How Many Resources and How to Identify Them? Commentary on Boles et al. and Vidulich and Tsang

Abstract

One goal of human factors is to identify theory-based mechanisms of human performance that can account for meaningful performance differences in real-world tasks and settings. Boles, Bursk, Phillips, and Perdelwitz (2007, this issue) set out to do this by demonstrating that mechanisms of multiple resources, as assessed by a method different from the more conventional multiple resource model with which I have been associated (Wickens, 1980, 1984, 2002), can account for differences in dual-task interference between complex tasks that can generalize to real world tasks. (In the preceding characterization of their research, I focus only on Experiment 2, for reasons I will describe later.) Their research provides data that appear to support this generalization. Vidulich and Tsang (2007, this issue) offer a number of criticisms of the research, with two in particular focusing on (a) the fact that sequential processing of two tasks, resulting because of widely separated displays and discrete responses, may limit the contributions of multiple resources to time-sharing in the paradigm chosen; and (b) concerns about the structure of the three tasks in Experiment 2 ‐ one pair mandating more continuous demands and therefore concurrent processing, and the other two pairs probably prohibiting it. They suggest that this feature could readily account for the observed data, thereby eliminating the need to postulate multiple resource mechanisms. Boles and Phillips (2007, this issue) in turn argue in defense that sequential processing is common in most time-sharing applications and that asynchronous processing “was considered to more closely mimic dual tasks as used in the real world” (p. 51). Although both of Vidulich and Tsang’s (2007) concerns have some merit, I also believe that the data are consistent with a multiple resource interpretation, and so I will not reemphasize their concerns. Rather, I wish to highlight the importance of six theoretical and methodological issues of applied attention research, which the current paper brings to the forefront. In doing so, I hope to extend the message of the experimental research reported in ways that the authors may not have highlighted. Issue 1. Total demand versus resource similarity. In both my older (Sarno & Wickens, 1995;