PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING2005 2119

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 other domains. In spite of its potentially broad

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
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 task-
related 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
PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING2005
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 5150.
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.
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
2120
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 surround-
sound (as shown in Figure 2).
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
PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING2005
task stimuli); 3) shooting accuracy; 4) accuracy on the
secondary task; and 5) subjective measures.
The experiment included the following within-
participant 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.
2121
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, Wilks = .98, F (1, 36) =
.38, p = .69, partial 2 = .11, and Wilks = .99, F(1, 36)
= .03, p = .87, partial 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.
82
80
Percent Corre
78
Secondary
76 Task
74
No task
72
70
0 2 4
Working Memory Demand
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, Wilks = .61, F (1, 18) =
11.47, p = .003, partial 2 = .39, and Wilks = .76, F(1,
18) = 5.61, p = .03, partial 2 = .61, respectively. These
results show that participants were significantly more
accurate on the recognition task when there was not a
PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING2005 2122

target present than when a target was present and when DISCUSSION
the working memory demand level was two (see Figure
4). This study is the first to investigate processing
efficiency in a military context and in a high fidelity
95 simulator. The use of a high fidelity simulator is
93 important in this case because the target aiming and
91 target acquisition task are performed using a real weapon
89
which adds to the physical and the hand-eye
Percent Corre

87
Target coordination demands of the task. As can be seen from
85
No target the results there are changes in performance for both the
83
primary and the secondary task as the global demands
81
change, clearly indicating that it is not the case of one
79
task dominating the other. However, contrary to our
77
hypothesis, when the task was most demanding (both
75
2 4
primary and secondary tasks were present in the highest
working memory demand condition), participants were
Working Memory Demand
more accurate on the shooting task when the secondary
was present than when there was no secondary task. One
Figure 4. Relationship between the presence of a target
possible explanation for these results is that there were
and working memory demand.
combined practice and fatigue effects. Hence, the
variability among participants increased, and possibly
Perceived Workload. A 3(working memory demand)
the analysis of the anxiety data could contribute to the
X 4(block) within-participants repeated measures
understanding of these changes. Additionally, because
ANOVA was conducted on the subjective ratings of
participants were significantly less accurate on the
mental workload (RSME). There was a significant main
recognition task in the same high demand condition, it is
effect of working memory demand and block, Wilks
possible that participants stopped attending to the
= .35, F (2, 31) = 29.32, p = .001, partial 2 = .65, and
secondary task and focused all of their attention on the
Wilks = .55, F(3, 31) = 8.08, p = .001, partial 2 =
primary, more salient shooting task. This would be
.45, respectively. Specifically, participants reported that
consistent with the Hancock and Warm (1989) model
a higher level of mental effort was required as a function
indicating that participants were no longer in their
of working memory demand; as the working memory
comfort zone and were unable to maintain an optimal
demand increased, the ratings of mental effort increased
level of performance on both tasks. Furthermore, such
(see Figure 5). Additionally, and consistent with our
pattern could also be explained using Hockeys (1997)
hypotheses, participants reported that a lower level of
compensatory control model. Specifically, it may be that
mental effort was required as a function of time in each
as attentional demand increased participants regulated
working memory demand condition.
their effort downward on the secondary task to free
75
resources required to successfully engage the primary
task.
65 Furthermore, the mental workload measures
Block 1 correlate with this global demand, indicating a
Correct

55
Block 2 performance-workload association (Hancock, 1996; Yeh
45 Block 3 & Wickens, 1988). Further investigations of this
Block 4 relationship are necessary. Current efforts are underway
35 to examine the hypothesized effects of trait anxiety on
performance and perceived workload. In addition, the
25
0 2 4
other form of workload assessment, physiological
measurement, will be examined to test whether there is
Working Memory Demand
physiological evidence that participants reduced their
effort in the face of high working memory demand.
Figure 5. Relationship between subjective mental effort These analyses will provide a more coherent
and working memory demand. understanding of the results. Future studies will be able
to refine these findings and better explain the tradeoff in
performance as task demand increases.
PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 49th ANNUAL MEETING2005 2123

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 DAAD19-01-1-0621, 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. 99-104). Scottsdale, AZ:
Pabst Science Publishing.
Eysenck, M.W., & Calvo, M.G. (1992). Anxiety and performance:
The processing efficiency theory. Cognition and Emotion, 6,
409-434.
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), 73-93.
Hancock, P.A., & Warm, J.S. (1989). A dynamic model of stress and
sustained attention. Human Factors, 31(5), 519-537.
Jones, G., & Swain, A.B.J. (1992). Intensity and direction dimensions
of competitive state anxiety and relationships with
competitiveness. Perceptual and Motor Skills, 74, 467-472.
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, 533-542.
Woodman, T. & Hardy, L. (2001). A case study of organizational
stress in elite sport. Journal of Applied Sport Psychology, 13(2),
207-238.
Yeh, Y., & Wickens, C.D. (1988). Dissociation of performance and
subjective measures of workload. Human Factors, 30(1), 111-
120.
Yerkes, R.M., & Dodson, J.D. (1908). The relation of strength of
stimulus to rapidity of habit formation. Journal of Comparative
Neurology & Psychology, 18, 459-482.
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.