|
Article Excerpt
Tasks demanding vigilance or sustained attention require observers to monitor displays for the appearance of infrequent signals for action over prolonged periods of time. The central issue, in almost 60 years of vigilance research, has been the decline in performance with time on watch termed the vigilance decrement (See, Howe, Warm, & Dember, 1995). Although vigilance has been and remains an important part of many contemporary work environments, it has assumed significantly greater importance with the introduction of many 9/11-mandated security systems such as by the Transportation Security Administration's (TSA) need to screen 100% of all baggage at all US airports (Harris, 2002 see also Hancock & Hart, 2002). The advent of these, and many similar, security requirements has increased the need for effective training procedures for system operators. In vigilance, training has traditionally employed feedback in the form of knowledge of results (KR) regarding an operator's accuracy expressed in terms of correct detections, false alarms, or missed signals. KR has been shown to improve performance on sustained attention tasks (for a review, see Warm & Jerison, 1984). However, there is experimental evidence that the form of KR can impact its effectiveness. For example, Dittmar, Warm, and Dember (1985) asked participants to monitor increments in a vertical line presented on a video display terminal (VDT) and provided observers with feedback, in the form of a 1,000 Hz tone, regarding their correct detections, false alarms or missed signals. They observed that hit-and false alarm-KR enhanced perceptual sensitivity (A') and attenuated the vigilance decrement, but miss-KR had no such effects. They explained these findings in terms of the information provided by each KR format. Hit- and false alarm-KR provide information regarding responses to specific stimuli but miss-KR is not linked to the observer's noticing of any particular characteristic of a signal event or to the observer's subsequent response.
Although Dittmar and her colleagues (1985) reported performance effects of KR format, they did not examine the stress effects associated with the use of KR. Several experiments have reported that vigilance tasks are stressful. Observers report feeling more boredom, strain, irritation, distress and less task engaged after the task than prior to its start (e.g. see Szalma et al., 2004; Warm, 1993). In addition, research has indicated that manipulations of task characteristics that influence performance also influence individual subjective state (e.g. Szalma et al., 2004; Temple et al., 2000; see also Warm, Dember, & Hancock, 1996). Therefore, it is feasible that the format in which KR is provided to observers can influence their stress state in addition to influencing their performance. Given that Dittmar et al. observed performance differences as a function of KR format, and the associations between performance and subjective state (Warm et al., 1996), one might expect that hit- and false alarm-KR would be associated with reduced stress symptoms and that miss-KR would be associated with greater stress, relative to a no-KR control condition. That is, the effect of KR format on stress should mirror the effects previously reported by Dittmar and her colleagues with regard to performance. Previous studies using a composite format have shown that used together the three forms of KR enhance performance (e.g. Szalma, Hancock, Warm, Dember, & Parsons, in press; Szalma, Miller, Hitchcock, Warm, & Dember, 1998), but these studies did not examine the effects of composite KR on observers' stress states. It may be that providing all three forms of KR will show the greatest performance enhancement and reduction in stress, since all information regarding detection accuracy is explicitly provided. One goal for the current experiment was to test these propositions.
Individual differences in performance and stress response: Feedback intervention theory (FIT)
From existing evidence, it appears highly likely that the effects of KR format on performance and stress associated with vigilance will depend upon the characteristics of the individual. Indeed, it has been long recognized that general individual differences among observers can influence performance in vigilance (see Berch & Kanter, 1984; Davies & Parasuraman, 1982). Recent evidence has confirmed that responses to the stress of sustained attention also depend on characteristics of the individual observers (e.g. Helton, Dember, Warm, & Matthews, 1999; Szalma, 2002a). Such findings are consistent with transactional models of stress that emphasize the importance of appraisal processes in stress response (Lazarus & Folkman, 1984; Matthews, 2001).
An individual differences variable that has recently emerged as an influential factor on stress responses to sustained attention is dispositional pessimism/optimism. Helton et al. (1999) investigated the effect of false failure feedback on the performance and stress of both pessimists and optimists in a vigilance task. Although false failure feedback did not significantly impact performance, individuals high in pessimism and low in optimism exhibited steeper vigilance decrements across conditions relative to individuals low in pessimism and high in optimism. Helton and his colleagues also observed that pessimistic individuals reported changes in stress-related parameters such as greater tension, lower hedonic tone (i.e. a more negative mood state) and less confidence relative to optimists. These findings suggest that the high stress associated with vigilance may be moderated by an observer's relative level of pessimism/optimism. Further evidence for the influence of pessimism and optimism on observers' subjective stress states was reported by Szalma (2002a) who found that higher degrees of pessimism were associated with greater distress and lower task engagement after a vigilance task. Szalma also reported that pessimism predicted greater reliance on emotion-focused coping, consistent with prior research indicating use of this coping strategy by pessimistic individuals. Optimism, in contrast, was less predictive of stress and coping.
With regard to individual differences in the effectiveness of KR for vigilance, predictions can be made based upon a theory of feedback intervention (FIT), derived from a meta-analytic study of the feedback intervention literature (Kluger & DeNisi, 1996). FIT serves as a general framework for understanding the processes that influence feedback effectiveness. Drawing on control theory (see Carver & Scheier, 1981, 1998), FIT assumes that behaviour is regulated by comparison of feedback to goals or standards that are hierarchically organized. Two further assumptions of this theory are that (1) only those feedback interventions that receive attention can influence performance, and (2) feedback interventions have the capacity to alter the locus of attention within the goal hierarchy. Although hierarchies can vary in complexity, Kluger and DeNisi describe three general levels relevant for feedback effects. At the lowest level are task-learning processes consisting of 'hypotheses' for improving performance that individuals test by utilizing feedback. Attention to this level of the hierarchy constitutes attention to the details of focal task performance. These processes are regulated by the next highest level in the hierarchy, task-motivation processes, which serve as standards or comparators for the task-learning processes. At the highest level are meta-task processes involving the self and general but highly salient goals. Specifically, meta-task processes relate task outcomes to higher-level goals that are important to the individual. Kluger and DeNisi argued that most affective reactions to feedback interventions, when they occur, result from evaluation of feedback with respect to salient self-goals specified at the meta-task level.
According to FIT, feedback is effective in improving performance only to the extent that it serves to focus attention on the task (i.e. task-motivation and task-learning processes). Thus, an FIT interpretation of the results of Dittmar and her colleagues (1985) would be that the hit- and false alarm-KR directed attention to task-learning processes (e.g. learning the nature of a signal/non-signal) and perhaps task-motivational processes (e.g. see Warm, Kanfer, Kuwada, & Clark, 1972; Warm, Epps, & Ferguson, 1974), while miss-KR failed to adequately direct attention to relevant task characteristics. Miss-KR may fail to direct attention appropriately because it is not linked to a specific, overt response from the observer, so that observers are unable to use the feedback to associate their response to the stimulus characteristics (see Szalma et al., in press). Moreover, one of the propositions of FIT (see Proposition 5, Kluger & DeNisi, 1996, p. 269) is that feedback cues that match salient personal goals associated with a given personality trait will direct attention to meta-task processes and thereby debilitate performance. This may subsequently induce stress effects in such individuals. Kluger and DeNisi suggested a general mechanism for these effects whereby individuals' traits influence their ability to redirect their attention to the details of the task in response to feedback. They proposed that this occurs via an inhibition mechanism that protects working memory space from competing demands imposed by the goal hierarchy, and that traits such as the tendency to blame oneself for failures may exert debilitative effects on feedback by interfering with this inhibition mechanism. Hence, individual differences relevant to an observer's response to feedback may influence the effectiveness of training with feedback interventions such as KR.
Given the propositions of Kluger and DeNisi (1996), traits such as pessimism and optimism, which affect general expectancies regarding performance and affective responses to performance success and failure (see Chang,...
|
