TESTING THE EFFICIENCY AND EFFECTIVENESS OF SOLDIER PERFORMANCE UNDER INCREASING TASK DEMAND

Transcription

1 PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING TESTING THE EFFICIENCY AND EFFECTIVENESS OF SOLDIER PERFORMANCE UNDER INCREASING TASK DEMAND K.A. Burke, J. Szalma, T. Oron-Gilad, A. Duley, and P.A. Hancock University of Central Florida Orlando, FL Combat is inherently a demanding situation which may increase stress, heighten arousal, and increase anxiety. The Processing Efficiency Theory (PET) was specifically developed to account for how anxiety influences performance. Therefore, the PET provides a potential theory to explain the positive and negative changes in performance in a combat environment. This study is the first attempt to examine PET in the military domain. Using the Small Arms Simulator Testbed (SAST), we investigated the relationship between processing efficiency and performance, considering mental workload, stress, and anxiety effects. Shooting performance effectiveness measures were target acquisition and friend/foe discrimination. Changes in processing efficiency were manipulated by varying degrees of working memory demand and sustained information transfer. The results indicated that shooting performance and processing efficiency, as well as mental workload demands, decreased as the global demand of both tasks increased. Further analyses for anxiety and stress and future directions are discussed. INTRODUCTION Combat is inherently a demanding situation which may increase stress, heighten arousal, and increase anxiety especially when the outcome is important and uncertain (Jones & Swain, 1995). Additionally, in a combat environment, maintaining a high level of performance, both of cognitive skills (i.e., decision making tasks) and motor skills (i.e., shooting accurately) is critical and may mean the difference between life and death. Although the relationship between anxiety and performance has been investigated extensively (Eysenck & Calvo, 1992; Yerkes & Dodson, 1908), the mechanisms underlying the facilitative or debilitative effects of anxiety on performance are not yet clearly understood. The Processing Efficiency Theory (PET; Eysenck & Calvo, 1992) was specifically developed to account for how anxiety influences performance. Traditionally, anxiety has been considered detrimental to performance, however a significant body of literature suggests that anxiety has both facilitative and debilitative functions (e.g., Jones, Hanton, & Swain, 1994; Jones & Swain, 1992; Jones, Swain, & Hardy, 1993). The PET (Eysenck & Calvo) seeks to explain how individual differences in trait anxiety interact with variations in state anxiety and attentional mechanisms to affect performance. The PET has primarily been applied to the examination of test anxiety (Eysenck, 1996), but the theory is broad enough to have wide application in other domains. In spite of its potentially broad applicability, the theory has not been empirically validated in other domains (Woodman & Hardy, 2001) with the exception of one study in the sport psychology domain (Murray & Janelle, 2003). As far as we know, our study is the first to investigate the PET in the military domain and in a high fidelity environment. The SAST II (Small Arms Simulator Testbed II) is a high fidelity simulation allowing for superior assessment of shooting performance in simulation. In this study the SAST II was utilized to test hypotheses based on PET. In many instances performance on a primary task may be maintained in the face of anxiety or stress only at the cost of increased effort. That is, performance may remain effective but less efficient. Primary task performance may therefore mask detrimental effects of anxiety and stress. The aim of the current study was to identify changes in processing efficiency under varying degrees of working memory demand and sustained information transfer and to examine the interaction between taskrelated stress. METHOD Participants Thirty-eight participants (mean age = 20.4, SD = 4.6, range = 18-43) were recruited on a voluntary basis from undergraduate psychology classes at the University

2 PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING of Central Florida and were paid $22.50 for their participation (based on an hourly rate of $7.50). Materials Questionnaires. Participants completed several subjective questionnaires to measure individual differences including workload, stress, and anxiety. All subjective measures were administered via Inquisit, psychological data collection software, on a Dell Inspiron Experimental Tasks Primary Task. Target acquisition task and a Friend/foe discrimination task: two different types of stimuli/targets appeared on the screen (Figure 1): the non-target was an M-16 (friend), the target was an AK- 47 (foe); both stimuli/targets were superimposed over the silhouette stimulus; the discrimination was based upon the weapon. The spatial location of each stimulus (targets/non-targets) was randomly generated to prevent the predictability of the appearance of a target. pressed in response to the presentation of a valid cue word and true association word, the second trigger was pressed in response to the presentation of a valid cue word and false association word, and the third trigger was pressed in response to an invalid cue word. Apparatus Small Arms Simulator Testbed II (SAST II). (Naval Air Warfare Center Training Systems Division, 2000) is a single user research testbed for the evaluation and study of both current and new weapon concepts. The SAST controlled the presentation of targets (target type, location, time of onset, and exposure time), and recorded weapon aim-point data and shot results from an M16 rifle equipped with an infrared emitting diode and collimator lens. Weapon recoil was simulated with an electromechanical recoil system and weapon sounds were simulated via digital audio surroundsound (as shown in Figure 2). AK47 M16 Figure 1. Screenshot from the SAST II displaying the two types of targets superimposed on the silhouette stimuli. Secondary Task. Memory recall task: consisted of the auditory presentation of cue words and association words which simulated unit battalion numbers (e.g., 21, 23, 27) and locations (north, south, east, west), respectively, and were designed to provide varying demands on working memory. There were three secondary task conditions which varied according to working memory task demand, these being zero (control condition), and two or four items to be retained in memory. During execution of the primary task the participant was presented with the secondary task. The participants responded to the secondary task stimuli via three triggers mounted to the weapon. One trigger was Figure 2. The SAST II facility including the weapon and the visual display. LabVIEW 7.0. (National Instruments, 2004). LabVIEW is a graphical development environment for creating flexible and scalable test, measurement, and control applications rapidly and at minimal cost. For the current study, LabVIEW was used to integrate the SAST II system, and the presentation of the secondary task. Design Overview The dependent measures of interest were: 1) response time for the primary task (mean response time participants required to pull the trigger on the weapon after the appearance of the targets); 2) response time for the secondary task (mean response time participants required to press a trigger in response to the secondary

3 PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING task stimuli); 3) shooting accuracy; 4) accuracy on the secondary task; and 5) subjective measures. The experiment included the following withinparticipant factors: 1) presentation of secondary task (present, not present); 2) increasing levels of working memory demand tasks (recognition/recall; three levels); and 3) friend/foe discrimination task (half of the trials included a target (foe), half of the trials did not include a target. Procedure Each participant was randomly assigned to one of the three working memory presentation orders, with the constraint that an equal number of participants were tested in each order. For each participant, the sand bags enabling a seated, supported firing position were adjusted for differences in height, the weapon was calibrated for accurate tracking and aiming vectors, and the triggers attached to the weapon (to allow the participant to respond to the secondary task stimuli) were adjusted for handedness and finger position. The participants received training and practice so that they would become familiar with firing the weapon, identifying the targets, and the secondary task. Each session (approximately three hours) consisted of administration of the subjective measures, training and practice on both the primary and secondary tasks, and of the presentation of the three conditions of the secondary task. Each participant completed shooting scenarios for each condition of the secondary task (each working memory demand condition) which consisted of four blocks of 16 trials totaling 64 trials per working memory condition. Each trial consisted of the simultaneous presentation of one of two types of target stimuli (target, or non-target) and in half of the trials, the presentation of the secondary task. The trials were organized such that each participant was presented with an equal number of trials which included: the secondary task with an enemy target, secondary task without an enemy target, no secondary task with an enemy target, and no secondary task without an enemy target. The presentation of trials within each working memory demand condition was randomized so that each condition was presented to each participant an equal number of times. Additionally, the order of presentation of each working memory demand condition (0, 2, and 4) was counterbalanced. Following each block, participants completed the RSME and between conditions, participant received a five-minute break. RESULTS Primary Task Performance. A two-way 2(secondary task present or not ) X 3(working memory (0,2, and 4)) within-participants repeated measures analysis of variance (ANOVA) was conducted on performance in the friend/foe discrimination primary shooting task. The independent variables consisted of presentation of the secondary task in which the battalion number and location combinations were presented to the participant and varying levels of working memory demand. The dependent variable was the percent correct (number of times the participant correctly identified and hit the target) at each level of working memory demand. The main effect of presence of the secondary task and working memory demand on shooting performance was not statistically significant, Wilk s Λ =.98, F (1, 36) =.38, p =.69, partial η 2 =.11, and Wilk s Λ =.99, F(1, 36) =.03, p =.87, partial η 2 =.001, respectively. However, as depicted in Figure 3, in working memory conditions 0 and 2, participants were more accurate in the shooting task when there was no secondary task than when the secondary task was present. This was not the case in working memory condition 4; possibly due to the combination of practice and fatigue effects. Percent Corre Secondary Task No task Figure 3. Relationship between presence of the secondary task and working memory demand. Secondary Task Performance. A two-way 2(primary task target present or not) X 2(working memory demand (2, 4)) within-participants repeated measures ANOVA was conducted on secondary task performance (percent correct-- number of times the participant correctly recognized and responded to the presentation of the battalion number and azimuth location). There was a significant main effect of presence of a target and working memory demand, Wilk s Λ =.61, F (1, 18) = 11.47, p =.003, partial η 2 =.39, and Wilk s Λ =.76, F(1, 18) = 5.61, p =.03, partial η 2 =.61, respectively. These results show that participants were significantly more accurate on the recognition task when there was not a

4 PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING target present than when a target was present and when the working memory demand level was two (see Figure 4). Percent Corre Target No target Figure 4. Relationship between the presence of a target and working memory demand. Perceived Workload. A 3(working memory demand) X 4(block) within-participants repeated measures ANOVA was conducted on the subjective ratings of mental workload (RSME). There was a significant main effect of working memory demand and block, Wilk s Λ =.35, F (2, 31) = 29.32, p =.001, partial η 2 =.65, and Wilk s Λ =.55, F(3, 31) = 8.08, p =.001, partial η 2 =.45, respectively. Specifically, participants reported that a higher level of mental effort was required as a function of working memory demand; as the working memory demand increased, the ratings of mental effort increased (see Figure 5). Additionally, and consistent with our hypotheses, participants reported that a lower level of mental effort was required as a function of time in each working memory demand condition. Correct Block 1 Block 2 Block 3 Block 4 Figure 5. Relationship between subjective mental effort and working memory demand. DISCUSSION This study is the first to investigate processing efficiency in a military context and in a high fidelity simulator. The use of a high fidelity simulator is important in this case because the target aiming and target acquisition task are performed using a real weapon which adds to the physical and the hand-eye coordination demands of the task. As can be seen from the results there are changes in performance for both the primary and the secondary task as the global demands change, clearly indicating that it is not the case of one task dominating the other. However, contrary to our hypothesis, when the task was most demanding (both primary and secondary tasks were present in the highest working memory demand condition), participants were more accurate on the shooting task when the secondary was present than when there was no secondary task. One possible explanation for these results is that there were combined practice and fatigue effects. Hence, the variability among participants increased, and possibly the analysis of the anxiety data could contribute to the understanding of these changes. Additionally, because participants were significantly less accurate on the recognition task in the same high demand condition, it is possible that participants stopped attending to the secondary task and focused all of their attention on the primary, more salient shooting task. This would be consistent with the Hancock and Warm (1989) model indicating that participants were no longer in their comfort zone and were unable to maintain an optimal level of performance on both tasks. Furthermore, such pattern could also be explained using Hockey s (1997) compensatory control model. Specifically, it may be that as attentional demand increased participants regulated their effort downward on the secondary task to free resources required to successfully engage the primary task. Furthermore, the mental workload measures correlate with this global demand, indicating a performance-workload association (Hancock, 1996; Yeh & Wickens, 1988). Further investigations of this relationship are necessary. Current efforts are underway to examine the hypothesized effects of trait anxiety on performance and perceived workload. In addition, the other form of workload assessment, physiological measurement, will be examined to test whether there is physiological evidence that participants reduced their effort in the face of high working memory demand. These analyses will provide a more coherent understanding of the results. Future studies will be able to refine these findings and better explain the tradeoff in performance as task demand increases.

5 PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING ACKNOWLEDGEMENTS The views expressed in this work are those of the authors and do not necessarily reflect official Army policy. This work was supported by the DoD Multidisciplinary University Research Initiative (MURI) program administered by the Army Research Office under grant DAAD , Peter A. Hancock, Principal Investigator. REFERENCES Baddeley, A.D. (1986). Working memory. Oxford, UK: Clarendon Press. Eysenck, M.W. (1996). Anxiety, processing efficiency theory and performance. In W. Battmann & S. Duke (Eds.), Processes of the molar regulation of behavior (pp ). Scottsdale, AZ: Pabst Science Publishing. Eysenck, M.W., & Calvo, M.G. (1992). Anxiety and performance: The processing efficiency theory. Cognition and Emotion, 6, Hockey, R.G. (1997). Compensatory control in the regulation of human performance under stress and high workload: A cognitive-energetical framework. Biological Psychology, 45(1-3), Hancock, P.A., & Warm, J.S. (1989). A dynamic model of stress and sustained attention. Human Factors, 31(5), Jones, G., & Swain, A.B.J. (1992). Intensity and direction dimensions of competitive state anxiety and relationships with competitiveness. Perceptual and Motor Skills, 74, Jones, G., Swain, A.B.J., & Hardy, L. (1993). Intensity and direction dimensions of competitive state anxiety and self-confidence. Journal of Sports Sciences, 11, Woodman, T. & Hardy, L. (2001). A case study of organizational stress in elite sport. Journal of Applied Sport Psychology, 13(2), Yeh, Y., & Wickens, C.D. (1988). Dissociation of performance and subjective measures of workload. Human Factors, 30(1), Yerkes, R.M., & Dodson, J.D. (1908). The relation of strength of stimulus to rapidity of habit formation. Journal of Comparative Neurology & Psychology, 18, Zijlstra, F.R.H. (1993). Efficiency in work behavior. A design approach for modern tools. PhD thesis, Delft University of Technology. Delft, The Netherlands: Delft University Press. Zijlstra, F.R.H. & Van Doorn, L. (1985). The construction of a scale to measure perceived effort. Delft, The Netherlands: Department of Philosophy and Social Sciences, Delft University of Technology.