Situation Awareness: Does It Change with Age?

Proceedings of the Human Factors and Ergonomics Society Annual Meeting http://pro.sagepub.com/

Situation Awareness: Does it Change With Age? Cheryl Actor Bolstad Proceedings of the Human Factors and Ergonomics Society Annual Meeting 2001 45: 272 DOI: 10.1177/154193120104500401 The online version of this article can be found at: http://pro.sagepub.com/content/45/4/272

Published by: http://www.sagepublications.com

On behalf of:

Human Factors and Ergonomics Society

Additional services and information for Proceedings of the Human Factors and Ergonomics Society Annual Meeting can be found at: Email Alerts: http://pro.sagepub.com/cgi/alerts Subscriptions: http://pro.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://pro.sagepub.com/content/45/4/272.refs.html

>> Version of Record - Oct 1, 2001 What is This?

Downloaded from pro.sagepub.com by Cheryl Bolstad on April 23, 2012

272

PROCEEDINGS of the HUMAN FACTORS AND ERGONOMICS SOCIETY 45th ANNUAL MEETING- 2001

SITUATION AWARENESS: DOES IT CHANGE WITH AGE? Cheryl Actor Bolstad* North Carolina State University Raleigh, North Carolina *Now at SA Technologies,

Inc.

The aim of this research project was to study age-related differences in our ability to perceive and abstract important information from the environment and to determine what physical and cognitive components are related to this ability. Young, middle and older adults completed several questionnaires, a battery of psychological tests, a standardized vision measure as well as several trials using a realistic driving simulator. A concurrent memory probe technique was used to measure participant's ability to attend to important information while driving. Their probe answers were checked against the actual simulator data and a composite situation awareness (SA) score was created. Results confirm the hypotheses that older adults have lower situation awareness when compared to younger and middle-aged adults. Factors that are related to this ability include useful field of view (UFOV), perceptual speed, driving experience and self. reported vision. INTRODUCTION During this century, we have been wimessing the "graying of America," with the average age of the population steadily rising (Moody, 1994). Along with this aging population is an increase in the number of licensed older drivers, In general, good drivers are more aware of their surroundings, in part, because the ability to attend to important and necessary information while driving is essential for good driving performance. This ability is embedded in the construct of Situation Awareness (SA), which is defined as the internal conceptualization of the current situation (Endsley, 1997). I believe a person's ability to attend to important information during driving is related to the work on SA. Borrowing from the research in SA, we need to know if the driver is capable of perceiving, comprehending and understanding all that is around them, and if they are able to guide their future actions based on their understanding of the current situations. Lastly, does this ability decline with age? Bolstad and Hess (2000) believe that age differences in SA will be greatest during its initial formation: in the ability to perceive what is important in the surrounding. For instance, while driving a car, such important information can be lane position, speed and location of other cars. If drivers are not attending to the necessary information, it seems logical that performance will ultimately suffer. However, SA is closely related to experience. Thus, it seems likely that an experienced older driver in certain situations and with certain tasks can perform as well as younger driver. There may come a time; however, when the experience will no longer insulate the driver from performance decrements. We can hypothesize that several cognitive and physical abilities are needed for the ability to attend to important information. These include vision, perception, memory and attention. We know that these abilities change with age (Laux, 1995; Salthouse, 1985; Smith & Earles, 1996) and are more pronounced as the task difficulty increases (Tun & Wingfield, 1997).

We can therefore speculate that the ability to attend to important information (initial formation of SA) will also change with age due to changes in these cognitive abilities and these differences will be more apparent as task complexity increases. Therefore, older drivers that have deficits in certain cognitive and physical abilities will also be deficient in their ability to attend to important information. The purpose of this study was to gain a better understanding of how normative aging changes affect a driver's ability to attend to important information in the driving environment and to determine cognitive factors that are related to this ability. A concurrent memory probe technique was used in this study to measure SA (the ability to attend to important information while driving). This research tested the efficacy of this technique to discriminate between different aged individuals. METHOD Participants Drivers in three different age groups were recruited for this study. The young adult group consisted of 16 participants (]_ age 19.5, range = 16,25, M driving experience = 3.2 years). The middle-aged adult group consisted of 16 participants (M age 45.1, range = 40 - 50, M driving experience = 29. I years), and the older adult group consisted of 16 participants (M age = 70.4, range = 65 - 80, M driving experience = 54.1 years). Design A 3 x 2 x 2 mixed factor design with age (young, middle and old), trial and complexity level (moderate and high) was used. Complexity level and trial were within group variables, such that all participants drove in the moderate and high complexity scenarios.



273

Simulator Trials

Situation Awareness Measure

This research used a PC-based, high fidelity, fully interactive driving simulator called STISM (Systems Technology Interactive Simulator). The simulator provided both immediate visual and auditory feedback to the user as well as steering torque. It contained driving hardware, a steeringwheel and brakesystem (see Figure 1). Three 1l-rain test trials and one 15-rain practice simulator trial was created for this study. The test trials contained both moderate and high complexity conditions as well as a one-minute low complexity scenario. The practice trial began with 5 minutes of low complexity followed with 5 minutes of moderate complexity and 5 minutes of high complexity. Complexity was manipulated by varying the content of the scenarios. In the low complexity condition, the scenario was a four-lane road with no other traffic, buildings, turns or pedestrians in the scene (see Figure 2). In the moderate complexity condition, the scenario contained a nominal number of other cars, buildings and turns (see Figure 3). In the high complexity condition, the road was also a fourlane road, but it contained 4 times the amount of traffic, buildings, turns and intersections with crosswalks, It was hypothesized that SA performance would decline from the moderate to the high complexity scenario and

The first partof the study involved the development of the concurrent memory probe. The queries used in the probes were created using a cognitive task analysis with the aid of experienced drivers. The analysis occurred in two stages; the first stage used unstructuredinterviews and the second stage used a goal-directedtask analysis. Unstructured interviews were conducted with five experienced drivers (_M_M driving experience = 18 years). They were asked, "What would you want to know to have perfect driving performance?" The goal-directed task analysis was created to assure that the information obtained through the unstructured interviews was complete and representative of the requirements needed for good driving behavior. Two experienced drivers performed the goal directed task analysis. From these two analyses a structured questionnaire was created that contained a comprehensive list of possible important driving elements, such as weather, speed and lane position. Twenty-five participants (_M_ age = 55.0 and M__ driving experience = 38.8 years) rated the importance of these items in terms of good driving behavior. The items were compared to the simulator data output list and the nine highest ranked items that could be collected by the simulator were used to create the situation awareness queries (see Table 1 ).

than the other age groups in this study.

older adults would exhibit greater decreases in performance

iliii

i_iii_ili}i_iii

iiiiiiii!iiiiiiiiii ............ ill

Figure 2: Low Complexity Scenario.

iiiiiiiiii!iiiiiiiiiiiii

'"_ii

ii

ilii _'?'i'i'iii_'i!iiiii'iii' _%_;_:=_i_::_::_:_:::::_::_::_:::_;:_:_:,_::_: 'i"ii"i""_"_'"i""i""iiiiiiiiii i i;I '_i'_

Figure 1"

etup.

Figure 3: Moderate Complexity Scenario.


274


_: i:" S_.-__O_AW_NES_ QUERIES ......... = How fast is the closest driver in front of you going? A. Fasterthan me B. Slower than me C. About the same speed D. No cars in front What is currentcolor of the closest traffic signal in front of you? A. Red B. Yellow C. Green Can you legally get through the next intersectionwithout stopping? A. Yes B. No C. No intersection in sight Are you currentlytravelling? A. Above the speed limit B. At the speed limit C. Below the speed limit ...... T_ible ii Sampie Situation AwarenessQueries.

workstation to have their UFOV measuredusing the Visual AttentionAnalyzer. After the UFOV assessment, participantsreceived a half hour of trainingon use ofthe STISM simulator,including the situationawareness measure. During the training session, participantscompleted the 15 rain practicetrial. This trial began with the no complexity scenario and increasedin difficulty. Participantswere instructedto follow all the roadwaysigns (speed limits, stop lights) and warnings,as they would in the actualworld. After training,participantsdrove in the three 11 min driving trials.Twice during each complexity level (4 times per trial)-- 2 1/2 and4 I/2 minutes into the scenario--the trialwas stopped to measure their situationawareness. Right before each stop, the experimenterquickly covered the computer screen with black poster board andfroze the simulation. The quick movement was used to reduce the effect of visual after images. The stops took no more than 2 rain andparticipants answered all nine queries with regardsto what was important in their surroundings. During each probe stop,the queries were randomized so the participantscould not anticipatethe queries. Upon completion of these queries the simulation was resumed.

Background Measures RESULTS Background Questionnaire. A background questionnaireand a standardizedhealth survey (SF-36; Ware, 1993) were used to collect data on health, age, driving abilities and accidentrates. Some of the data from these questionnaires was used to gather descriptive data on the participantswhile other data was used to determineif self-report driving data was related to a person's situationawareness, UFOV. The Visual AttentionAnalyzer was used to measureUseful Field of View (UFOV) (Owsley, Ball & Keeton, 1995). The Visual Attention Analyzer instrument consisted of a computer display with white targets againsta black backgroundto provide for high contrast. Three different subtasks are used to measure UFOV: 1) Speed of information processing, 2) Divided attention, and 3) Selective attention. Perceptual Speed. Salthouse and Coon's (1994) letter comparison task was used to measure processing speed, Dynamic Working Memory. Dynamic working memory was measured using the Weschler Adult Intelligence Scale- III (WAIS-III). Dynamic working memory is dependent on the role of the central executive (Morris & Jones, 1990). This component is one of three that Baddley and Hitch (1974) propose make up working memory. Baddley (personal communication, 1997) recommended using the {WAIS-III) letter-number sequencing test to measure the central executive memory component,

PROCEDURE Testing took place in one day over a two-hour period. Participants completed an informed consent sheet and a background questionnaire. This was followed by a perceptual speed and dynamic working memory task. Once the paper and pencil tests were completed, participants moved to a computer

Participantsin this study were presentedwith three driving trials with 4 stops per trial(2 atthe moderate complexity level and 2 at the high complexity level) for a total of 12 concurrentmemory probe stops. During each of these stops, participantsanswered a set of nine queries. The answers to these queries were compared againstthe simulatordata file to determine what was actuallyhappening in the condition at the momentof the stop. Participantsreceived a score of either 1 or 0 depending uponthe correctness of their answer. The scores for all nine queries were then summedand averaged within each complexity level and trial for a total of six situationawarenessscores per participant. Participants were encouraged to complete all three trials of the driving task. Many older adults could not completely finish the task due to several uncontrollable events, including simulator sickness and overrunning the allotted test time. Due to these problems and in order to optimize the power of the statistical tests to identify the anticipated effects, only the f'wsttwo trials were included in the following analyses. Therefore, the analysis was performed using 4 situation awareness scores for each participant. Concurrent Memory Probe Performance (Sit). An Age Group X Trial X Complexity ANOVA was performed on the situation awareness score. As expected, there was a significant age group effect, _F(2, 42) = 3.39, t2= .043 (see Figure 4). A follow up Tukey HSD test revealed that older adults reported significantly less information (M = 60.51) than younger (M = 66.23) and middle-aged adults (M = 68.85) (l_S< .042), whose performance on the probe did not differ from each other.


275


Factors Affecting Performance. in order to iii_i 8¢ _

........... i:

iii-_ ......... _= _ 6_ ....... ...........

:

determine what factors have an effect on an individual's situation awareness, correlations were calculated between the

,_

7_

baekgrotmd probe score. measures It was hypothesized and the average that concurrent all measuresmemory would

• _ _

69 5_ 4_ 3_ _ I0 _

have a smallconditions, correlationand withasperformance moderate complexity performance inthe demands increased (high complexity scenario), the correlations would also correlated increase.However, nomeasures weresignificantly withperformance usingtheconcurrent memory probemeasure from the moderate complexity condition (see Table 2) In the high complexity condition, four measures were moderately correlated with probe performance: age group (-.41), UFOV divided attention task (-.32), UFOV selective attention task (-.40) and driving experience (-.40) (see Table 2). However, onlythecorrelation withtheUFOVselective attention task was significant (12= .048, one-tailed). None of the other tests that were hypothesized to be predictors reached a significant level.

¥_g

_dd_e

O_der

Figure 4: Percent of Queries Answered Corre_tly By Age Group. Overall, all participants answered fewer queries correctly when driving in high complexity conditions than in the moderately complex conditions, F(1, 42) = 85.79, tZ= .029. However, the Age Group X Complexity interaction was not significant, F(2, 42) = .37,12= .696, with the impact of

M__i I _GH_ I]ii!iiiiiii_AS__,_=,_i=11,i CO_LEX_y_I=It iCO_LEX_Ti_==i i,,]t

performance oftheolderadults,middle-aged adultsand younger adults did decline from the moderate complexity to the high complexity conditions. Whereas the difference in probe performance was in the hypothesized direction, older adults did not experience a greater decrease in probe performance than the other two groups. A significant trial effect, F (i, 42) = 5.12, p_= .029,

-" Gender Vision (se_f,report) _Medicat [_ProNems Dr_v!ng

wasthepresent indicating that with performance of participants improved (M practice, = 63.4 vs.probe M = 66.55) from trial 1 to trial 2. A significant Age Group X Trial interaction, F__ (2, 42) = 6.70, p_=.003, was also present (see Figure 5).

i Accident Number ................ 'Ticket (,_Umber Miles Driven ]

When the interaction was examined using a Tukey HSD test, the younger adults exhibited significant improvement performance from trial 1 to trial 2 (12< .001), whereas in theprobe

WM$ [ Perceptual S_¢d ] _ : ,: .. LTOV-Percept_al - UFOV*Divided ' UFOVoSe_ecfive

differences between trials 1 and 2 for both the middle-aged and older adults were not significant. This indicates that there was a significant practice effect only for the young adults as the performance of the middle-aged and older adults remained stable between trials.

......

: 1_. _[} 7_:

_

_

6_

s_.36 _en_al

.30"_ _ -.30"

-.04 -26

-.07

.09

-.21

-.40"

- "16 -,05 .21

'.24

i

_

..

SF"36 Physical *significant, I_< f_;

tt .3i*

-.01 .11 .t .26

-.02 -._7 "23

-.20 -.32" -.40"

i02

-.t 9

.t4

-.12

-.09

.03 "

Table 2: Correlationsof the BackgroundVariables with the :_-,]:iiiiiiii_ii:iliiiiii _'.

::i_

- _TOV-R_sk

:

___

ii!i?i:i_(:.:i! i_ii 2

__

Complexity. Percent of Correctly Reported Information By Task

20:

_........ iii._i-_iii:_i:_i_i_i-_i:i-_i_i.ii_;::::::::::::::::;:::::::::::::::::::::::::::::: _ _ Overall, the results from this study support the : _ _:_ .... primary hypothesis that older adults have lower SA, as .. :: ,_:-_:_:_:_:_ .... :_ ,,:-