Abstract
Objective:
We sought to determine the influence of message presentation rate (MPR) and sensory modality on soldier cognitive load.
Background:
Soldiers commonly communicate tactical information by radio. The Canadian Army is equipping soldiers with a battle management system (BMS), which also allows them to communicate by text.
Method:
We varied presentation modality (auditory vs. visual) and MPR (fast or slow) in an experiment involving a tactical scenario. Participants (soldiers) received messages and periodically provided situation reports to higher level command, and the scored reports were used to provide a measure of situation awareness (SA). The detection response task (DRT) and NASA-TLX were used to measure cognitive load.
Results:
The fast MPR reduced DRT accuracy and increased response times relative to slow MPR. The NASA-TLX results also showed higher subjective workload ratings for several subscales with fast MPR. Messages presented visually produced greater cognitive load, with slower DRT response times for the visual than the auditory condition. SA scores were higher with slower MPR and auditory presentation. There was no statistical interaction of presentation modality and rate for any measure.
Conclusion:
Fast MPR and visual presentation increased cognitive load and degraded SA.
Application:
These findings show that the DRT can be used to measure workload effectively in a tactical military context and that the method of information presentation affects how soldiers process information in a BMS.
Mental workload is one of the most important concepts in human factors research and application (Bailey & Iqbal, 2008; Moray, 1979; Parasuraman & Hancock, 2001; Vidulich & Tsang, 2012; Wickens, 2008a). Workload is a pervasive issue in sociotechnical systems, in which designers and managers ask questions like: How busy is the operator with the primary task? Can any additional tasks be handled? How does the operator feel about the tasks being performed? Despite interest in the topic for the past 40 years, a universally accepted definition of workload remains elusive. Mental workload is thought to be multidimensional, including multiple types or sources. Thus, the NASA-TLX (National Aeronautical and Space Administration Task Load IndeX; Hart & Staveland, 1988) has six different subscales, including physical demand and frustration level.
The motivation for the current work comes from the Canadian Army’s interest in the potential for cognitive overload that new battle management systems (BMSs) and other information technologies could impose on the dismounted soldier (Bell, 2012). We focus here specifically on cognitive load, which we define as the information-processing demands imposed on an operator when performing a task requiring the interpretation and comprehension of data. Primarily, such tasks will require working memory resources, including the short-term storage of spatial, temporal, and numeric information; manipulation of that information; and relating the information to knowledge stored in long-term memory. Such working memory resources are commonly considered to be of limited capacity (Baddeley, 2000; Cowan, 2000).
The use of the secondary task technique to measure workload has a long history in engineering psychology, gauging the spare capacity available when a primary task of interest is being performed (Norman & Bobrow, 1975; Wickens, 1984). Typically, one task (the primary task) is varied in difficulty, and performance on the secondary task is measured as an index of spare capacity. One particular version of a secondary task—called the detection response task (DRT)—has been extensively utilized in studies of driver distraction (e.g., Strayer et al., 2015) and has become an international standard in that domain (International Organization for Standardization [ISO], 2016). With the DRT, an LED (light emitting diode) is mounted on the participant’s head to appear in the visual periphery. The LED is timed to flash at random intervals, and the participant presses a button whenever the light is detected. The DRT offers a method for assessing the cognitive load associated with the primary task: As the load of the primary task increases, the participant is typically slower to respond to the light and more likely to miss the light altogether.
Another prevalent human factors topic is the concept of situation awareness (SA; Banbury & Tremblay, 2004; Durso & Sethumadhavan, 2008; Endsley, 1995; Endsley & Garland, 2001). Engineering psychologists have taken a considerable interest in the concept (Wickens, 2008b) because of its importance in designing effective displays that support SA and understanding the causes of accidents in which SA has been lost. Endsley (1988) defined SA as the perception of critical elements in the environment, comprehension of their meaning, and projection of their status into the future. In the military context, a tactical commander needs accurate data on the enemy and friendly forces in a given area and their relationship to other points of reference. From this information, the commander can deduce certain likely objectives and project that enemy forces are more or less likely to attack in a given manner. In a complex environment, the information can arrive quickly and exceed an operator’s capacity to interpret it, producing both high cognitive load and reducing SA, which will likely impede effective decision making.
The Canadian Army has identified both cognitive overload and commander situation awareness as contributors to persistent “Army Hard Problems” (Bell, 2012). Currently, Canadian soldiers are being equipped with a BMS called the integrated soldier system (ISS), which allows soldiers to communicate via multiple text and radio channels. (It also displays geospatial information, including the locations of individual friendly soldiers, in real time.) ISS has the potential to reduce cognitive load and improve SA for tactical decision makers. Alternatively, there is the potential that it could increase cognitive load and worsen SA, especially if the human-machine interface is designed poorly. For ISS and systems like it to be truly effective, they should communicate the information needed by soldiers effectively and be easy to use.
Currently, Canadian soldiers communicate tactical information (e.g., grid reference values, target identity and range, information about casualties) by radio. ISS allows the same information to be communicated visually. For instance, the grid reference for an enemy location can be typed into the texting system, and the recipient reads the text visually. In some cases, the use of text could provide an advantage over auditory presentation, improving communication accuracy and potentially reducing cognitive load, at least for the recipient. An important question is whether the presentation of tactical information in visual format (relative to the standard auditory format) provides a demonstrable advantage in terms of reduced cognitive load and improved SA.
To assess this question experimentally, we presented tactical scenarios to each participant, who acted as a platoon commander. The scenarios were presented in the form of a series of messages from subordinates; as these are presented, the participant seeks to build a tactical picture of the scenario. The messages were presented either visually (text) or in an auditory format (as radio messages). The message presentation rate (MPR) was manipulated to be either fast (approximately 7.2 messages per minute) or slow (approximately 3.7 messages per minute). These MPRs are representative of rates soldiers would see in the field with a platoon under attack, as in our scenarios. In each experimental condition, we measured cognitive load using DRT response time (RT) and accuracy. The DRT offers a method for assessing the cognitive load associated with the soldier’s primary task of maintaining SA and preparing reports. We also took subjective workload ratings from the NASA-TLX.
To assess SA, the participant was periodically tasked with providing a simulated situation report (SITREP) to higher levels of command. During the SITREP, incoming messages ceased, following the SAGAT (situation awareness global assessment technique) simulation halt procedure (Endsley, 1988). The SITREP is a necessary and familiar task element to a platoon commander. Although the cessation of incoming messages during the SITREP is not necessarily realistic, continuing to present messages would likely interfere with the SITREP content, complicating the SA assessment.
We predict that a fast MPR should increase cognitive load, and therefore DRT performance should degrade relative to a slow MPR (i.e., we should observe reduced accuracy and longer RTs). If cognitive load is greater with fast MPR, then we should also expect to see higher subjective mental workload ratings in the NASA-TLX scores. Further, the faster MPR should degrade SA relative to a slower MPR because if the messages are not processed properly, some information will be lost or misremembered.
An advantage of text presentation is that messages can be stored until such time as they can be read and interpreted. In contrast, radio messages are “pushed” and need to be processed right away. We might therefore expect better SA with visual than auditory presentation, although it is not clear what effect the flexibility of visual message presentation will have on cognitive load. However, the DRT requires visual detection, which is the same perceptual modality as the primary task in the visual conditions; as a result, there could be interference between the two visual tasks. In this case, we would expect better SA performance and lower cognitive load with auditory than visual message presentation.
Our design also allows an assessment of whether the effect of increased cognitive demand (MPR) is the same under visual and auditory message presentation or whether it differs. The former result would be consistent with a model in which demand is independent of modality. Alternatively, we might expect that the increase in cognitive load with MPR will be greater for auditory presentation of messages over visual because of the immediate demand the auditory messages place on working memory. In contrast, reading text is generally faster than speech (Carver, 1970; Yuan, Liberman, & Cieri, 2006), and so even with fast MPR, participants may be able to read the messages more quickly. If so, we would expect an interaction between presentation modality and MPR.
Method
Participants
Twenty-four soldiers (23 men and 1 woman aged 21–55 years; M = 36 years, SD = 10) were recruited as participants. The male-female ratio of our participants is representative of the combat arms occupations with the Canadian Army (about 2.4% female). Ten of the participants were warrant officers or master warrant officers. The rank of the remaining participants varied, including the following: master corporal, sergeant, second lieutenant, lieutenant, and major. Sixteen of the participants were warrant qualified or platoon commander qualified; that is, each had completed the requisite training to enable them to understand the responsibilities of a platoon commander. Those participants not being formally qualified were assessed by a subject matter expert to determine whether their previous experience was sufficient to enable them to accomplish the task. The participants had an average 17 years of service and were solicited from Army Reserve and Regular Force units (primarily infantry and engineer) in the Toronto area using the Canadian Forces Tasking, Plans, and Operations process. Participants were reimbursed $13.37 CAD (in addition to reserve pay as necessary) for their participation. The research complied with the Canadian Tri-Council Policy Statement on Ethical Conduct for Research Involving Humans and was approved by the Defence Research and Development Canada Human Research Ethics Committee, Protocol No. 2018-020.
Stimuli and Apparatus
The scenario for the experiment was loosely based on the Battle of Kamdesh, an attack on the United States Command OutPost (COP) Keating by Taliban insurgents, which took place in 2009 in Eastern Afghanistan (Tapper, 2012). For purposes of a cognitive task analysis, Tack and Nakaza (2016) produced a scenario based on COP Keating, dubbed Platoon House Pearson. This scenario describes a fictional attack on a Canadian Army base placed near a Somalian town, with typical insurgent tactics and weapons and Canadian Army doctrine, units, and weapons for the defending force. We used the Platoon House Pearson scenario to generate four sets of sequential events for such an attack. Specific details changed across the four sets, but each set represented a plausible sequence of events within the scenario. These were described in a series of 30 messages, each one representing an informal situation report from an observer (e.g., a section commander) to the platoon commander. Examples of the messages are shown in Table 1.
Examples of the Messages Presented to Participants
The auditory messages were recorded as .wav (wave sound) files, recorded on a Roland R-05 Wave/MP3 Recorder. The speakers were five soldiers with appropriate experience, to simulate the voices of different platoon members (none of the speakers later served as a participant). The messages varied in duration from less than a second to 12 s, with a mean of 4.86 s. Audacity audio software (v 2.2.2) was used to edit and organize the .wav files. Prerecorded battle sounds (Hawkins, 2018) were mixed onto the recordings using Audacity to enhance the participant’s presence within the tactical environment.
Messages were presented sequentially, and the same sequence was used for auditory and visual conditions. In the auditory conditions, the messages were played on AKG K141 Monitor headphones. The time between the end of one message and the beginning of the next was determined by the interstimulus interval (ISI), which was either 3 s for the fast or 10 s for the slow conditions.
In the visual condition, the messages were presented on a cell phone display. The messages appeared in a list format as they would on the ISS system. The message appeared in the list at the time determined by the auditory sequence described previously. That is, the sequence of times the message first appeared in the visual condition corresponded to the start times for each message in the auditory condition. The participant could select a particular message to read by clicking on it and then click on the “back” icon on the upper left of the screen to close the message to return to the list. The participant could choose when to open each message. The messages were retained in the list for the duration of the trial and could be opened (and closed) again. This emulated the behavior of the ISS system.
Three devices were used, requiring cross-device communication: a cell phone, the DRT, and a laptop. The cell phone was a Samsung Galaxy S6 running Android L (Lollipop). Android ran a messenger service that emulated the ISS interface on the cell phone, sending messages to the participant at predetermined times. The messenger service was written using Android SDK, which uses Java. The Red Scientific DRT+ (which follows the ISO 17488 standard) was used to present the LED, control its timing, and collect responses from a micro-switch. The experiment was controlled by a 64-bit Dell Latitude 14 laptop running Windows 7 Professional. The DRT was connected to the laptop through a USB serial port, and the cell phone was connected to the laptop using a Wi-Fi network. A custom C# application (developed in Visual Studio, 2013 using .NET framework version 4.5.1 and C# version 5.0) on the laptop initiated the experiment on both the cell phone and the DRT nearly simultaneously (any time difference was less than 12 ms). The program logged events that occurred on either of the two devices and output the time-stamped events to a .csv file. Events included the presentation of and response to a DRT stimulus, presentation of messages, and opening and closing of messages on the cell phone in the visual conditions.
For the purpose of the SA report, a template was provided that required participants to identify all enemy combatants by location. Enemy was defined as individuals, groups, vehicles, or groups of vehicles that had engaged the platoon or the platoon had fired on. Along with location, participants were asked to report the numbers, weapons, and actions of each enemy group. Finally, participants were tasked with reporting ammunition, actions, and casualties for friendlies (Canadian soldiers). An example SA report is shown in Table 2.
Sample Situation Awareness Report: Enemy Group Identified by Location, Number, Weapons, and Actions
Subjective workload ratings were collected after each trial using NASA-TLX (Hart & Staveland, 1988). Participants responded to each of the six items on a 21-point Likert scale. Participants did not weight the subscales (i.e., the resulting scores were unweighted).
Design and Procedure
The experiment had a 2 × 2 within-subjects design. The independent variables were: presentation type (visual or auditory) and MPR (slow or fast, 3.74 or 7.14 messages/min). Dependent variables were SA accuracy, DRT accuracy and RT, and subjective workload measures (NASA-TLX). The order of the four conditions was counterbalanced, and each participant was randomly assigned to a participant number (which determined the order of conditions).
The experimenter briefed participants about the goals of the study, the experimental scenario, and the measures. Participants signed an informed consent form and filled out a demographics form. Each participant was seated at a desk in a partially lit room, and the experimenter fitted the DRT to the participant’s head. The DRT LED was positioned approximately 15° to the left and 7.5° above the left eye and was held in a fixed position using a headband so that the DRT did not occlude the direct field of view. The DRT micro-switch was placed on the thumb of the participant’s nondominant hand.
In each of the four conditions, participants played the role of a platoon commander (PC) in the Platoon House Pearson scenario (Tack & Nakaza, 2016). Specifically, participants were instructed that they would command a platoon occupying an outpost at the edge of a village. They were provided with details including the outpost dimensions, its distance from the village, and the length of open ground extending from the outpost (OP). They were told that as a result of a major sandstorm, support (in the form of personnel, weaponry, or otherwise) would not be available. The participant was told to imagine that the send capability of the radio has been damaged and as such the PC could not speak to the section commanders. Finally, the participant was provided with recent information about defensive positions along the outpost walls and an inventory of ammunition. A simplified map of the OP and the village was also provided (Figure 1).
Simplified map of the outpost and village with north indication provided to participants for the experimental conditions.
Prior to the experiment, participants were given instructions on how to operate the messaging interface of the cell phone. Then they were presented with a practice scenario (10 messages) for familiarization with the visual and auditory interfaces, the SA report, and the DRT. The practice scenario differed from the experimental scenario in two ways: In the practice scenario, the same message sets were used for both visual and auditory formats, and all elements reported by the section commanders were considered to be enemy.
After the practice trials participants began the experiment, completing the four conditions in blocks of 30 messages each. In the auditory conditions, the recordings were played through the headphones connected to the laptop audio jack at 80% volume; during this time, the cell phone was placed on the table and not used. All four scenarios used the same terrain, but the locations and actions of enemy and friendlies differed. The experimental session took approximately 2 hr to complete.
Depending on their preferences, participants either recorded information on blank sheets of paper or the map provided. If they chose to use a map, they were provided with a new map on request.
After the 15th and 30th message (halfway and end point of each block, respectively), participants were asked to complete the SA reports. During this time, incoming messages and DRT stimuli ceased, following the SAGAT simulation halt procedure (Endsley, 1988). When filling out the SA reports, participants could refer to their own notes and the map but no longer had access to the visual messages or audio recordings. They were instructed to enter the most recent status information. Participants notified the experimenter after completion of the SA report, and the experimenter either resumed the scenario (if they were at the halfway point) or presented the participant with a NASA-TLX survey before proceeding to the next scenario (if they were at the end point).
For the DRT, the LED (a red light) lit up at random intervals between 3 s and 5 s. Participants were instructed to respond to the light as quickly as they could by depressing the micro-switch. The LED remained on until a response was made or 1 s had elapsed. RT was recorded with millisecond accuracy. Participants were instructed to treat the DRT as of equal importance to the primary task of processing tactical information, assume that the DRT represents a realistic task appropriate to the situation (e.g., monitoring a thermal sight), and perform both tasks to the best of their abilities.
Results
SA Report
An SA report answer key was created establishing the number of enemy groups for each SA report and the maximum number of correct answers for each of the participant’s eight SA reports (two for each of the four conditions). Answers were scored using a point system. For example, 2 points were assigned to a location response of North, 200 m: 1 point for North and 1 for 200 m. There were no deductions for incorrect answers. The scores were determined by the highest possible scoring matches between the answer key and the responses entered by participants. For each participant in each condition, the SA score was computed by summing the points obtained and dividing by the total number of points possible, expressed as a percentage. The mean SA scores are shown in Figure 2.
Proportion correct on the situation awareness measure, as a function of message presentation rate (MPR) and modality (auditory vs. visual). Error bars indicate 95% confidence intervals based on pooled mean square error for the within-subjects main effects in all graphs (Jarmasz & Hollands, 2009).
The SA scores were submitted to a 2 × 2 repeated-measures analysis of variance (ANOVA). There was a main effect for MPR: The slow MPR led to greater SA than the fast MPR, F(1, 23) = 64.03, MSE = 0.00954, p < .0001. There was also a main effect for modality: SA was greater in the auditory than the visual condition, F(1, 23) = 12.27, MSE = 0.01245, p < .005. The interaction was not significant, p > .45.
DRT Accuracy
For the DRT, the participant was required to press a button each time the LED light went on. A failure to respond was also scored as a miss. Following the ISO (2016) scoring guidelines, upper and lower cutoffs (2500 and 100 ms from stimulus onset, respectively) were used. That is, responses were scored as correct if the button press took place in the interval between the cutoff values. If the RT exceeded 2500 ms, it was scored as incorrect (a miss); if it was less than 100 ms, the response was considered premature and was discarded. For each participant in each condition, the number of correct responses was summed and divided by the number of signals to produce a DRT accuracy score. The mean DRT accuracy scores are plotted in Figure 3. There was a main effect for MPR: The slow MPR led to greater DRT accuracy than the fast MPR, F(1, 23) = 23.84, MSE = 0.01448, p < .0001. There was also a main effect for modality: DRT accuracy was greater in the auditory than the visual condition, F(1, 23) = 5.23, MSE = 0.01882, p < .05. The interaction was not significant, p > .40.
Proportion correct on the detection response task (DRT), as a function of message presentation rate (MPR) and modality (auditory vs. visual). Error bars indicate 95% confidence intervals.
DRT RT
A mean DRT RT was computed on the accurate trials for each participant. The mean DRT RTs are plotted in Figure 4. There was a main effect for MPR: The slower MPR led to a smaller RT, F(1, 23) = 28.24, MSE = 9831, p < .0001. There was a main effect for modality: RTs were smaller for auditory than visual presentation, F(1, 23) = 53.17, MSE = 5407, p < .0001. The interaction was not significant, p > .70.
Mean response time for the detection response task (DRT) as a function of message presentation rate (MPR) and modality (auditory vs. visual). Error bars indicate 95% confidence intervals.
NASA-TLX
We computed the mean of all six NASA-TLX subscales for each condition. These values are plotted in Figure 5. The Likert scale workload rating is represented on the ordinate (maximum value is 21 points). There was a main effect for MPR: Subjective workload was greater for fast than slow MPR, F(1, 23) = 28.85, MSE = 5.13, p < .0001. There was also a main effect for modality: Subjective workload was greater for the visual than the auditory modality, F(1, 23) = 6.42, MSE = 5.13, p < .05. There was no interaction, p > .25.
Mean subjective workload scores from NASA-TLX as a function of message presentation rate (MPR) and modality (auditory vs. visual). Error bars indicate 95% confidence intervals.
A repeated-measures ANOVA was conducted on each subscale. In general, the same pattern was obtained as for the combined subscale. For brevity, rather than consider each result in detail, we present the F and p values for each effect in Table 3. The effect of MPR was significant in every case. The effect of modality was significant for the temporal demand subscale but not otherwise. The interaction came close to conventional significance levels for the temporal demand and frustration subscales but was never significant, p > .05 in every case.
The F and p Values for Each NASA-TLX Subscale for Message Presentation Rate (MPR), Modality, and Their Interaction
Note. All reported F values have (1, 23) degrees of freedom.
p < .05. **p < .01.
Discussion
This experiment was designed to determine the joint influence of MPR and visual or auditory presentation of battle management information on cognitive load and SA. We predicted that the greater cognitive load under a fast MPR should degrade DRT performance relative to slow MPR and also increase NASA-TLX scores, and this prediction was confirmed. We also predicted that SA performance should be degraded with fast MPR relative to slow, and this prediction was also confirmed. Presumably, the fast MPR did not allow each message to be processed as well as with a slow MPR.
We were equivocal on whether auditory or visual message presentation would produce better SA performance and higher cognitive load, noting arguments for both possibilities. The results showed greater SA accuracy and reduced cognitive load (DRT performance more accurate and faster) in the auditory than the visual condition. This may have resulted from interference between the visually presented messages and the visual DRT. This auditory advantage occurred equally for both fast and slow MPRs (there was no interaction), so there was no support for the hypothesis that being able to read messages more quickly would be more effective (reduced cognitive load or improved SA) than auditory presentation under a fast MPR.
The DRT measures indicated that for both DRT accuracy and RT, the cognitive load with visual message presentation and a slow MPR was comparable to that obtained with auditory message presentation and a fast MPR. If there is more information being transmitted per unit time with a fast MPR, it suggests that auditory presentation allows a higher bandwidth of information transmission without an appreciable increase in workload relative to visual presentation.
The subjective workload (NASA-TLX) measures decreased with a slower MPR. Modality reliably affected subjective ratings for the temporal demand subscale only. Again, there was no evidence of interaction between the two factors.
Generally, the results for modality are consistent with Wickens’s (2008a) multiple resource theory. Wickens classified mental resources using a set of four dichotomies of information processing; one of these included a distinction between auditory and visual resources. Importantly, Wickens noted that “to the extent that two tasks use different levels along each of the . . . dimensions, time-sharing will be better” (p. 450). When the resource demands of a task exceed the supply, further demand will degrade performance. This might explain why we observed poorer performance with the visual presentation modality: Both the DRT and visual task competed for resources from the same visual resource pool. Future work should include the implementation of an auditory DRT to further explore task interference on the modalities dimension and see if auditory presentation would then suffer as a result.
An alternative explanation is that soldiers are simply more familiar with voice over radio (auditory message presentation) than text (visual message presentation). This could be examined by using an auditory version of the DRT and seeing if the current results persist. On a multiple-resource argument, the best performance should occur with text-based presentation, but on a familiarity argument, the best performance should occur with voice.
We noted earlier that additive results would be consistent with a model in which demand is independent of modality. The two manipulations may have affected different processing stages, which is another dimension of multiple resource theory. On this view, the manipulation of modality affects attentional resources (early processing), whereas the manipulation of MPR increases the demand on working memory (central processing), which would be consistent with independent effects. The manipulation of MPR would be seen to affect working memory load the same amount regardless of modality and as such is independent of the modality change.
Caveats
Experimenters must choose to hold certain factors constant while varying others. In this study, we focused on comparing cognitive load under visual and auditory conditions while also varying MPR. In simplifying the task, not all aspects of the experiment were fully realistic. Normally, a platoon commander would be able to communicate with section commanders, but this would have been difficult to simulate. To address this, participants were told to imagine that the send capability of their radio has been damaged. Similarly, it seems unrealistic to assume that urgent messages would be entered in text form rather than through an audio channel (e.g., “Receiving RPG fire!”). Nonetheless, to allow fair comparison across conditions, we did present such messages in visual text form. Further, a platoon commander using ISS might have the text messages available while preparing an SA report, but we did not permit our participants to do so to allow a fair comparison of visual and auditory conditions. Perhaps an alternative SA measurement technique like SPAM (situation present assessment method; Durso & Sethumadhavan, 2008) might enable an online SA assessment without the simulation halt.
To represent the way text information is presented on the ISS, messages were shown in a list, and the participant had to select a message to view it. The participant later needed to close the message to return to the list. In contrast, in the auditory conditions, messages were “pushed”; that is, they were simply played at the appropriate time. Thus, all messages in the auditory conditions were necessarily heard, whereas the messages in the visual conditions were not necessarily seen. A missed message would likely reduce the accuracy of a participant’s SA report, which could have reduced SA accuracy in the visual conditions.
Conclusions
Our objective was to examine factors affecting cognitive load and SA for a soldier in a tactical military situation. We saw that MPR affected both cognitive load and SA and also affected subjective mental workload. We also saw that the presentation of tactical information in auditory format both improved SA and reduced cognitive load.
In summary, we showed that not only does the rate of message presentation affect cognitive load, as predicted, but it also can adversely affect SA. In complex and dynamic environments such as those encountered in military settings, rapid transmission of information can quickly overload a soldier’s capacity, resulting in a loss of SA. However, we did not find that the visual presentation of information provided an improvement in SA beyond that afforded by auditory, or radio, messages. An unanswered question is whether redundant presentation of auditory and visual information (using speech recognition technology) might benefit beyond the auditory case. We might also find that some specific types of information are better transmitted as visual text (e.g., grid reference codes).
Beyond the military context, we believe that our results could generalize to other information-processing situations in which the environment is dynamic and complex, requires responsive action, and the information is presented in visual or auditory form. These might include air traffic control, medical operating rooms, or the trading floor of a stock exchange.
Key Points
There is potential for cognitive overload and a loss of situation awareness as battle management systems are introduced to dismounted soldiers.
We sought to determine the influence of message presentation rate and sensory modality on soldier cognitive load.
Participants (24 soldiers) served in the role of a platoon commander presented with tactical information about an unfolding scenario.
Workload was greater at faster message presentation rates and visual presentation (vs. auditory).
Footnotes
Acknowledgements
This work was conducted under the Defence Research and Development Canada (DRDC) 02ab Soldier Systems Effectiveness (SoSE) project. We thank Dorothy Wojtarowicz for help in coordinating and scheduling participants and Sakshaat Choyikandi for equipment setup and for designing and coding the messenger service on the Android device. We thank David Tack for initial discussions about the experimental scenario, Matt Lamb for coordination and planning, and Elaine Maceda and Nada Pavlovic for help running participants.
Justin G. Hollands is a defense scientist in the Human Effectiveness Section at Defence Research and Development Canada (Toronto Research Centre). He obtained a PhD in psychology from the University of Toronto in 1993.
Tzvi Spivak is a human factors engineer at Pacific Science and Engineering Group in San Diego, California. He obtained a BASc in industrial engineering from the University of Toronto in 2018.
Eric W. Kramkowski is a defense and security contractor with QTAC Inc., in Toronto. He has served 18 years with the Canadian Armed Forces and is a student of kinesiology in a bachelor’s program at York University.