PERCIEVED WORKLOAD UNDER DIFFERENT ATTENTION DEMAND DURING AN EXPERIMENTAL STUDY

Transcription

1 PERCIEVED WORKLOAD UNDER DIFFERENT ATTENTION DEMAND DURING AN EXPERIMENTAL STUDY Adel Mazloum 1, Hiroyuki Izumi 2, Yoshiyuki Higuchi 3, and Masaharu Kumashiro 4 1, 2, 4 Department of Ergonomics, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Japan 1-1, Iseigaoka, Yahatanishi-ku, Kitakyushu, Fukuoka, , Japan Corresponding author s amazlomi@med.uoeh-u.ac.jp 3 Department of Community and International Nursing, Faculty of Nursing, Fukuoka Prefectural University, Japan 4395, Oaza-Ita, Tagawa, Fukuoka, Japan Abstract: Attentional demands and individual s cognitive failure are hypothesized to be detrimental factors to workload assessment, although previous research has focused merely on the characteristics of one type of attention demand. The objectives of this study were to investigate the degree to which various attentional paradigms would be demanding to the participants, and the effect of their cognitive failure on this demands. A total of 24 participants completed two 15-min and one 60 min sessions representing three paradigms of attentional demand. Attentional paradigms were created as divided, selective, and sustained attention tasks and three groups of CFQ s were instructed to manipulate those tasks. Outcomes were measured in subjective assessment of workload, stress, arousal, and anxiety level, along with performance measures. In general, divided attention contributed to an increased level of workload with a higher level of arousal for high s, while sustained attention created an increased level of frustration with a decreased level of arousal. Different paradigms of attention demand induced diverse pattern of workload with a worse level in individuals with high CFQ scores Attention demand 1. INTRODUCTION One imperative aspect of the tasks in safety and health domain is Attention Demand. In safety and health domain the attempts are often toward investigating the root causes of occupational accidents, added mental/and or physical workload, fatigue, and so on in any dynamic and static work situations. Realizing the role of attention demand might be informative for such investigation. In other word, a critical aspect of designing systems for dynamic environments, such as driving or flying, is the amount of visual and cognitive attention required to complete a task. This required attention is termed "attention demand" by researchers and designers. For static environments such as desktop computing, attention demand is more important for usability issues than for safety issues. For some devices such as monitors and displays for operating systems, devices in vehicles, the safety aspects of attention demand are paramount; as, failing to adequately address attention demand issues in a desktop software program may lead to poor usability, user confusion, and loss of revenue (Monk et. al., 2000). Attentional resources can be thought of as a pool from which all tasks and mental activities are drawn. These attentional resources are used to safely perform cognitive tasks such as driving the vehicle, piloting task, process control, medical monitoring and so forth on. Different aspects of attention may be investigated with the three main paradigms (Matthews et. al. 2000): 1- Selective attention: The participants must respond to same stimuli, or stimulus properties, whilst ignoring others. 2- Divided attention: The participants must perform two (or more) tasks, simultaneously, which may lead to performance breakdown (dual-task interference).

2 3- Sustained attention: The participants must maintain the focus of attention over a relatively long timeperiod. Each of the above-mentioned attention demand tasks induces different level of mental workload Individual differences in error proneness and accident liability A multitude of studies have examined the cognitive aspects of accidents (Matthews et. al. 2000), with most of this research being focused on differentiating worker characteristic. Relatively recently, research has examined the relationship between the construct of cognitive failure and accident rates (Larson, et. al., 1997; Larson & Merritt, 1991). A cognitive failure has been describe as breakdown in cognitive functioning that results in a cognitively based mistake or error in task execution that a person should normally be capable of completing (Martin, 1983) with some people being more prone to experiencing cognitive failure than others (Wallace, et. al., 2002). Figure (1) shows a model of the impact of cognitive failure proneness on attentional focus and performance breakdowns and it was developed in this study on the bases of Reason s (1988) premise. As Reason further stated, people prone to cognitive failure might have a rigid attentional focus, thereby creating a cognitive management style might allow for the occurrence of cognitive breakdowns in dealing with intervening and concurrent stimuli. For instance, individuals with high level of cognitive failure may be relatively unaware of dynamic or novel conditions or that incoming information exceeds their attentional capacity while performing tasks. This might be especially true in the presence of task automization (Heckausen et. al., 1990). Additionally as mentioned before, accidents are commonly the result of faulty cognitive/attentional processes such as selective attention, mental errors, and distractibility. Therefore, it appears plausible that the occurrence of cognitive failures and unsafe behaviors and accidents may be significantly related (Wallace et. al., 2003). Attempts have made to develop self-report measures of information processing abilities, of which probably the best know is the 25-item cognitive failure questionnaire CFQ (Broadbent et. al., 1982). Some specific characteristics of the CFQ include: (I) the best known self-report measures of information processing abilities, (II) An index of the efficiency for distributing attention over multiple inputs under stressful conditions. (III) Predictive of accident involvement (Matthews et. al. 2000). In sum, considering attentional aspect of task demands and individual differences in cognitive failure, we attempted to apply three classes of workload measurement techniques, as proposed by O Donnell and Eggemeier (1986), includes: (I) performance-based assessment; (II) subjective workload assessment techniques; and (III) physiological workload assessment in this study (figure 1). Much of the previous studies related to attention demand have concentrated merely on a simple aspect of attention, either sustained, divided, or selective attention demand; while no/or little attention has been given to the comparative investigation of these demand. Yet, such a comparison would make a new prospective on how to deal with these demands in real-world tasks to minimize mental workload and to optimize task performance. To address the need for more empirical data in this concern, the current study was designed to examine the degree to which various attentional paradigms would be demanding to the participants. Furthermore, the influence of individuals cognitive failure on subjective and objective outcomes was probed while performing each task Design of experiment 2. Methods A 3 3 full factorial experiment was conducted to test the effects of three levels of cognitive failure and three levels of attentional demand on perceived workload as well as task performance variables. A total of 24 participants were conducted three kinds of different cognitive task within three groups of low, medium, and high s (8 people for each group). The order of three tasks was randomized between participants.

3 Cognitive Failure Low High Rigid attentional focus - Incapable of flexible attention in response to changing task demands -Incoming information exceed attentional capacity -more vulnerable to the effects of stress Different Attention Demand tasks Cognitive breakdown (impairment) Perceiving greater mental workload Higher Psycho-physiological variation Higher error commission (lapse and Figure 1: Model showing the impact of cognitive failure proneness on task performance, perceived workload and psychological functions. A 15-minute duration was allocated for each of the selective and divided attention tasks plus a 60-minute for sustained attention task. Between each trail a 5-minute break was given to the participants Independent variables Three kinds of attention demand were defined as independent variables: (I) Selective attention: The participants must respond to same stimuli, or stimulus properties, whilst ignoring others. (II) Divided attention: The participants must perform three tasks, simultaneously. (III) Sustained attention: The participants must maintain the focus of attention over one hour time-period. Moreover, CFQ scores which representing participants cognitive failure state was another independent variable in this study Dependent variables Performance-related variables In terms of sustained attention task these variables were: response time which was the time lapse between the first appearance of each target letter and clicking mouse by the participants. And, accuracy rate means the number of correct answers divided by the total number of answers was measured. The result is expressed as a percentage. For selective attention task two independent variables were defined: response time that is a time lapse between first appearance of the color or name stimulus and clicking of mouse. And, Accuracy rate: which is the number of correct answers divided by the total number of answers. The result is expressed as a percentage.

4 In case of divided attention task, reaction time for significant deviations in the four vertical gauges in the monitoring task, average of fuel levels of tanks A, B, C, and D (sum of these four tanks) in the fuel management task and RMSE ( root mean squared error) in the tracking task were recorded Subjective measures Subjective workload assessment A self-reported quantitative measure of mental workload, the NASA-TLX inventory (Hart, & Staveland, 1988; Hill, et. al., 1992) was applied for this assessment. It includes a scale consists of six subscales: (i) mental demand (ii) physical demand (iii) own performance (iv) temporal demand (v) effort, and (vi) frustration. They include visual analogue scales and participants are asked to mark each subscale on lines, on which are written low / high or good / bad on each end. These marked points are quantified from 0 to 100 for a raw score for each scale. Then the raw scores for workload on the six subscales are weighted according to paired comparisons and combined into a single measure called Weighted Workload (WWL) score. However, in this study for overall subjective workload assessment, the Adaptive Weighted Work Load (AWWL) score was used instead of WWL. Although the participants were undergraduate medical students, the process of multidimensional comparison required by the original NASA-TLX could have been somewhat cumbersome and complex for them. Therefore, the overall work load score was derived from the AWWL to simplify and facilitate the multidimensional rating procedure. Its validity as an alternative index of mental workload in younger adults has been reported in a previous study by Miyake and Kumashiro (1993) with a high correlation of the traditional WWL and the AWWL scores. Additionally, a simple summation of the scores was calculated as R-TLX too. In the abridged Japanese version of the NASA-TLX, subscale names and explanations in the original NASA-TLX, which are translated into Japanese, are used Evaluation of stress and arousal level: Stress Arousal Checklist (SCAL) developed by Mackay (1978) was utilized at prestudy and at the end of each trail for evaluation of stress and arousal level over experimental runs. It consists of the stress scale, reflecting individual perceived preference about physical and psychological conditions, and the arousal scale, reflecting physical activities, especially autonomic nervous activity. It has seventeen adjectives relating to perceived stress and thirteen relating to arousal, for a total of thirty questions, and operators respond to each adjective on a fourpoint analogue scale Evaluation of anxiety For this mean, the State-Trait Anxiety Inventory (STAI) was adopted, as it initially conceptualized as a research instrument for the study of anxiety in adults (Spielberger, 1983). It is a self-report assessment device which includes separate measures of state and trait anxiety. According to the author, state anxiety reflects a "transitory emotional state or condition of the human organism that is characterized by subjective, consciously perceived feelings of tension and apprehension, and heightened autonomic nervous system activity." State anxiety may fluctuate over time and can vary in intensity. In contrast, trait anxiety denotes "relatively stable individual differences in anxiety proneness" and refers to a general tendency to respond with anxiety to perceived threats in the environment. 2.4 Tasks Three kinds of attentional demand tasks were introduced to the participants in the current study SART (sustained attention response test): is designated as vigilance or sustained attention type of task (Appendix 1) including go and no-go Roman alphabets which the subjects should click upon appearing the letter p or q on display and no click for the letters (Robertson et. al., 1997) SCWT: (Stroop Color Word Test) this task is a psychological test of mental (attentional) vitality and flexibility (Appendix 2). The task takes advantage of our ability to read words more quickly and automatically

5 than we can name colors. The cognitive mechanism involved in this task is called selected attention (Matthews et. al., 2000). In this task, the participants have to manage their attention, inhibit or stop one response in order to say or do something else MATB: A revised version of the Multi-Attribute Task Battery (Comstock et. al., 1992) was used for divided attention demand (Appendix 3). This task consists of tracking (first order, one dimensional compensatory tracking task, (Appendix 3, top center), monitoring (Appendix 3, top left) and fuel management (Appendix 3, bottom center). The tracking task of the original version was a two-dimensional tracking task. Participants were instructed to keep fuel levels of four tanks above specific fuel levels (2000 for tanks A and B; 1000 for tanks C and D) in the fuel management task. This instruction was different from the original one in which the participants were required to maintain the levels of tanks A and B at the optimum level of 2500 units each. Other tasks were almost identical to those in the original version Participants In all, twenty four participants were recruited from the University of Occupational and Environmental Health, Japan. They were undergraduate male students with the mean age of (SD=1.16). They had no history of or current conditions that would affect their perceptual, cognitive, or motor functions during manipulation of the tasks in general and they were required to have normal or corrected-to-normal vision, have had sufficient sleep the night before the experiment and have abstained from drinking alcohol Design and Procedure A total of 24 students were screened among 120 volunteers who have initially responded to the first call of the study. After final selection, a written informed consent was obtained from all participants following a detailed briefing on the purpose of the experiment. Then, after completing demographic and medical questionnaire the participants were required to perform the tasks in a random base. Before each task they had a practice session till getting familiar to the task and adaptation to the experimental condition. At the end of practice session they required to complete SCAL and STAI, and then a 10 minutes break time were given to them before beginning the experimental session. Over the experimental session they were assigned to complete MATB and Stroop test for 15 minutes and SART task for 60 minutes. A five minute break was giving to them between each trail and during this period they were asked to complete NASA-TLX, STAI, and SACL questionnaires Data analysis A multivariate analysis of variance (MANOVA) was used to test statistical differences of the dependent variables based on the independent variables (type of tasks and CFQ scores). Furthermore, a pairwise comparison with Bonferroni confidence interval adjustment was used to determine specific differences between levels for any dependent variables that showed statistical significance. All results showed normal distribution of data, and a 5% significant level was adopted in all tests Attentional demand 3. Results Mean of the subjective measures for the experimental trails summarized in table (1).Overall score of the NASA- TLX showed that the participants did perceive divided attention task more demanding than selective or sustained attention task (F (2, 69) = 13.64, p < 0.001). Specifically, except for temporal demand and performance (F (2, 69) = 0.57, p = 0.57; F (2, 69) = 3.1, p = 0.58), they experienced different level of mental and physical demand as well as different level of effort and frustration when performing the instructed attentional demand tasks. Wherein, the participants showed an increased level of mental and physical demand (F (2, 69) = 9.42, p > 0.001; F (2, 69) = 16.94, p > 0.001) and a high level of effort (F (2, 69) = 5.45, p = 0.006)

6 needed for performing divided attention task. On the other hand, they evaluated sustained attention task as more frustrating task (F (2, 69) = 8.48, p = 0.001) with a higher level of effort (F (2, 69) = 5.45, p = 0.006) required for continuing this task. Finally, subjective measures indicated that selective attention was less demanding, concerning all categories of subjective workload perception, than divided and sustained attention tasks. The results of the SACL questionnaire indicated that performing sustained attention task contribute to an increased level of stress (F (3, 68) = 4.49, p = 0.006) and a decreased level of arousal (F (3, 68) = 5.07, p = 0.003) as compared to the initial psychological state of the participants in pre-study stage. Nonetheless, MANOVA test showed no significant differences in state anxiety level of the participants between the tasks and the baseline levels (F (3, 68) = 2.22, p = 0.094) Cognitive failure The means of CFQ total scores (sum of responses on 25 items) was with a standard deviation of and a range of Based on this score the participants were divided into three groups of low (>25), medium (26-40), and high (<41), as 8 participants were assigned in each group. Then, all comparison was made between these groups. Table 1: Mean (SE) of the subjective measures for the experimental trails (n = 24) Task Type MANOVA Variables Prestudy SCWT MATB SART F p value Subjective evaluation of workload (NASA-TLX) AWWL N/A > (3.1) (3.61) (3.51) R-TLX N/A > (3.29) (3.67) (3.62) Mental demand N/A > (6.23) (3.56) (7.34) Physical demand N/A > (4.69) (6.14) (7.54) Temporal demand N/A (6.31) (7.36) (6.81) Performance N/A (4.43) (6.54) (5.07) Effort N/A (4.98) (4.38) (5.35) Frustration N/A (4.54) (6.81) (5.65) Stress Arousal Scale (SACL) Stress level (0.66) (1.11) (1.14) (0.89) Arousal level (0.75) (0.54) (0.58) (0.35) State Trait Anxiety Inventory (STAI) State Anxiety (1.54) (2.01) (2.68) (2.92) AWWL = Adaptive Weighted Work Load; R-TLX = Raw overall score; SCWT = Stroop color word test; MATB = Multi- Attribute Task Battery; SART = Sustained attention response test; N/A= Not Applicable.

7 The means of the subjective measures for each group of the CFQ s based on each type of the tasks summarized in table (2). Considering subjective evaluation of workload (table 2), sustained attention task was more demanding task to the high CFQ score individuals than the low and medium s level, but only in the overall and mental demand scores (AWWL F (2, 21) = 4.89, p = ; R-TLX F (2, 21) = 7.46, p = 0.004; MD F (2, 21) = 8.19, p = ). Likewise, the high score groups did perceive divided attention task more challenging (AWWL F (2, 21) = 3.48, p = ), although they showed no significant differences in any of the subscale scores of the NASA-TLX scores. In addition, considering participants psychological states, although SACL scores indicated no significant differences between the CFQ groups for either type of attention demand. But, pairwise comparison yielded a significant increased level of stress and anxiety for the high s as compared to the low s when they performed divided attention task (p = ; p = ). None of the stress, arousal, and sate anxiety level significantly differed between the CFQ groups when performing either sustained attention or selective attention task. Table 2: Mean (SE) of the subjective measures for each group of the CFQ s in each type of the tasks (n = 24) SCWT MATB SART Variables Low Medium High Low Medium High Low Medium High Subjective evaluation of workload (NASA-TLX) AWWL * * (5.7) (4.74) (6.25) (5.54) (6.25) (5.55) (4.66) (4.86) (5.62) R-TLX * (6.48) (4.67) (5.97) (5.86) (6.95) (5.32) (3.62) (4.22) (5.97) Mental demand * (12.79) (7.12) (12.66) (7.01) (5.25) (6.16) (8.86) (10.03) (9.78) Physical demand (6.23) (7.45) (10.09) (10.84) (12.5) (8.35) (6.95) (12.79) (13.36) Temporal demand (10.99) (10.37) (11.89) (13.76) (11.31) (13.79) (14.36) (12.88) (8.06) Performance (3.01) (8.75) (9.93) (9.83) (13.08) (11.6) (7.64) (10.74) (6.85) Effort (10.63) (9.7) (6.1) (3.72) (11.64) (1.097) (12.47) (7.68) (9.05) Frustration (7.68) (6.71) (6.34) (10.32) (11.9) (12.6) (6.6) (11.5) (8.53) Stress Arousal Scale (SACL) Stress level * (0.73) (2.33) (1.89) (1.24) (1.63) (2.35) (1.5) (1.28) (1.62) Arousal level (0.6 (1.2 (0.99) (1.16) (1.03) (0.8) (0.85) (0.66) (0.50) State Trait Anxiety Inventory (STAI) State Anxiety * (1.54) (2.03) (5.39) (1.95) (4.03) (6.21) (5.93) (3.52) (5.29) *P < 0.05; AWWL = Adaptive Weighted Work Load; R-TLX = Raw overall score; SCWT = Stroop color word test; MATB = Multi- Attribute Task Battery; CRT = Sustained attention response test; CFQ = Cognitive failure questionnaire.

8 Table 3. Means (SE) of performance-related measures for each group of the CFQ s (n=24) CFQ Group MANOVA Variable Low Medium High F p value Stroop Color Word Test (SCWT) Accuracy Rate (1.44) (1.9) (1.32) Reaction time (37.73) (38.19) (37.64) Multi Attribute Battery Test (MATB) Response time 1.58 (0.26) 1.7(0.086) 2.19 (0.49) Omission rate 1.63 (0.84) 2.75 (2.34) 2.71 (0.84) Fuel level 1653 (70.89) 1837 (62.47) 1749 (53.06) Sustained Attention Response Test (SART) Response time (18.39) (17.31) (12.46) Accuracy rate (0.12) (0.1) 99.8 (0.04) CFQ = Cognitive failure questionnaire. Furthermore, mean differences of performance related measures were analyzed between three groups of the CFQ s (table 3). As a result, high score participants showed a decreased level of accuracy F (2, 21) = 9.49, p = when they instructed to reply name in different color (incongruent stage of the Stroop test) in selective attention task, even though they carried out the task slower than the congruent (just name or color) stages (F (2, 21) = 5.49, p = ). But, in terms of sustained attention task, applying MANOVA test indicated no significant differences in response time as well as accuracy rate between the CFQ groups (F (2, 20) = 1.43, p = 0. 28; F (2, 20) = 2.92, p = 0.5). On the other hand, individual cognitive failure was a detrimental factor for performance variations in divided attention task. As in this case, low score individuals could manage the task easier with better performance than high score individuals. This group showed less omission rate F (2, 20) = 4.28, p = ) with a shorter response time F (2, 20) = 3.78, p = 0. 05). Fuel level was significantly higher in the low score group than the high score group F (2, 20) = 3.56, p = ). But, no significant differences were found for the RMSE changes between the groups. 4. Discussion The intention in this study was to put participants into three types of work situations, each of which requiring different kind of attention demand, and then to evaluate which one of these situations would be more demanding to the individuals through analyzing their subjective and objective measures under each situation. The three types of attentional demand resulted in different pattern of subjective workload and psychological reactions. In addition, individuals cognitive failure was found as a detrimental factor, to some extent, for variations in the subjective and objective measures. In the current study we found that different attention demand created different subjective rating of workload, as the highest rating was related to the divided attention task. Specifically, the highest rate of effort, mental and physical demand in this task might mirror greater amount of resource investment expended by the individuals. This is in line with the fact that maintenance of performance is achieved at the cost of greater effort/capacity expenditure, and this is reflected in the subjective workload ratings (Hancock et al., 1988). The MATB task is a multi complex task, which consists of tracking, monitoring, and fuel management tasks. Particularly the latest part of the task requires monitoring and making decision when the level of fuel in four tanks frequently were changed and to manage how it should be maintained in the optimum level. The tracking task and monitoring part needed to control simultaneously over the task implementation. Therefore, the observed rating of workload for this task might be coincided with the real demand of the task described here. Considering participants

9 cognitive failure background, high CFQ individuals perceived attention task much more challenging than the low s. This is consistent with the findings of Grubb et al. (1994), which they concluded mental workload is perceived as greater for individual obtaining high than for those obtaining low CFQ scores, even though performance scores may be much the same for the two groups. Yet, in the current study, high CFQ individuals showed not only an increased level of perceived workload, but their outcomes were also impaired further as compared to the low score individuals. Using NASA-TLX inventory, an increased level of perceived workload was observed in the current study when the participants undertaken sustained attention task. Evidences from studies employing the same method of subjective workload assessment suggest that although vigilance tasks are often considered rather undemanding measure, studies using the NASA-TLX workload measure consistently show that vigilance tasks impose a high workload to individuals (Warm, Dember, & Hancock, 1996; Warm & Dember, 1998). Also, higher rate of frustration observed in this study might be supported by their conclusion that rated mental demands and frustration are the most salient sources of workload. Furthermore, decreasing arousal level particularly is another impact of the vigilance task which was shown by the participants in this study. This confirming the arousal theory on sustained attention task in which prolonged task performance leads to a lowering of central nervous system arousal or activation, which in turn results in performance deterioration. However, considering CFQ scores, arousal level was not statistically different between the low and high s. This might be interpreted by the higher rate of anxiety and stress observed within this group. As the more anxiety and stress the participants experienced, the lower impact of the vigilance demand they may have in the arousal level. Finally, considering changes in performance related measures, except for the sustained attention task, in almost case of the tasks differences in performance measures were statistically significant between three groups of the CFQ s. Overall, the high score individuals showed worse performance than the low s. Although we found no experimental study to support the relationship between CFQ scores and performance, however, there are some field studies showing this relationship. As evidence, Larson and Merrit (1991) concluded that those individuals who had been cited for causing traffic accidents had significantly higher CFQ scores than those who had not. Furthermore, in a study to investigate the ability of the Cognitive Failures Questionnaire (CFQ) to predict accidents and job performance ratings, (Wallace et. al., 2004), both self-reports and objective reports of safety behavior were collected. Results demonstrated that both types of reports overlapped substantially and the authors concluded that the blunders and distractibility subscales of the CFQ could predict overall performance ratings. 5. Conclusion The current investigation indicated that different paradigms of attentional demand can create various levels of workload and subjective stresses. Specifically, highest rate of perceived workload for divided attention task and increased sense of frustration with decreased level of arousal for sustained attention were remarkable patterns showing the main characteristics of these tasks were found in the current study. Furthermore, individuals cognitive failure had a detrimental effect on both perceived workload and performance measures. In conclusion, the investigation of workload on the basis of attentional paradigms would make a new perspective in the domains of mental workload assessment and job analysis. Consequently, it would be useful to optimize the system, in real world tasks, by an attempt to minimize the attentional demand requiring for each element of system. In addition, the purpose of choosing between operators (low CFQ score individuals) or providing an operator with further training would be another solution to control the impact of individuals cognitive failure on lowering mental workload and performance decrements. 6. References Broadbent, D.E., Cooper, P.F., Fitzgerald, P., & Parkes, K.R. (1982). The cognitive failures questionnaire (CFQ) and its correlates. British Journal of Clinical Psychology, 21, 1-16.

10 Comstock J. R., and Arnegard R. J. (1992). Multi-Attribute Task Battery, NASA-tech. Memorandum , Hampton, VA: NASA Langley Research Center. Grubb, P.L., Miller. L.C., Nelson, W.T. Warm, J.S., Dember, W.N., & Davies, D.R., (1994). Cognitive workload and perceived workload in vigilance performance. In M. Mouloua & Parasuraman (Eds.), Human performance in automated systems: Current research and trends. Hillsdale, NJ: Lawrence Erlbaum Associates Inc. Hancock, P. A., and Meshkati, N. (1988), Human Mental Workload (pp.41-62). Amsterdam: North- Holland.Warm, J.S., Dember W.N., & Hancock, P.A. (1996). Vigilance and workload in automated systems. In R. Parasuraman & M. Mouloua (Eds.), Automation and Human performance: Theory and applications. Mahwah, NJ: Lawrence Erlbaum Associates Inc. Hart, S. G., & Staveland, L.E. (1988). Development of NASA-TLX (Task Load Index: Results of Empirical and Theoretical Research. In P. A. Hancock & N. Meshkati (Eds.), Human Mental Workload (pp ). Amsterdam: North-Holland. Heckausen, H. & Beckman, J. (1990). Intentional action and action slips. Psychological Review, 97, Hill, S. G., Iavecchia, H.P., Byers, J.C., Bittner, A. C., Zaklad, A. L., & Christ, R. E. (1992). Comparison of Four Subjective Workload Rating Scales. Human Factors, 34, Larson, GE., Alderton, D.L., Neideffer, M., & Underhill, E. (1997). Further evidence on dimensionality and correlates of Cognitive Failures Questionnaire. British Journal of Psychology, 88, Larson, GE., & Merritt, C.R. (1991). Can accidents be predicted? An empirical test of the Cognitive Failure Questionnaire. Applied Psychology: An International Review, 40, Mackay, C. (1978): An inventory for the measurement of self-reported stress and arousal. Br J Soc Clin Pschol, 17: Martin, M. (1983). Cognitive failure: Everyday and laboratory performance. Bulletin of Psychonomic Society, 21, Matthews, G., Davies, D.R., Westerman, S.J., & Stammers, R.B. (2000). Human Performance Cognition, Stress, and Individual Differences. London: Routledge. Miyake, S. & Kumashiro, M. (1993). Subjective Mental Workload Assessment Technique, An Introduction to NASA-TLX and SWAT and a Proposal of Simple Scoring Methods. The Japanese Journal of Ergonomics, 29, Monk C.A., Moyer M.J., Hankey J.M., Dingus T.A., Hanowski R.J., Wierwille W.W. (2000) Design Evaluation and Model of Attention Demand : A Tool for In-Vehicle Information System Designers DEMAnD, Public Roads, Vol. 64 No. 3. Reason J.T., (1988). Stress and cognitive failure. In S. Fisher & J.T. Reason (Eds.), Handbook of life stress, cognition, and health. John Wiley & Sons Ltd. Robertson IH, Manly T, Andrade J, Baddeley BT, Yiend J. (1997)'Oops!': performance correlates of everyday attentional failures in traumatic brain injured and normal subjects. Neuropsychologia, 35(6): Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (STAI). PaloAlto, CA: Consulting Psychologists Press.

11 Warm, J.S., Dember W.N. (1998). Tests of vigilance taxonomy. In R.R. Hoffman, M.F. Sherrick & J.S. Warm (Eds.), Viewing psychology as a whole: the integrative science of William N. Dember. Washington, DC: American Psychology Association. Warm, J.S., Dember W.N., Hancock, P.A. (1996). Vigilance and workload in automated systems. In R. Parasuraman & M. Mouloua (Eds.), Automation and Human performance: Theory and applications. Mahwah, NJ: Lawrence Erlbaum Associates Inc. Wallace, J. C., and Vodanovich, S. J. (2004) Can Accidents and Industrial Mishaps Be Predicted? Further Investigation into the Relationship between Cognitive Failure and Reports of Accidents, Journal of Business and Psychology, Volume 17, Number 4.

12 7. Appendixes a p a q b a Appendix 1: Sustained Attention Response Test (SART) 13) Stroop colored-word test Red Red Blue Appendix 2: Stroop Color Word Test (SCWT)

13 Appendix 3: Multi-Attributed Battery Task (MATB)