A model of the uncertainty effects in choice reaction time that includes a major contribution from effector selection.

Abstract

Hick's law describes the relation between choice reaction time and the number of stimulusresponse alternatives (NA). For over half a century, this uncertainty effect has been ascribed primarily to the time taken to map a stimulus to its associated response. Here, data from two experiments suggests that selection of the appropriate effector-the particular body part to make a response-also contributes substantially to the uncertainty effect. This insight is important both for our understanding of basic cognitive architecture and because many classic experiments studying stimulus-response mapping have confounded NA with the number of effectors. Our data also suggest that, when stimuli are spatial and linked to the responses in an intuitively simple layout, the time required for stimulus-response mapping depends minimally on the NA, independent of effector. Experiment 1 demonstrated that in order to account for the complex patterns of uncertainty effects observed when stimulus type (spatial versus symbolic), response mode (typing, with multiple effectors versus touching with a single, known effector), and participant population (skilled versus novice typists) are all manipulated a model is required that includes effector selection, along with stimulus-response mapping, and a proper treatment of stimulus-response repetitions. Using spatial indicator stimuli that minimized the contributions of stimulus-response mapping, Experiment 2 compared four effector conditions-the factorial combination of one or three fingers on one or both hands. The results showed that the increase in the uncertainty effect associated with the number of effectors is negatively accelerated and possibly additive across the variation of hands and fingers.

How Relevant is Hick's Law for HCI?

“…A recent study by Wright et al [51] also demonstrates that when stimuli are spatial and linked to the responses in an intuitively simple layout, reaction time minimally depends on the number of alternatives. In a mode-switching study using the non-preferred hand, Ruiz et al [40] report a logarithmic relation between the number of modes and reaction time but do not find a significant effect of the number of modes on reaction time.…”

Hick’s law equivalent for reaction time to individual stimuli

“…The utility of such an analogy is debatable and even Hick’s law itself has been the subject of some challenges over the years (see, for example, Laming, 1968, Ch 1; Luce, 1986, Ch. 10; Proctor & Schneider, 2018; Wright, Marino, Chubb, & Mann, 2019). Of course, as in the physical sciences, laws are not necessarily exact under all circumstances.…”

Visual selection and response selection without effector selection in tasks with circular arrays

“…In their words, "A unidimensional response code utilizes a single effector member and a specified number of points along one of the physical dimensions of a particular response continuum, such as the direction, the force, or the duration of a movement" (p. 200). As emphasized recently by Wright, Marino, Chubb, and Mann (2019), such situations allow isolation of response-selection processes at a level of spatial locations without including the level of effector selection, as required for key-press responses. Because Wright et al (2019) have made a strong case that effector selection is a distinct process that may affect results obtained in choice-reaction tasks, we focus mainly on tasks where selection of an effector is not an issue.…”

“…As emphasized recently by Wright, Marino, Chubb, and Mann (2019), such situations allow isolation of response-selection processes at a level of spatial locations without including the level of effector selection, as required for key-press responses. Because Wright et al (2019) have made a strong case that effector selection is a distinct process that may affect results obtained in choice-reaction tasks, we focus mainly on tasks where selection of an effector is not an issue. A reason for doing so is that any phenomenon evident in such tasks cannot be attributed to effector selection.…”

“…A study by Wright et al (2019) used logic similar to that of Fitts and Deininger (1954), but with the semicircular eightlocation target array. They assigned letters (or symbols) to the target locations corresponding to those on which the left and right hands would be placed when typing: A, S, D, and F to the left hand and J, K, L, and ; to the right hand, going in order from the leftmost to rightmost target locations.…”