneous" system) and to muscles, tendons and joints (the "kinesthetic" system)? For example, we are able to determine the minimal intensity required to just detect the presence of, say, a static force, a vibration, a gap between two points, or a cool surface, applied to the skin, or the smallest movement of the hand that can just be detected. We can also determine the minimum change in stimulus intensity that is required to just notice a change in one's perception. If we perform the same experiments with other sensory systems (e.g., haptics, vision, audition) , we may then compare how well each can sense minimum energy levels and resolve fine details.

But psychophysicists don't use only physical values that produce threshold-level responses. We also consider how people's sensations and perceptions grow in magnitude as a function of increasing physical intensity. With psychophysical methods, we have learned, for example, that when judging the roughness of metal gratings with different spatial periods, your perceptions turn out to be most strongly determined, not by the spatial period of the grating or by the width of the raised elements, but by the width of the spaces between these elements. Psychophysicists are very interested in the nature of the mathematical function that best describes the relation between physical and perceptual parameters. Continuing with our example, the results of our psychophysical experiments indicate that perceived roughness grows as a power function of interelement spacing, with an exponent of around 1.0 - 1.6. In fact, a power function seems to best describe the rate of growth of all sensations and pereptions, with the specific value of the exponent depending on several factors, such as the task and sensory system involved.

I have noticed that in the engineering and computing science fields, people sometimes use the term psychophysics inappropriately, by intending it to include results obtained with any behavioral technique. Actually, psychophysics provides a specific set of behavioral paradigms for addressing the types of physical parameters that I outlined above, namely sensitivity, resolving power, and the rate of growth of sensations. Behavioral techniques other than psychophysical ones can prove valuable as well in uncovering other important facts we need to know about human haptic processing. For example, by visually observing how people explore objects with their hands, Roberta Klatzky and I found striking evidence that people use highly stereotypical hand movement patterns ("exploratory procedures") during haptic search, selecting one or more that are specifically adept at extracting the particular property(ies) of interest, e.g. repetitive back-and-forth shearing motions for extracting texture information.

Psychophysical research can perform two general functions for those of you who actually design haptic interfaces. First, you can use the scientific information to guide initial design considerations. For example, the psychophysical results can be used to select appropriate physical parameters and associated values for your interface system. Results from basic research on human haptics highlight both the strengths and the limitations of using the haptic system to operate a sensory interface. Such information allows you to match the critical input/output parameters that underly human processing to the specific demands imposed by some haptic interface system. Therefore, to avoid subsequent design problems early consultation with a psychophysicist will prove very helpful. Some critical questions are: which type(s) of haptic information can or should be delivered? what are appropriate site(s) of delivery? and which are the best ways to display haptic information? While most designers of haptic interfaces have initially been forced to deal with relatively low-level sensory and motor issues, they should also be aware of additional cognitive influences (e.g., context, previous training, past knowledge, etc.) that may alter operator performance.

The second function provided by psychophysics and, more generally, by the scientific study of biological touch is the set of formal experimental tools that are available for evaluating how well operators perform with the haptic interface. Some important issues to think about relate to making sure you develop an adequately broad and appropriate set of performance tests for assessing your system, and to gathering data in a scientifically appropriate way to ensure the validity and generality of your conclusions.

Selected References

On psychophysical methodologies:
Gescheider, G. (1976). Psychophysics: method and theory. Hillsdale, NJ:Erlbaum.

A mini-tutorial on the scientific method:
Lederman, S.J. & Pawluk, D. (1992). Lessons from biological touch for robotic tactile sensing. In H. Nicholls (Ed.). Advanced tactile sensing for robots. World Scientific Series in Robotics and Automated Systems. Vol. V. World Scientific Publishing: Singapore.

Recent work on haptic perception and manipulation:
Lederman, S.J. & Klatzky, R.L. (1996). Haptic aspects of motor control. To appear in Boller & Grafman (series eds.); M. Jeannerod (section ed.). Handbook of Neuropsychology. Volume 11: Action and cognition. Amsterdam: Elsevier Science Publishers

A recent discussion of psychophysics as related to haptic interfaces:
Durlach, N. & Mavor, A. (Eds.). (1994). Virtual reality: Scientific and Tehnological challenges. National Research Council, National Academy Press, Washington, DC. Chapter 4 (Haptic Interfaces by M. Srinivasan).