Brown et al., 1969

Test track / instrumented vehicle

n/a¹

Cognitive²

Kames, 1978

Test track / instrumented vehicle

CON & IP³

none

Drory, 1985

Driving simulator

n/a

Cognitive

Stein et al., 1987

Driving simulator

CON & IP

Cognitive

Boase et al., 1988

Laboratory computer game

n/a

Simple / intense conversation

Zwahlen et al., 1988

Test track / instrumented vehicle

CON & IP

none

Hayes et al., 1989

On-road / instrumented vehicle

CON & IP

none

Alm & Nilsson, 1990

Driving simulator

IP

Cognitive

Brookhuis et al., 1991

On-road / instrumented vehicle

?⁴

Cognitive

Fairclough et al., 1991

On-road / instrumented vehicle

?

Simple / intense conversation

Nilsson & Alm, 1991

Driving simulator

IP

Cognitive

Kantowotz, et al, 1996

Driving simulator

IP

Simple task / Cognitive

Parkes, 1991

Laboratory, On-road/ instrumented vehicle

?

Cognitive

Green et al., 1993

On-road / instrumented vehicle

CON

Simple task / conversation

McKnight & McKnight, 1993

Driving simulator

IP

Simple / intense conversation

Nilsson, 1993

Driving simulator

IP?⁵

Cognitive

Serafin et al., 1993

Driving simulator

IP & HUD⁶

Cognitive

Pachiaudi & Chapon, 1994

Driving simulator

?

General conversation

Alm & Nilsson, 1995

Driving simulator

IP

Cognitive

Tijerina et al., 1995

On-road / instrumented vehicle

IP

Simple task / cognitive

Alm, H., & Nilsson, L. (1990). Changes in driver behaviour as a function of handsfree mobile telephones: a simulator study (DRIVE Project V1017, Report No. 47). Linkoping, Sweden: Swedish Road and Traffic Research Institute.

Type of Study: Driving simulator.

Keywords: hands-free cellular telephones, route difficulty, lateral position, workload, speed level, working memory span, subjective measures

Author's Abstract:

The effects of a mobile telephone conversation on drivers reaction time, lane position, speed level, and workload in two driving conditions (easy versus hard driving task) were studied in an advanced driving simulator. 40 subjects, experienced drivers in the ages 23 to 61 years, were randomly assigned to four experimental conditions. It was found that a mobile telephone conversation had a negative effect on drivers reaction time, when the driving task was easy. It led to a reduction of speed, when the driving task was easy. It had a negative effect on drivers' lane position, most pronounced when the tracking component of the driving task was hard. Finally, it led to an increase in workload for both the easy and hard driving task. The effects were discussed in terms of what subtask, car driving or telephone conversation, the drivers gave the highest priority. Some implications for information systems in future cars were discussed.

Sample and Methods:

• Drivers (20 male, 20 female) who had been licensed at least five years and who drove at least 10,000 km per year. Ages ranged from 23 to 61 (mean age 32.4).
  
  
• Subjects were randomly assigned to one of four experimental conditions:

  "easy" (straight road) driving

  "easy" driving and telephone task

  "hard" (curvy road) driving

  "hard" driving and telephone task
  
  

• The VTI Driving Simulator was used for the study.
  
  
• All subjects were given practice in the driving simulator. Those subjects in the telephone task conditions were also given practice with telephone calls.
  
  
• The cellular telephone used was an Ericsson Hot Line device mounted at the height of the steering wheel on the instrument panel.
  
  
• During testing, subjects in the telephone conditions had to answer an incoming call by pressing a telephone button which simulated a hands-free telephone operation. The telephone was mounted at the height and to the right of the steering wheel on the instrument panel. After answering, they then had to perform the cognitive task.
  
  
• Subjects in the two control conditions were only required to drive.
  
  
• The telephone task consisted of the Baddeley Working Memory Span Test where subjects had to answer "yes" if a sentence was sensible and "no" if it was nonsense. Each call to the driver on the telephone started with task instructions, followed by the presentation of five sentences. This task was paced.
  
  
• After all five sentences were read, subjects had to repeat the last word in each sentence in order of presentation.
  
  
• Each presentation (simulated telephone call) took roughly 60 seconds.
  
  
• A red square was used as a visual stimulus, acting as a hazardous object in the driver's path. Drivers were to brake as fast as possible whenever the red square appeared.
  
  
• Vehicle speed, lateral position, and brake reaction time to the visual stimulus were used as performance measures.
  
  
• Subjective measures of workload as measured by the NASA-TLX, and communication measures of the number of correct sentence judgments and the number of correctly recalled last words were also gathered.

Major Findings:

• For the easy (straight) route, drivers' brake reaction times were longer in the telephone task condition than in the control condition, while for the hard (curvy) route, drivers' reaction times were not significantly different in the telephone and control conditions.
  
  
• During the first distance measured (0-500 meters), drivers initiated the hands-free function to answer the mobile telephone.

  In the easy driving condition, no significant differences in mean lateral position were found between experimental and control groups.

  In the hard driving condition, mean lateral position increased when calls were received, compared to the control condition. In addition, when the randomly timed calls occurred during straight sections of the curvy road, mean lateral position decreased.

  Differences in lateral position for the control groups only in the easy versus hard driving conditions were not reported.

  The second distance (0-2,500 meters) measured driver performance during the entire period of the telephone task for each call.

  In the easy driving condition, drivers' mean lateral position was greater when using the telephone than in the control condition.

  In the hard driving condition, drivers' mean lateral position was greater when using the telephone than in the control condition. In addition, during calls, curvier sections of the road led to greater variation in lateral position.

  Differences in lateral position for the control groups only in the easy versus hard driving conditions were not reported.
  
  

• Subjective workload was measured by the NASA-TLX rating scales.

  Compared to the control groups, drivers who received calls had increased scores on the factors "mental demand," "physical demand," "time pressure," "operator performance," "operator effort," and "frustration level."

  An interaction between receiving calls and route was seen, so that drivers were more frustrated during telephone use, and this effect was further influenced by route difficulty.

  The hypothesis about higher workload due to telephone use was supported, but the hypothesis that workload should increase with the complexity of the driving task was refuted.
  

• Speed level was measured from the onset of receiving each telephone call and 80 seconds forward.

  Speed was lower for the drivers who used the telephone than for the control groups. For the easy route, speed was lower when the driver had to use the telephone. However, on the hard route, the difference in speed between telephone and non-telephone conditions was not significant.

  An interaction was seen between brake reaction time and route. For the easy route, the telephone task increased the subjects' reaction time to the visual stimulus. For the hard route, no differences were seen between the telephone task and control groups for brake reaction time.

  No differences in telephone task performance were found between the driving task complexity conditions.
  
  

Author's Conclusions:

• When the driving task was easy, a telephone task had a negative impact on drivers' ability to react quickly, but when the driving task was hard no negative impact was found.
  
  
• The demands of the hard-driving task may have induced drivers to concentrate more on the driving task, giving the telephone task secondary status, which limited its influence on drivers' behavior. In the easy driving task, drivers may not have had to allocate as much attention to the driving task, leading drivers to give the telephone task primary status.
  
  
• This explanation is supported by better memory retrieval scores during the easy task and greater frustration scores from the NASA-TLX during the hard driving task.
  
  
• The introduction of a non-driving task can have different effects, depending upon what priority drivers give the non-driving task.
  
  
• If the driving task is perceived as easy, the non-driving task may be treated as the primary task, and this may have negative effects upon drivers' ability to react quickly in an emergency.
  
  
• While a secondary task may have an alertness-arousing effect, this increase in alertness may not necessarily have a positive effect on driving. Instead, it may sometimes be used to improve performance on the secondary task, especially in easy driving conditions.
  
  
• The finding of increased physical workload scores may suggest that activation of the hands-free button should be improved, possibly by marking the hands-free button more clearly.
  

Critical Assessment:

• This study presents a dynamic picture of how drivers' priority, the road, and the secondary tasks mutually influence one another.
  
  
• Reaction time measures should also include a decision component following the detection of the stimulus. In a driving situation, the driver must not just detect the stimulus, but then he/she must make an appropriate response. Rarely is a single response the only appropriate response in a driving situation. The response (reaction) time is heavily influenced by the amount of uncertainty in the environment at the time of the stimulus (e.g., Hick-Hyman law). Traffic complexity should be a factor in the determination of driver reaction time.
  
  
• This study did not consider the influence of different types of conversations upon driver performance, which has been shown to have a significant influence in other studies. Furthermore, the "conversational" materials were demandingin terms of reasoning and memory requirements. This could induce a greater attentional distraction than that associated with normal cellular phone conversations. the paced nature of the taskswould also promote attention to the conversation rather than on the driving task. Presumably, paced dialogue is rare in normal cellular phone conversations.
  
  
• Drivers only received calls during driving. The act of placing a call is probably more demanding than receiving and should be a part of the experimental protocol.
  
  
• This study utilized a between subjects experimental design, comparing a group of drivers who performed the treatment task to those who did not (a control group). However, we cannot be certain that individual differences between the two groups were not present, which could explain the study's results. It is not known whether drivers in the two groups were matched to one another, or if other experimental control procedures were used to equate the individual characteristics of the drivers more closely, such as driving ability, memory recall, or simple reaction time.

Alm, H., & Nilsson, L. (1995). The effects of a mobile telephone task on driver behaviour in a car following situation. Accident Analysis and Prevention, 27(5), 707-715.

Type of Study: Driving simulator.

Keywords: hands-free cellular telephone, simulator, reaction time, mental workload

Author's Abstract:

The effects of a mobile telephone task on young and elderly drivers' choice reaction time, headway, lateral position, and workload were studied when the subjects were driving in a car-following situation, in the VTI driving simulator. It was found that a mobile telephone task had a negative effect upon the drivers' choice reaction time, and that the effect was more pronounced for the elderly drivers. Furthermore, the subjects did not compensate for their increased reaction time by increasing headway during the phone task. The subjects' mental workload, as measured by the NASA-TLX, increased as a function of the mobile telephone task. No effect on the subjects' lateral position could be detected. Taken together, these results indicate that the accident risk can increase when a driver is using the mobile telephone in a car following situation. The reasons for increased risk, and possible ways to eliminate it, are also discussed.

Sample and Methods:

• Forty subjects, 30 men and 10 women, participated in the study.
  

  The subjects were divided into two groups, younger drivers (<60 years of age), and older drivers ( 60 years of age).
  
  

• Apparatus:

  The VTI driving simulator was used for the study, with a Saab 9000 with manual gearbox used for the car body.

  The mobile telephone used was an Ericsson Hot Line device with hands-free facility, mounted on the instrument panel to the right of the steering wheel.
  
  

• Participants drove an 80 km test route where they were forced into a car following situation 16 times during the test session.

  Subjects in the experimental condition were exposed to a telephone task during eight of the car following situations.

  In four of the telephone tasks, something safety critical occurred on the roadway.

  Subjects in a control condition did not receive any telephone tasks.
  
  

• The Baddeley Working Memory Span Test was chosen as the telephone task where subjects had to determine whether sentences were sensible or nonsense. After 5 sentences were read, subjects were asked to recall the last word in each sentence, in order of presentation.
  
  
• Choice reaction time to the lead vehicle braking was used as a performance measure, as was headway and lateral position. Subjects were instructed to brake as fast as possible in response to the lead vehicle's braking.

Major Findings:

• Subjects in the experimental groups had a longer reaction time compared to subjects in the control groups.
  
  
• An effect was also present for age. Older subjects had longer reaction times than did younger subjects.
  
• Subjects in the experimental groups had a shorter minimum headway compared to subjects in the control groups.

  The minimum headway also differed between age groups. Elderly subjects averaged a minimum headway of 50m, while younger subjects averaged a 42m headway.
  
  

• Average headway only differed between age groups, with older drivers maintaining a larger headway than younger drivers. Subjects did not compensate for an increased reaction time by increasing their headway.
  
  
• No significant difference was found between experimental and control groups for lateral position during the eight telephone tasks.
  
  
• Using the NASA-RTLX, subjects in the experimental groups rated mental demand, time pressure, effort, and frustration as higher compared to subjects in the control groups.

Author's Conclusions:

• The fact that drivers in the experimental groups did not compensate for increased reaction time with increased headway can be interpreted in two ways.
  

  First, subjects may have been unaware of the impact of the mobile telephone task on their ability to react quickly.

  A second explanation may be that subjects already had a large enough headway during the telephone task to compensate for decreased reaction time.

   

• The mobile telephone task led to an increased level of mental workload.
  
  

  It is often argued that a driver can control the demands of the driving task, within reasonable limits. However, this study shows that drivers could not control the demands of the driving task, in terms of keeping the workload constant.

  Driving in a car-following situation, and using the mobile telephone concurrently, may increase the risk of an accident if something unexpected happens.

Critical Assessment:

• The use of a simulator may have influenced driver behavior. Because drivers were not in any real danger, they may have reacted differently than if they were in a real-life situation.
  
  
• Answering questions about whether a sentence is sensible or nonsense is not a realistic task for using a mobile telephone and consequently measuring workload.
  
  
• A key factor in this study was that the telephone tasks occurred while the driver was in a car following situation. However, this is not the only road situation which may occur. To extend the validity of this study, other scenarios should have been incorporated such as performing the telephone task during a lane change or other complex maneuver.

Boase, M., Hannigan, S., & Porter, J. M. (1988). Sorry, can't talk now... just overtaking a lorry: The definition and experimentation investigation of the problem of driving and handsfree carphone use. In E. D. Megaw (Ed.). Contemporary ergonomics (pp. 527-523). London: Taylor and Francis.

Type of Study:

-Structured interview of hands-free cellular telephone users (5 executives and 4 salesmen)

-Laboratory study using a computer game as the surrogate (driving) task (with a different group of participants).

Keywords: hands-free cellular telephones, dual tasks, self-assessments, age differences

Author's Abstract:

This study investigated the interactional effects of [simulated] driving and talking using a hands-free cellular telephone. Nine hands-free users were interviewed to provide preliminary information for the design of a laboratory study. The experimental data show that the quality of a complex dialogue suffered at higher simulated driving workloads. Simulated driving performance was also adversely affected with both simple and, to a greater extent, complex dialogues. These findings contrast with the comments from many users who state that their driving does not suffer because they have the option to reduce driving speed, to not answer a call, or to stay silent during a conversation.

Sample and Methods:

• 24 drivers, 6 young males, 6 young females, 6 older males, 6 older females. Age ranges not given.
  
  
• Participants played a "squash" type computer game involving tracking, prediction, and some decision making. The game was scored by tallying the number of balls hit as a percentage of total hits possible in the 90 second periods before and after the start of the dual task situation (omitting the 30 second band in which the telephone call began).
  
  
• Workloads within the squash game were calibrated prior to the experiment according to each subject's ability, which was determined after their "driver" performance on the surrogate task had reached a certain criterion level.
  
  
• An experimenter was present beside the participant during the testing sessions.
  
  
• A hands-free telephone was used but the model was not specified by the authors.
  
  
• Two participants were involved at a time. One operated the computer game while the other called the first participant five times. Callers played the role of an office worker and engaged drivers in five simple Information Dialogues (IDs), e.g., asking about favorite foods or past education. The first call occurred before task performance, the next three occurred during task performance and were timed to coincide with high, low, and mixed (both high and low) workloads, and the last call occurred after task performance.
  
  
• The entire task was then repeated with the experimenter taking the role of the caller and making calls to engage the driver in more difficult Negotiation Dialogues (NDs). The six NDs involved situations such as returning faulty goods to a shop or having a holiday booking altered by the company.
  
  
• At some point, the second participant, i.e., the caller in the ID condition, switched roles and received calls. (It is unclear whether this happened before or after the ND condition began, or what the second participant did while the experimenter became the caller in the ND condition).

Major Findings:

 
• Task performance, as measured by percent of balls missed, deteriorated significantly (approximately 11%) at the high and mixed workloads, when conversation was introduced.
  
  
• The simple information dialogues produced a greater decrement in missed balls (14.5%) than did the more difficult negotiation dialogues (6.5%).
  
  
• Age differences showed decrements in task performance (percent of balls missed) in the dual task situation, with older participants performing significantly worse in the mixed workload condition.
  
• Recordings indicated that conversation length and pause length increased significantly in the dual task situations.
  
  
• Comparing post-task questionnaires to game scores showed that participants accurately assessed their own decreased game playing ability while conversing.

Author's Conclusions:

• Simulated driving performance [i.e., task performance] deteriorates significantly at the high and mixed workloads when conversation is introduced.
  
  
• While structured interviews with a separate population (5 executives and 4 salesmen who used hands-free cellular telephones) showed that drivers believe that their driving does not deteriorate when using the hands-free cellular telephone, the laboratory study revealed that participants perceived and stated that their game-playing ability deteriorated.
  
  
• The immediate feedback of missing squash balls may have shown experimental participants the extent of their performance deterioration while roadway drivers have no such clear-cut indication of how their driving performance in a car may be suffering.

Critical Assessment:

• Simulated squash playing while using a hands-free cellular telephone is not a good representational task of actual driving while using a hands-free cellular telephone. Since the safety of playing squash while using a hands-free cellular telephone is not really an issue, construct validity is questionable, and the results may not tell us much about the effects of cellular telephone operations or conversations on driving.
  
  
• Why omit the 30 second band in which the call began? It would be useful to know how participants performed while initially adjusting to the new dual task situation.
  
  
• The study did not consider participants' previous squash-playing skills or familiarity with computer games. It is likely that an age difference in these skills obscures any age difference found in task performance.
  
  
• In the ID situation, using participants as callers allowed too little experimental control over the conversation and may have aided the callers' ability to handle the dual task once it became their turn to "take the wheel."
  
  
• It does not appear that the ID and ND situations were counterbalanced across participants (only the presentation order of WL on the squash game was counterbalanced). Therefore, learning effects may explain the better task performance in the ND situation since skill on the computer game could probably be learned very easily.
  
  
• The authors do not address the finding that performance decreased less with NDs than with IDs, which were supposedly easier. Other research suggests that driving performance may be more impaired with more intense conversation.
  
  
• Additionally, the authors did not provide a rigorous justification regarding the equivalence of the computer game to driving. In this regard, the study's relevance to real-world driving is suspect.

Briem, V., & Hedman, L. R. (1995). Behavioural effects of mobile telephone use during simulated driving. Ergonomics, 38(12), 2536-2562.

Type of Study: Laboratory study using pursuit tracking task as surrogate driving task.

Keywords: hands-free mobile telephones; driving performance, gender differences.

Authors' Abstract:

The effects on driving performance of using a hands-free, mobile telephone were investigated in a pursuit-tracking task that simulated driving. Twenty subjects in two age groups, 19-26 years (median=21 years) and 40-51 years (median 45.5 years) participated, with five males and five females in each group. The primary task was driving safely. The subjects drove for 20 min in each of three secondary task blocks with (i) simple telephone conversation about a familiar topic, (ii) a difficult telephone conversation, incorporating a test of working memory, and (iii) car radio tuning and listening. Half of the driving was done on a simulated firm road surface and half on a slippery road surface. The subjects' behavior was subsequently observed and classified in four activity categories, two without and two with a secondary task, with driving (i) on a clear road, and (ii) with obstacles, and with driving involving the secondary task components of (iii) communication, and (iv) instrument manipulation. The results show different patterns of driving performance on the two road surfaces. For driving on the slippery road, a deterioration was especially marked during manipulation of the instruments, in particular the radio, which required more prolonged manipulation than the hands-free telephone. Driving during an easy telephone conversation was associated with the least performance decrement, and could, in some cases, be seen as facilitatory. The female subjects tended to perform less well than the male subjects while driving on a slippery road. Some of this difference could be attributed to less previous driving experience. In general, the males drivers exhibited better control while driving under difficult conditions. There was no difference in driving proficiency between the age groups. It is concluded that simply conversing over a hands-free telephone while driving does not in itself impair performance. However, a difficult conversation may affect the driving adversely, and any prolonged manipulation of the telephone is liable to produce a performance decrement, particularly under conditions that put heavy demands on the driver's attention and skill.

Sample and Methods:

 
• Twenty (20) test participants in two age groups, 19-26 years (median=21 years) and 40-51 years (median 45.5 years) participated, with five males and five females in each group.
  
  
• A pursuit tracking task said to simulate driving was used as the primary task. In this task, test participants watched a computer monitor that displayed a curved line (the "road") and a triangle (the "car") on the upper half of the screen and an analogue speedometer on the lower half of the screen. The task was to keep the triangle on the line by means of a steering wheel, control speed with a pedal, and avoid obstacles (red horizontal bars that appeared next to the road and sometimes intruded). Various road signs (speed signs, phone-now signs, slippery road signs, end-of-slippery-road signs, and tune-radio-now signs) appeared on the screen as well. Firm road and slippery road conditions were simulated by applying velocity and acceleration dynamics, respectively, to the steering wheel controller.
  
• Test participants "drove" for 20 minutes in blocks associated with each of three secondary task conditions. A "radio" task required turning the radio on, tuning into and listening to at least four channels with a quick-tune (but not preset) facility, then turning the radio off. "Easy" telephone conversations were 2 minute dialogues about current events (War in Bosnia, Child prostitution in Thailand, Clinton's economic policies in the US, and the unemployment situation in Sweden). "Difficult" telephone conversations were 2-minute working memory tests similar to those used by Alm and Nilsson (1990).
  
  
• Within the 20-minute blocks, firm and slippery road surface conditions were presented in alternate 5-minute sub-blocks. During half of each 5-minute sub-block, the test participant engaged in the assigned secondary task. During the other half of each 5-minute sub-block, the test participant either drove only or drove only with obstacles that had to be avoided (i.e., no secondary task was ongoing).
  
  
• The independent variables analyzed were: two levels of Road surface (firm vs. slippery); three levels of Secondary task (radio, easy conversation, difficult conversation); and four levels of Activity (driving only, driving only but with obstacles to avoid, driving with communications over the mobile telephone (2-way) or the radio (conceived of as one-way communication in the study), and driving with manipulation of the radio or telephone). The analysis also examined the effects of between subject-variables of gender, age group, number of kilometers per year a test participant drove, and number of previous incidents that a test participant had been involved in while driving.
  
  
• The three sets of dependent variables were identified. First, road position was measured as a) Root Mean Square (RMS) position deviation off of the curved line), b) percent of time the car triangle was on the "shoulder of the road", i.e., between 25 and 50 pixels from the road, during which a green point was presented on the screen; c) percent of time the car triangle was "off the road", i.e., greater than 50 pixels from the road, during which red pointer was presented on the screen; and d) number of corrective steering movements that the driver performed. Second, the number of collisions with obstacles that appeared twice in every 5 minute sub-block was measured. Third, speed of driving was measured as a) mean driving speed relative to the given speed limit; b) RMS (absolute) deviations from the speed limit; and c) fast and slow violations, the former defined as the number of times that the test participant's speed went more than 10 km/hr above the speed limit, the latter defined as the number of times that the test participant's speed went more than 10 km/hr below the speed limit.

Major Findings:

 
• Preliminary assessment of the test participant variables resulted in only Gender being retained as a between-variable in subsequent analysis.
  
  
• Road position measures showed the following results. For position deviation, slippery road surfaces were the single most difficult condition. Among secondary tasks, the radio secondary task degraded performance most, followed by difficult conversation, then easy conversation. Among Activities, manipulation and driving only but with obstacle avoidance led to the greatest position deviations and small position errors. Males on average performed the tracking task better than females. Significant interactions among the three factors were reported as well for the RMS position deviation measure. Steering wheel movements revealed no effects.
  
  
• In terms of collisions, females had substantially more collisions than males on simulated slippery surfaces but males had more collisions on simulated firm or dry surfaces.
  
  Speed measures revealed the following. The only significant effect on mean speed revealed, somewhat paradoxically, that the test participants "drove" slower on the firm than on the slippery surfaces. In terms of speed deviation, there was significantly less speed deviation, on average, on firm surfaces than on slippery surfaces. Furthermore, speed deviation differed in statistically significant ways as a function of levels of secondary tasks, with greater speed deviations during difficult conversation than during simple conversation, with the radio task falling in between. A complex pattern of two-way and three-way interactions is reported for speed violations

Authors' Conclusions:

The authors state that the results show that communication in the form of conversing over a hands-free mobile phone or listening to a car radio during a simulating driving task affects driving performance in various ways. Given the current state of the art in mobile communications equipment, the authors advise that the driver park the vehicle before attempting to initiate telephone calls under difficult driving conditions, such as in-town traffic, on an unstable road surface, or with reduced visibility. They conclude cautiously by pointing out that the extent to which their findings can be generalized should be determined more closely in field studies of driving while using hands-free and hands-on mobile telephones.

Critical Assessment:

 
• The authors present a useful characterization of the car driving task in terms of modern theory from cognitive science.
  
  
• No data are provided to support the assertion that the pursuit tracking task simulated driving. The validation of driving simulators has proven to be a difficult undertaking and the low fidelity of the tracking task to real-world driving suggests that this may not be a faithful simulation. The demand characteristics of the laboratory situation may have been substantially different from that of real world driving.
  
  
• Because the pursuit tracking task is of unknown validity with respect to real world driving, the practical significance of the statistically significant results cannot be determined.
  
  
• The authors state that the primary task for the test participants was to drive safely. However, there were no consequences to degraded performance on the tracking task.
  
  
• No description is provided for the mobile telephone used in the experiment. This makes it impossible to appreciate the nature of the manipulation demands associated with telephone. Thus, though mention is made of a manipulation task associated with the hands-free phone (something of an oxymoron), no details are provided.
  
  
• · It is also unclear how the radio task data was parsed into manipulation components and "one-way communications" components. For example, there was no driver test administered to examine the extent to which the driver appreciated the contents of a radio broadcast. If so, there may have been no communications going on at all.

Brookhuis, K. A., de Vries, G., & de Waard, D. (1991). The effects of mobile telephoning on driving performance. Accident Analysis and Prevention, 23(4), 309-316.

Type of Study: On-road research with an instrumented vehicle.

Keywords: hands-free cellular telephones, hand-held cellular telephones, on-road measures, instrumented vehicle, human information processing

Author's Abstract:

The effects of telephoning while driving were studied in three different traffic conditions, i.e., in light traffic on a quiet roadway, in heavy traffic on a four-lane ring-road, and in city traffic. Twelve subjects, unfamiliar with cellular telephones, drove an instrumented vehicle for one hour each day during three weeks and while in each of the three traffic conditions, had to operate the cellular telephone for a short while.

To ensure a fixed "heavy traffic load" in the second condition, the subjects were instructed to follow another instrumented vehicle (at a safe distance). The results showed a significant effect of telephoning while driving as opposed to normal driving (i.e., not involving telephone conversation), on the effort subjectively measured by an effort scale and objectively measured by heart rate indices and on some of the measured parameters of driving performance.

One half of the subjects had to operate the cellular telephone manually, the other half performed the cellular telephone task with a hands-free cellular telephone set. The subjects who operated the hands-free cellular telephone showed better control over the test vehicle than the subjects who operated the hand-held cellular telephone, as measured by the steering wheel movements. Also, a clear improvement over time in the course of the 15 test days was found for some of the measurements. As a consequence of the results, some advice concerning mobile telephoning can be given to authorities, manufacturers, and users.

Sample and Methods:

 
• Drivers who had been licensed for at least five years, who drove at least 5,000 km annually, and who had not previously used a cellular telephone.

  12 Dutch participants (10 male, 2 female).

  4 participants were between 23-35 years old, 4 were between 35-50 years old, and 4 were between 50-65 years old.
  
  

• The apparatus used was a Volvo 245 GLD with redundant controls, modified to measure lane tracking, steering wheel movements, speed, following distance, and drivers' rear view mirror checking. An event recorder also measured drivers' cardiac inter-beat-intervals.
  
• A driving instructor was present during the testing sessions.
  
  
• Both hand-held and hands-free cellular telephones were used, however, the models for both types of telephones were not specified by the authors.
  
  
• For three weeks the driving performance of each subject was tested on every working day [although not specified by the author, it is assumed to be Monday through Friday], with and without operating a cellular telephone under three traffic conditions, i.e., light traffic on a quiet road, heavy traffic on a 4-lane road, and city traffic. To ensure heavy traffic, subjects were to follow an instrumented lead vehicle.
  
  
• Subjects both placed and received calls, although it was not specified whether or not they did both under all experimental conditions. The cellular telephone task consisted of a 3 minute paced serial addition task that was a fairly hard combination of a memory test and a mental arithmetic test.
  
  
• Half of the subjects had to operate the cellular telephone manually by picking up the handset. The other half of the subjects performed the task with a hands-free set.

Major Findings:

 
• Talking on a cellular telephone while driving significantly decreased standard deviation of lateral position, i.e., swerving, particularly on quiet roads.
  
  
• Talking on the cellular telephone significantly delayed reaction time in adapting to speed of variations of a car in front by 600 msec, only when the car in front's speed changes were not easily discriminated, e.g., when the front car slowed without lighting its brake lights.
  
  
• In city driving, steering wheel movements were affected by cellular telephone use. An interaction was found between cellular telephone type (hands-free/hand-held) and movements before versus after telephone contact:

  For drivers who dialed, the amount and amplitude of steering wheel movements were "violent" during the 20 sec. period before contact with the second party was reached, while for drivers who simply received a call, the standard deviation of steering wheel movements was elevated during the 20 sec. period after the cellular telephone rang.
  
  

• Whether or not they were using the cellular telephone, drivers checked the rear view mirror less often on the busy ring-road (heavy traffic condition) than on the quiet road (light traffic condition). However, an interaction was found between type of road and subsidiary cellular telephone task, so that telephoning did not lower rear view attention in heavy traffic any further (it "bottomed-out"), while an ample margin of rear view mirror attention was available in light traffic.
  
  
• Heart rate increased when the subsidiary cellular telephone task was carried out, compared to driving alone. This finding was in accord with drivers' self-reports of workload. Heart rate variability was also affected by increasing workload; however, practice over the three weeks of the study had a habituating effect. A significant linear trend towards slower heart rate and decreased heart rate variability was found.
  
  
• The performance on the cellular telephone task itself, i.e., the percentage of correct answers to the paced serial addition task (PASAT), significantly improved with practice over the course of the three weeks, and a strong learning effect was apparent in the first week.

Author's Conclusions:

 
• On quiet roads, cellular telephone use may decrease automatic information processing, and this may have an alerting effect on the driver. The author conceptualizes this effect to take place at the operational level of driving. This lowest, operational level refers to the driver's automatic action patterns and basic control over the vehicle.
  
  
• Manual dialing of the cellular telephone has adverse traffic safety consequences, leading to steering wheel amplitudes that were ten times the amplitude of those on the quiet road. This effect is comparable to that of tuning a radio while driving found in Stein, et al., (1987).
  
  
• The tactical level of the driving task is most affected by telephoning, as opposed to the strategical and operational levels. This tactical level involves reactions to the maneuvers of other cars, including checking the rear view mirror, adapting to the speed of the car in front, and braking in reaction to speed changes of other drivers. In fact, drivers seemed to employ some minimum strategy that was not affected by the subsidiary task. Evidence for this minimum strategy was found in the fact that the increase in drivers' reaction time to braking lights while telephoning was not statistically significant.
  
  
• The highest, strategical, level of driving involves route planning and speed level. As measured by speed maintenance, this level was not affected by telephoning in the present study, since drivers did not change their speed when they had the chance on the quiet road.
  
  
• Traffic safety may be decreased by cellular telephone use under certain circumstances, especially dialing by hand while driving in city traffic.
  
  
• Hands-free cellular telephones, preferably equipped with voice-activated dialing systems, are recommended.
  
  
• During a conversation it is recommended that drivers keep ample distance from other drivers and drive at moderate speed in the slower lanes.

Critical Assessment:

• This study benefits from multiple measures, i.e., physiological, behavioral, and self-report. However, correlations between these measures were not performed.
  
  
• A small sample size may have led to a failure to reveal age differences. In fact, age may have a differential effect upon strategic, tactical, and operational levels of information processing during the driving task.
  
  
• The finding of less rear-view mirror checking in the heavy as opposed to light traffic conditions is not surprising. During the heavy traffic condition, drivers had to follow a lead instrumented vehicle. Priority would necessarily be given to devote attention to this vehicle as opposed to traffic following behind, especially since the driver had to react to the lead vehicle from time to time.

Brown, I. D., Tickner, A. H., & Simmonds, D. C. V. (1969). Interference between concurrent tasks of driving and telephoning. Journal of Applied Psychology, 53(5), 419-424.

Type of Study: Closed-course test track with an instrumented vehicle.

Keywords: hands-free cellular telephones, gap judgment, spatial judgment, speed-accuracy tradeoff

Author's Abstract:

Twenty-four men were given the task of judging whether to drive through gaps which might be larger or smaller than the car. They were also given a telephoning task of checking the accuracy of short sentences. Interference between the concurrently performed tasks was investigated. Telephoning mainly impaired judgments of 'impossible' gaps (p < .01). The control skills employed in steering through 'possible' gaps were not reliably degraded, although speed of driving was reduced (p < .01). Driving increased errors (p < .01) and prolonged response times (p < .005) on the sentence-checking task. It is concluded that telephoning has a minimal effect on the more automatized driving skills, but that perception and decision-making may be critically impaired by switching between visual and auditory inputs.

Sample and Methods:

 
• 24 males within the age range 21-57 years old (median age 41).
  
  
• Car driving experience of these participants ranged from 3-37 years (median time 15.5 years).
  
  
• 22 participants were volunteers from the U.K. Ministry of Transport, and the remaining two were drawn from the Cambridge Applied Psychology Research Unit's Research Panel.
  
  
• Participants were alone in the test vehicle during the testing sessions.
  
  
• Participants communicated with the experimenter via a radiophone, consisting of a loudspeaker and telephonist headset. No manipulation of controls was necessary to perform the telephoning task.
  
  
• For the driving task, participants drove a 5 ft. wide test car around a track that required them to judge whether or not to drive through 20 gaps (4 of each size: 3 inches < car, 0 inches < car, 3 inches > car, 6 inches > car, 9 inches > car). In cases where drivers opted not to drive through a gap, the alternate route posed a comparable delay.
  
  
• In addition to judgments, successful clearing of accepted gaps, speed of performance, and lateral and longitudinal accelerations were measured.
  
  
• The telephone communications task was a paced grammatical reasoning task in which the driver heard a short sentence followed by the letters "A" and "B," where each sentence claimed to describe the order of the letter pair that followed. The driver decided whether the sentence was true or false and responded accordingly. Speed and accuracy of responses were recorded. Examples are provided below:
  

  Incoming Phone Message: "A follows B...BA"

  Driver (Correct) Response: "True"

  Incoming Phone Message: "B precedes A... AB"

  Driver (Correct) Response: "False"

   

• The telephone task was presented over a loudspeaker in the vehicle. The subjects responded via a telephonist's headset. A second experimenter sat in the back seat of the car and controlled the transmit/receive selector by a footswitch.
  
  
• All participants had 5 minutes practice on sentence-checking while stationary, then one practice trial of driving through the course and making gap judgments without the sentence task. Finally, participants had one practice trial of driving and telephoning concurrently. Six test trials followed in this order: 10 minutes of the sentence task, first driving-only trial, first concurrent trial, second concurrent trial, second driving only trial, and another 10 minutes of the sentence task only. Drivers met the various sizes of gaps in a different order on each trial.

Major Findings:

• In impossible gaps (width either 3 inches narrower than or width equal to the car) and in the largest possible gap (width 9 inches wider than the car), the percentage of errors of gap-judgment was significantly higher when drivers also had to telephone (19.2% increase for -3 inch gaps, 22.2% increase for 0 inch gaps, and 10.8% increase for 9 inch gaps). An impossible gap error judgment was defined as attempting to go through a gap that was narrower than or equal to the width of the car. A possible gap error judgment was defined as failing to go through a gap that was wider than the car.
  
  
• Skill in steering through possible gaps, as measured by clearing the gap, was not reliably impaired by telephoning, although there was a tendency for performance to be degraded when clearance was reduced to 3 inches wider than the car.
  
  
• A 6.6% increase in driving time per circuit was observed when participants telephoned concurrently. However, drivers' speed reduction was insufficient to prevent mutual interference between gap-judgment and sentence-checking.
  
  
• Average speed (.79 sec. longer) and average accuracy (21.2 more errors) of telephoning performance (sentence-checking) were both substantially impaired when participants also had to drive.
  
  
• There was a significant positive correlation between increase in driving time and increase in errors of gap-judgment.
  
  
• There was a significant negative correlation between increase in driving time and increase in errors on sentence-checking.

Author's Conclusions:

• Concurrent telephoning may have produced both a relaxation of gap-judgment criteria and an impairment of perception, resulting from switching between sensory modes.

  The relaxation of gap-judgment criteria would explain the increase in errors on impossible gaps, but would not explain the fact that drivers accepted fewer possible gaps while telephoning.

  When telephoning, drivers' perception was impaired as a result of switching between sensory modes (auditory and visual). However, a sensory impairment could not have been the sole source of interference, or the most difficult judgment (of 3 in. clearance) would probably have been degraded most rather than least.

  During judgments of possible gaps, the tendency for impaired perception to produce errors of rejection would have acted in opposition to the tendency for relaxed criteria to produce errors of acceptance.

  During judgments of impossible gaps, both impaired perception and relaxed criteria would have produced errors of acceptance. This would account for the finding that divided attention had the differential effect of causing a significantly large increase in acceptance of impossible gaps, but a smaller and mainly insignificant increase in rejections of possible gaps.

   

• The increase in driving time observed when drivers were telephoning could have resulted from at least two possible explanations:

  When telephoning, drivers made more gap-judgment errors, so they drove through the additional impossible gaps more slowly. Thiscautiousness would account for the finding that errors of gap-judgment were positively correlated with driving time.

  Alternatively, drivers deliberately reduced speed in order to handle the additional load of the sentence-checking task. Results suggest that drivers used the additional time to maintain performance on the telephoning task, at the expense of increased errors on gap-judgement.

   

• Drivers' age may determine the priority given to concurrent tasks of telephoning and driving, although the present study's sample size to detect this age difference was not sufficient.
  
• Calls while driving will take longer than ordinary calls, since messages such as the Baddeley (1968) that contain little redundancy are substantially affected by driving; therefore repetitions would be necessary to transmit all information, in practice.
  

Critical Assessment:

• This study assumes people utilize limited capacity, single channel information processing. However, one could argue that auditory and visual inputs may be processed in parallel.

  While an elegant explanation for results obtained, the author acknowledges that the claim that drivers suffer from a perceptual impairment during concurrent telephoning and driving deserves further study to determine whether this impairment is a function of the driving task, the type of conversation, or driver characteristics.
  
  

• The Baddeley (1968) task is a highly demanding cognitive task, and may be atypical of normal cellular telephone conversations. More recent research (e.g., Parkes, 1991) has found that type of conversation may differentially affect driver performance.
  
  
• After repeated hits, the gaps on the driving route may have "worn in" a little. This possibility of instrument decay may affect the ability to say with certainty that the concurrent tasks were solely responsible for producing the measurements obtained.
  
  
• Generalizability of results to real-life settings may be questionable since most gaps with a width of consequence are to a driver's side when passing, rather than straight ahead. Moreover, the closed course without traffic does not resemble normal roadway conditions.
  
  
• Results of this study should be generalized with caution to modern cellular telephones since the tasks associated with modern technology may be different than what was obtained/tested in 1969.
  
  
• With an unfamiliar test vehicle, drivers may not be as proficient at judging gaps than in a familiar vehicle
  .
  
• Task order should have been randomly counterbalanced between participants, even though results show that learning on the driving and telephoning tasks was negligible.

Department of the California Highway Patrol. (1987). A special report to the legislature on the findings of the mobile telephone safety study. California State Senate Concurrent Resolution

No. 8.

Type of Study: See Stein, Parseghian, & Allen (1987) for a full review.

Keywords: manual hand-held cellular telephones, memory-dial hands-free cellular telephones, voice-activated hand-held cellular telephones, cellular telephone mounting location, lane position, radio tuning, age differences

Abstract:

The impact of cellular telephone use on driver performance was tested using an interactive driving simulator. While drivers followed a 15-mile route containing curves, obstacles and signs, they sent and received mobile calls and tuned a radio. Type of cellular telephone varied (i.e., manual dialing/hand-held; memory dialing/hands-free; or voice-activated/hand-held), as did cellular telephone mounting location (dash or center console). Performance was compared between driver age groups, genders, and driving conditions, i.e., baseline driving, mobile phoning while driving, and radio tuning. Results suggest that manual dialing leads to driver performance that is worse than radio tuning; that performance deficits related to a secondary task increase with drivers' age; and that cellular telephones that are mounted within the driver's peripheral vision may reduce accident exposure compared to cab locations that are beyond peripheral vision.

Author's Abstract:

The study was designed to examine the effects of extra task simulation and extra rest on performance and fatigue of [heavy] haul truck drivers engaged in a simulated driving task. Sixty male subjects, randomly selected from the population of truck drivers in a large mining company, operated a driving simulator for a period of 7 h[ours]. A 2 x 3 experimental design was employed including two levels of rest conditions and three levels of secondary-task manipulations. The results show that performance and perceived fatigue were significantly higher when a secondary task involving voice communications was added to the basic driving task, but an added vigilance task had less effect. An extra 30-minute rest period in the middle of the experimental session significantly alleviated the reported experience of fatigue but did not affect performance. The results are discussed in terms of their relevance to actual industrial driving tasks.

Sample and Methods:

• Sixty (60) professional heavy-haul truck drivers randomly selected from a total population of 300 drivers employed by a large mining firm.

  Participants' ages ranged from 24-55 years old, with a mean of 39 years old.

  Participants had an average of three years of haul truck-driving experience with the company.

• A modified Redifon light motor vehicle simulator was used, showing a dimly lit road with no roadside features (corresponding to boring, rural driving conditions). A data logger switched lights, timed events, and monitored driver behavior.
  
  
• Participants drove for seven hours (21 blocks of 15 minutes duration each). Between each block was a six-minute rest.
  
  
• The vigilance task required that drivers turn off a set of four lights with corresponding switches. These lights illuminated in a random pattern, with an average frequency of one every 40 seconds.
  
  
• Participants communicated with the experimenter via a speaker and intercom device. No manipulation of controls was required to initiate the voice communication task.
  
  
• The voice communication task required that drivers respond to an incoming call over the intercom by reading the two least significant digits of the odometer (e.g., if the odometer showed 47,268.3 miles, participants were to respond "8.3". Drivers were contacted four times during each 15-minute block at random intervals.
  
  
• Participants practiced until they were capable of maintaining a steady position on the simulated road and were able to operate the vigilance or voice equipment as required for the particular experimental condition. It took 15 minutes to train the average participant.
  
  
• Five measures of driving performance were continuously recorded by the simulation apparatus: steering wheel reversals (SRT), tracking error (TET), number of brake responses to the appearance of tailgate lights (BRT), average brake reaction time (BTT), and control light response time (CRT). In addition, a subjective fatigue checklist was administered 15 minutes before the end of the experiment.

Major Findings:

• In general, the voice communication condition (driver is contacted four times during each 15-minute block at random intervals via speaker and asked to report current odometer reading) tended to yield the best performance whereas the basic driving condition (no secondary tasks) yielded the poorest performance. These results are presented in more detail, as follows:

  Brake reaction time was significantly lower in the voice communication condition than in the basic driving condition.

  Drivers made significantly fewer steering reversals in the voice condition than in the basic condition.

  Drivers made significantly fewer lane-tracking errors in the voice condition than in the vigilance condition (a switch-operated light-canceling task).

  Drivers made quicker and more accurate responses to a simulation of handling control lights (to operate the hauling mechanism?) in the voice condition than in the vigilance condition. Drivers' responses in the vigilance condition, in turn, were quicker and more accurate than the basic condition.

   

• No significant differences in performance were found with regard to the rest factor (30 minutes rest after the first three hours) for any of the five dependent variables.
  
  
• Self-reported fatigue was significantly lower on the extra-rest condition.
  
  
• The level of fatigue reported by drivers in the voice communication condition was significantly higher than those in either the vigilance task or the normal driving condition.

Author's Conclusions:

• A secondary task consisting of voice communication stimulus, a simple odometer screening and verbal response was more effective in maintaining good driving performance than a secondary vigilance task that required turning off lights by choosing the correct button.
  
  
• Secondary tasks should not be so demanding as to distract the operator's attention from driving and yet still have the potential of forcing him or her to maintain a higher level of alertness.
  
  
• The voice communication task forced drivers to maintain a relatively higher level of alertness and concentration which, on one hand, improved their performance, but at the same time increased their subjective reports of fatigue.
  
  
• In industrial driving situations, both performance and fatigue should be considered as independently important. Extra rest may help to alleviate increased sensations of fatigue introduced by a performance-improving secondary task; however, extra rest alone does not seem to aid performance.
  
  

Critical Assessment:

• This study highlights the possible benefit of increased performance when a secondary task, such as making verbal reports of odometer readings, is introduced into the monotonous task of driving. Results obtained by studying haul driving may not be generalizable to the less repetitive and more variable driving patterns found in non-industrial, non-simulator settings. However, the results do provide empirical evidence to support professional truck drivers' intuitions that conversation (currently accomplished mostly via the citizens band radio) can break the monotony of driving and help keep the driver awake.
  
  
• The voice communication condition did not involve a cellular telephone--rather it was handled through a speaker. While these results may help to illustrate the effect of communication on fatigue and alertness, this study's design did not consider performance effects of manipulating the cellular telephone equipment itself or the effects of engaging in different types of communication.
  
  
• Although this article's focus on heavy-haul drivers differs from the cellular telephone literature, Drory makes a useful contribution to the number of articles that find secondary tasks sometimes have what could be described as a general alerting or arousing effect on driving performance.

Author's Abstract:

Previous HUSAT research estimates that 65% of all cellular telephone conversation involves verbal negotiation. The present study required subjects to drive an experimental vehicle in a real road environment over three different experimental conditions. Two of the conditions involved a secondary task of engaging in a role-play negotiation whilst driving.

One condition involved negotiation with an experimental 'stooge' via a hands-free (i.e., no dialing or holding the handset required) cellular telephone and in the other, negotiation with an experimental 'stooge' sat in the front passenger seat. The third condition was an experimental control with no verbal negotiation task involved.

Twenty-four subjects took part in the experimental study. Driving behaviour was measured both in terms of objective data (time to complete experimental route, heart rate, eye movement behaviour) and subjective data (NASA-TLX and post-experimental questionnaire). The results of the study indicate that subjects found the secondary task conditions more difficult than the experimental control.

In the speaking and driving conditions, average speed decreased, heart rate increased and the questionnaire responses revealed an equivalent increase in perceived mental workload/stress compared with the control condition. The differences between the two speaking conditions were less pronounced yet significant at the physiological level. The implications of these results for cellular telephone users and cellular telephone design are discussed.

Sample and Methods:

• 24 drivers participated.

  All were familiar with the area of the experimental route (however road type, e.g., curvy or urban, is not described).

  Mean age of participants was approximately 45 years old with a mean of 25 years driving experience. (Participants' age ranges and gender were not described).

  Extraversion and introversion measures were taken with the Eysenck personality test.

• Three miniature video cameras were installed in an experimental vehicle to record drivers' face and eye movements, a view of the road ahead, and a view out of the right side of the vehicle.
  
  
• Each driver was exposed to three experimental conditions: normal driving (CONTROL), driving while conversing on a hands-free cellular telephone (CARPHONE), and driving while conversing with a person [the experimenter] in the front passenger seat (PASSENGER). Presentation order of experimental conditions was counterbalanced across subjects.
  
  
• The driving condition difficulty was not specified.
  
  
• A hands-free cellular telephone was used but the specific model was not reported by the authors.
  
  
• In both the CARPHONE and PASSENGER conditions, drivers negotiated a predetermined topic, e.g., booking a summer holiday or negotiating a partial exchange deal for the purchase of a car, until they reached a conclusion that was satisfying to them.
  
  
• Subjective data was gathered through the NASA-Task Load Index of perceived workload, and a rating scale questionnaire was developed to collect attitudinal responses to the experimental tasks.
  
  
• Driving performance data was gathered through videotape analysis of frequency and duration of eye movements. In addition, the time taken to complete each circuit of the route was recorded in seconds, indicating the average speed over the whole route.
  
  
• Psychophysiological data was gathered through a radio device strapped to the headrest that monitored electrodes on the driver's chest. Data was logged every 4 seconds, and was converted into the standard beats per minute measure of heart rate.
  
  

Major Findings:

• TLX questionnaire results revealed the highest overall workload in both CARPHONE and PASSENGER conditions compared to the CONTROL condition.
  
  
• Analysis of factors showed that mental demand, mental effort, and frustration were all higher in CARPHONE and PASSENGER conditions compared to the CONTROL condition.
  
  
• The subjective questionnaire revealed that 58% of subjects indicated that they felt more stressed in the CARPHONE and PASSENGER conditions and found these conditions more difficult.
  
  
• Participants also reported the most awareness of the extra load imposed by the secondary task in the PASSENGER condition.
  
  
• Using (what sounds like) a semantic differential questionnaire, 75% of participants indicated
  
  
• that speaking via the cellular telephone felt "unnatural."
  
• Drivers took 5% more time to complete the experimental route in the CARPHONE and PASSENGER conditions than in the CONTROL condition.
  
  
• Female participants took significantly longer than the male participants to complete the experimental route in all conditions.
  
  
• No significant differences in eye movements were found between the three conditions.
  
  
• Heart rate was highest in the CARPHONE condition, and this rate was significantly higher than CONTROL and PASSENGER conditions.
  
  

Author's Conclusions:

• The reduction in speeding while speaking, either on a cellular telephone or to a passenger, represents a means to cope with the attention tradeoff between driving and conversation tasks.
  
  
• Increased difficulty of the two speaking conditions was also reflected in the results of the subjective data.
  
  
• Since no significant differences were found in drivers' eye movements between the three conditions, the interference effects introduced by the speaking task appear to be modality-specific.
  
  
• The finding that drivers' heart rate was significantly higher in the cellular telephone condition may be explained at least two ways. First, the novelty of the cellular telephone may have induced either stress or heightened arousal. Or, second, the cellular telephone conversation may be fundamentally more demanding than speak ing to the passenger, and heart rate provides a physiological index of increased effort that went undetected by either the subjective or performance-based measures of driving behavior.
  
  
• The fact that heart rate was comparable between the passenger condition and the control condition may be explained by the fact that passengers can monitor the driver's task demands and converse appropriately.
  
  
• Additional driver stress while telephoning may be attributed to subjects' perceived lack of control over the continuation and initialization of the discourse. An interface should be developed to allow the driver to break from the conversation in case of any sudden rise in the demand of the driving task. This function would be augmented with machine-generated speech output informing the caller that the conversation has been interrupted because of the driving task, thereby taking the onus from the driver in the potentially embarrassing task of excusing themselves from the conversation.

Critical Assessment:

• In the description of participants, gender information and age ranges are not given. Without this information, it is hard to know for which kind of population the results may be pertinent. The finding that females took longer than males to complete the route might be more useful if we knew how many participants were of each gender and age.
  
  
• Heart rate may be increased by participants' reactance to wearing electrodes. As a measure of workload, the heart rate measure may be artificially high.
  
  
• In order to rule out the novelty explanation for heart rate results, future studies may compare workload between experienced and novice cellular telephone users. Such a study may help to show whether cellular telephone conversation is "fundamentally more demanding," as the authors suggest.
  
• It would be interesting to see if the experimenters-as-passengers could be trained to not monitor the driver's task demands and converse appropriately.
  
• This study would have benefited from an additional measure of standard deviation of lateral position.

Author's Abstract:

In this experiment 8 drivers (4 younger, 4 older) drove a 19 turn, 35-minute route. The route included sections through residential neighborhoods, on city streets, and on expressways. They were guided by an experimental navigation system that provided turn-by-turn instructions via a display mounted on the instrument panel. During the trip each driver was asked to dial a telephone number and participate in a simulated telephone conversation. At the end of the trip drivers were asked to rate the difficulty of a variety of driver-information-system-related tasks.

The instrumented car recorded lateral position in the lane, speed, throttle position, steering wheel angle, eye fixation location, and other measures. Typical lateral standard deviations were 0.5 feet and decreased with speed. Speed standard deviations were slightly in excess of 1 mile per hour. Using the cellular telephone and navigation systems resulted in slight increases in the standard deviation of throttle position and the standard deviation of steering wheel angle.

There were 8 navigation errors in this experiment, comparable to the 25 errors from 30 drivers in a previous experiment, a fairly low number. This experiment demonstrated that repeatable and reliable measures of driver performance and behavior could be obtained using the test protocol employed in this experiment.

Sample and Methods:

• Eight licensed drivers participated (four under 30 years old and four aged 60 or older). An equal number of men and women were in each age category. The corrected visual acuity of all participants ranged from 20/17 to 20/20 on a Titmus vision test.
  
  
• The vehicle used was a 1991 Honda Accord station wagon modified to measure lane tracking, steering wheel position, speed, accelerator/throttle percent declination, road scene, driver scene, and audio.
  
  
• A Motorola cellular telephone lying on the passenger seat was used for the cellular telephone tasks.
  
  
• Drivers engaged in three types of secondary tasks over the hand-held cellular telephone. In the listening task, drivers listened to a 30-second description of a scenario and then were prompted to make a decision based on the information they heard (e.g., drivers heard a description of three options for dining out, and then had to decide where to go). In the talking task, drivers were asked to describe something (e.g., what they did last weekend) for 30 seconds. For the listing tasks, a category was named (e.g., fruits) and drivers listed as many items in that category as possible in 30 seconds.
  
  
• Drivers made a total of 12 cellular telephone calls, including 3 practice calls while stationary, 3 practice calls while driving, and 6 calls during the test session while driving. Drivers made each of the three types of cellular telephone calls (listening, talking, and listing) first while driving on a 50 mph road, and then on a 65 mph expressway. The 12 calls were made in a fixed order by all drivers at the same locations along the test route. Each call included dialing a familiar number on a hand-held cellular telephone and completing one of the three conversation tasks.
  
• Data was also collected for the same driving performance measures while drivers used the navigation system. These results are not reported here.

Major Findings:

• Baseline driving, defined as driving along the seven straight segments of the road without using either the cellular telephone or the navigation system, revealed the following:

  Older drivers had larger standard deviations of steering wheel angle than younger drivers.

  Younger drivers had a mean lateral position further to the left than older drivers.

  For both younger and older drivers, mean lateral position was further to the left when the speed limit was over 55 mph than when it was 50 mph.

  Standard deviation of lateral position decreased as speed limit increased.

  For both younger and older drivers, mean speed increased as speed limit increased.

   

• The following effects of cellular telephone use on driving on straight roads were found:

  Young drivers had a smaller standard deviation of steering wheel angle than older drivers, similar to the baseline and navigation data.

  As measured by increases in steering wheel angle standard deviation, all three conversation tasks were equally difficult, and dialing was more difficult (distracting) than the conversation tasks.

  Throttle position varied the most during the talking task, suggesting that talking was most difficult. Throttle position varied the least during dialing, suggesting that dialing was least difficult.

  Measurements of mean lateral position showed that older drivers positioned the test vehicle closer to the center of the lane (3.0 feet to the right of the left edge verses 2.5 feet for younger drivers). This bias occurred for all road segments.

  Differences in mean lateral standard deviation were found between road types. However, no differences were found between tasks, drivers' age, or their interaction.

  Differences in mean speed were found as a function of road segment. In addition, on the expressway, younger drivers had a greater mean speed than older drivers, but not on the two-lane rural road.
  
  

• The following comparisons between baseline and cellular telephone task conditions were made:

  As measured by mean speed, participants drove more slowly while using the cellular telephone than in the baseline or navigation conditions.

  Participants drove more steadily, as measured by standard deviation of speed, in the baseline condition than in the cellular telephone condition.

  As measured by standard deviation of lateral position, drivers had less lateral variability when using the cellular telephone (dialing or conversing) than when driving alone.

  There was no differences between task conditions for absolute lateral position. An effect was seen between young and old participants, with older drivers positioning the vehicle closer to the center lane marking.

  When the vehicle was driven at a fairly steady speed, mean throttle position was greater while telephoning than in the baseline condition.

Author's Conclusions:

• Except for the standard deviation of steering wheel angle and throttle position, the cellular telephone task conducted concurrently with driving did not lead to differential effects in driving performance. This lack of significant differences could be due to the short sampling period or small sample size. This makes sense in that throttle and steering wheel measures are direct driver inputs while speed and lateral position are the results of those inputs as smoothed by vehicle inertia.
  
  
• The finding that age and road segment led to occasional differences, with the pattern that older drivers had larger values and more stable performance on higher speed roads, is consistent with previous studies.
  
  
• The main road-related factor in concurrent driving and cellular telephone use is speed.
  
  
• The standard deviation of lateral position does not seem to make sense, with lower standard deviations occurring while the cellular telephone was used. This could reflect a tradeoff with speed.
  
  

Critical Assessment:

• It is unclear how findings of differences in throttle position or standard deviation of steering wheel angle translate into practical safety decrements.
  
  
• Types of calls and roadways should be counterbalanced between drivers to control for practice effects.
  
  
• The sampling interval for the dialing task (10 seconds) was one-third of that for the other tasks (30 seconds), which may explain some of the differences.
  
  
• Although results for use of the cellular telephone itself and performance on the subsidiary task are given, they were not analyzed to show tradeoffs in driver attention between driving and cellular telephone tasks.
  
  
• The finding that lower standard deviations of lateral position were found while drivers used the cellular telephone could make sense, given the possibility of overall increased activation (e.g., Brookhuis et al., 1991).
  
  
• Drivers' lack of familiarity with the test vehicle may lead to poorer, or at least less consistent, driving performance.
  
  
• The listening and talking tasks used materials that were more realistic than the complex "intelligence test" materials used by other researchers.

Author's Abstract:

This report and the associated conference paper contain the results of a simulator study conducted to serve as a supplement to a NHTSA heavy vehicle driver workload field study. Its purpose was the evaluation of effects of cellular phone and text message display use tasks on driver-vehicle performance. Fourteen truck drivers participated and were asked to engage in three cellular phone dialing tasks (auto-dialing; local 7-digit dialing, and long distance dialing), two cognitive cellular phone tasks (responding to questions of a biographic nature or involving mental arithmetic), and seven CRT message reading tasks (tachometer checking, time checking, radio tuning, 4-line reading, auto-dial, local dial, and long distance dial).

Driver-vehicle performance was also evaluated relative to traffic density. Results indicated that driver-vehicle performance varied with respect to each of the three kinds of in-cab tasks.

Performance was also differentiated with respect to traffic density, although to a lesser extent. Of note is that the CRT reading tasks had a relatively more noticeable impact on driver-vehicle performance than either the dialing or cognitive tasks. This report concludes with a comparison of simulator and on the road data collection results and prospects for future heavy vehicle driver workload assessments.

Sample Studies:

• 14 professional heavy vehicle drivers. Participant ages ranged from 26 to 68 years old, with a mean age of 47.1 years. Twelve of the drivers had previous cellular phone experience.
  
  
• The STISIM driving simulator developed by Systems Technology Inc. was the research apparatus. The closed-loop, low fidelity simulator is fully interactive and includes a 5-speed transmission, variable vehicle dynamics, simulated road noise, tire squeal, and wire frame rendering of displayed objects.
  
  
• The truck cab mock-up replicated the interior of a Kenworth truck cab, including a truck seat, steering wheel and turn-indicator assembly, and a neutral gear shift.
  
  
• The windshield of the cab was divided into two sections. The left front housed a 20 inch color monitor to present the various driving scenarios. The right section of windshield and driver side window were covered to reduce ambient lighting.
  
  
• Verbal responses were used for most of the in-cab tasks. Manual responses were required for a tachometer task and an object detection task.
  
  
• A Motorola Attache Cellular Telephone, Model #TX400 was installed to the right of the driver.
  
  
• A text message display consisting of a 7-inch diagonal VGA-compatible green-phosphor CRT was mounted on top of the instrument panel to the right of the seated driver.
  
  
• Driving scenarios consisted of six separate modules, each approximately 100,000 feet in length and required about 30 minutes to complete at 55 mph.
  
  
• Modules were defined by high and low traffic volume conditions and manual dialing events or no manual dialing events.
  
  
• At preset distances within a scenario, messages appeared on the text message display and were announced by an auditory alarm.
  
  
• When prompted by the text message display, drivers dialed a number and reached a recorded message presenting a set of paced questions, seven seconds apart, allowing time for the driver to respond. Dialog tasks lasted about 60 seconds and included biographic (easy/low difficulty) and arithmetic (moderate difficulty) questions.
  
  
• Manual dialing was completed by pushing a "send" button.
  
  
• All curves built into the scenarios were easy (i.e., shallow)and included nine to the right and nine to the left.
  
  
• In cab tasks were divided into cellular-telephone dialing and dialogue, text message reading, tachometer reading, time reading, manual radio tuning and object detection (pedestrian). All but object detection tasks were initiated by a text message.
  
• Manual task data (auto-dial, 7-digit, and 10-digit manual cellular telephone dialing, plus radio tuning as a control condition) and cognitive task data (the two question-and-answer dialogue tasks) served as measures of driver performance.

Major Findings:

• For manual tasks, mean lane position was closest to lane center for all dialing tasks and farthest for the radio tuning task relative to driving only without in-cab tasks.

  Lane deviations during the manual tasks were smaller in high density traffic than in low density traffic.

  Lane exceedences were greater during dialogues of either type than when driving only.

   

• Drivers reduced their speed for the local-dialing task.
  
  
• Standard deviations of vehicle speed were higher for radio tuning and local-dialing tasks.
  
  
• Latency to detect a pedestrian was slower when no phone call was in progress.
  
  

Author's Conclusions:

• Results indicate some disruption of simulated driving due to complex tasks.
  
  
• The standard deviation of lane position was affected by inserting secondary tasks.
  
  
• Lane exceedence was also influenced by secondary tasks.
  
  
• Time to detect a pedestrian in the roadway was improved by adding a complex secondary task.
  
  
• There were no effects of traffic density considered over the entire secondary task.
  
  
• During data entry, drivers were more careful about lane position when traffic was heavy.
  
  
• Comparison of results with the companion on-road study (Tijerina, et al., 1995) indicated a number of differences.

  The on-the-road study found the greatest lane position standard deviation associated with message reading taks followed by the long-distance phone task. The simulator study found no differences between all reading and all dialing tasks.

  Variability was greater in the simulator study.

  No differences in lane exceedence were found in the on-the-road study for dual task conditions versus driving only. However, such differences were found on the simulator.
  

Critical Assessment:

A comparison of results with the companion on-the-road study (Tijerina, et al., 1995) yields a number of differences in findings. Despite the smilarity in tasks, materials and procedures, many differences existed between the two studies that could account for the discrepant findings and highlight the potential limitations of simulators for carrying out studies involving "risk taking" behaviors.

• Road scene characteristics were very different between the two studies.
  
  
• The cab layout for the two studies were different.
  
  
• The pattern of differences between the two studies suggested that heavy vehicle drivers in the simulator adopted a more lax attitude toward the driving task, perhaps as a result of the no-risk consequences of degraded lanekeeping in the simulator.
  
  
• Limitation in simulating driving condition variables (e.g., ambient lighting, road type) and the actions of other vehicles has the potential to limit the applicability of results where driver secondary task behavior (e.g., dialing a number, tuning the radio) might increase the risk of a crash.
  

Nilsson, L., & Alm, H. (1991). Elderly people and mobile telephone use--effects on driver behaviour? Proceedings of the Conference Strategic Highway Research Program and Traffic Safety on Two Continents. Gothenburg, Sweden, and DRIVE Project V1017 (BERTIE, Report No. 53), March 1991.

Type of Study: Driving simulator.

Keywords: hands-free cellular telephones, age differences, brake reaction time, lateral position, speed

Author's Abstract:

The effects of a cellular telephone conversation on driving were studied in the advanced driving simulator at VTI. Twenty subjects, 10 men and 10 women, between 60 and 71 years and 20 subjects, also 10 men and 10 women, between 23 and 58 years participated in the study. The road the subjects drove could be characterized as "easy." It was straight and not expected to cause the subjects any problems with speed choice and steering strategy.

The workload imposed on the subjects by the driving task was thus supposed to be very low. The cellular telephone task included handling of the cellular telephone and a conversation, containing a working memory part and a decision part. The handling task consisted of pushing the hands-free button to activate the cellular telephone when it was calling. During the conversation the subjects were asked to listen to pre-recorded sentences and for each sentence to judge if they experienced it as "sensible" or "nonsense." After a number of sentences they were required to recall the last word in each sentence, in the order they were presented.

Sample and Methods:

• Twenty licensed drivers, 10 males and 10 females, ages 60 to 71 years with a mean age of 65.9 years participated. All drove at least 10,000 km per year and had been driving at least 5 years.
  
  
• Half of the subjects were assigned to an experimental condition and the others were assigned to the control group. Subjects in the experimental group drove and performed the cellular telephone tasks consisting of receiving calls only, while subjects in the control group only drove.
  
  
• Whenever age was used as a variable in the analysis, data for young drivers was added from an earlier study (see Alm & Nilsson, 1990).
  
  
• The apparatus used was the VTI Driving Simulator.
  
  
• A hands-free cellular telephone mounted on the instrument panel was used for the cellular telephone tasks.
  
  
• All subjects received practice driving in the simulator on a 20 km long, straight road. The actual test route was 80 km long.
  
  
• A red square appeared on the left shoulder of the road to simulate an abruptly emerging event on the traffic scene. Participants were instructed to brake as soon as they saw the square.
  
  
• During the practice session, all subjects had an opportunity to experience the visual stimulus (red square) to which they had to brake as fast as possible. Subjects in the experimental condition received practice on the cellular telephone task as well.
  
  
• The cellular telephone was mounted on the instrument panel to the right of and at the height of the steering wheel. When a driver answered the cellular telephone by pressing the hands-free button, a tape recorder was activated and "read" the cellular telephone task to the driver. The presented cellular telephone tasks and the participants' answers were recorded on a second tape recorder.
  
  
• The Baddeley et al. (1985) Working Memory Span Test was chosen for the cellular telephone task. It contained a working memory part and a decision part.
  
  
• Eight cellular telephone calls were presented to drivers in the experimental group. Calls occurred at eight randomly chosen locations along the route. At four randomly chosen positions, the red square appeared in connection with the cellular telephone calls. For two of these four occasions (again randomly chosen) the visual stimulus appeared 1 second after the cellular telephone had rung, while for the remaining two occasions the visual stimulus appeared 30 seconds after the ring signal.
  
  
• The following driver performance measures were taken: speed (km/h), lateral position on the road (meters), variation in lateral position (meters), brake reaction time (seconds), number of correct sentence judgments (sensible/nonsense), and number of correctly recalled last words (in the order of presentation).
  
  
• Drivers' subjective workload was measured by the Task Load Index (NASA-TLX).

Major Findings:

• The mean brake reaction time for the four simulated danger situations (red square stimulus) was calculated for each subject.

  Regardless of age, use of the cellular telephone while driving resulted in longer brake reaction times compared to driving without using the cellular telephone.

  Regardless of cellular telephone use or non-use, older drivers had longer brake reaction times than younger drivers.
  
  

• Mean absolute lateral position was recorded for the 0-500 meter and 0-2500 meter distances after the initiation of each cellular telephone call. The 500 meter segment covered the distance during which drivers had to activate the cellular telephone, while the 2500 meter segment corresponded to the entire cellular telephone task.
  
  

  No effect of cellular telephone use on mean lateral position was found.
  
  

  When the 0-500 meter distance was considered, the "no effect" result obtained for the elderly drivers was in agreement with the result for the young drivers. However, when the 0-2500 meter distance was considered, the young participants drove significantly more to the right while performing the cellular telephone task, compared to young drivers that only drove.
  
  

• Variation in lateral position was also recorded and analyzed.

  Elderly drivers varied their lateral position more when they activated the hands-free function while driving compared to elderly drivers that only drove.

  When the distance covering completion of the entire cellular telephone task was considered, the variation in lateral position was larger for elderly drivers who used the cellular telephone while driving than for elderly drivers who only drove.

  During the phase of driving when the hands-free function had to be activated (0-500 meters), the effect of the cellular telephone was influenced by age, i.e., elderly participants varied their lateral position more than the younger participants.

  The effect of cellular telephone use was influenced by age during performance of the entire cellular telephone task (0-2500 meters). The variation in lateral position increased for the elderly drivers compared to the elderly driving only condition, while it decreased for the younger drivers compared to the younger driving only condition.
  
  

• NASA-TLX ratings were used to analyze drivers' subjective workload.

  Elderly drivers who performed the cellular telephone task while driving rated mental demand higher than elderly participants who only drove.

  Independent of their age, drivers who were engaged in the cellular telephone task rated mental demand and effort higher, were less pleased with their performance, and were more frustrated than drivers in the respective control groups.

  Younger drivers reported more frustration than elderly drivers when they used the cellular telephone while driving.
  
  

• Mean speed values were calculated from initiation of the hands-free function and 80 seconds forward, covering the entire cellular telephone task. Corresponding calculations were made for drivers in the control group.

  For every cellular telephone call, mean speed for elderly participants using the cellular telephone was lower than the mean speed for elderly participants who only drove.

  Independent of age, participants drove more slowly (9.2 km/h on average) while using the cellular telephone compared to those that only drove.

  Independent of cellular telephone use, the elderly participants drove on average 4.7 km/h faster than the younger participants.

  Performance on the Working Memory Span Test showed that young drivers made significantly more correct judgments and more correct recalls compared to the elderly drivers.

Author's Conclusions:

• Both cellular telephone use and age have a negative impact on drivers' ability to react quickly to a suddenly appearing event.
  
  
• Young, experienced drivers' brake reaction time on an easy road while engaged in a cellular telephone conversation coincides with that for elderly experienced drivers, driving on the same road but without using a cellular telephone.
  
  
• A cellular telephone conversation seems to have a more severe effect on elderly as opposed to younger drivers' tracking ability. While young drivers move to the right on the road and keep a more steady course, elderly drivers increase their variation in lateral position. Thus, elderly drivers using a cellular telephone while driving, run a greater risk to intrude into the wrong lane or to leave the road.
  
  
• Drivers do not try to keep up a constant level of workload. Instead of reducing the demands added from the secondary cellular telephone task, they seem to work harder to cope with the new situation.
  
  
• Although elderly drivers decreased their speed while using the telephone, they still drove faster than the young drivers. This finding may be due to the simulator environment's tendency to create "speed blindness."
  
  
• As elderly drivers usually are aware of their reduced capacities and receptive to procedures and devices that would make them safer and better drivers, it may be worthwhile to inform them about obtained behavioral effects and suggest strategical solutions.

Critical Assessment:

• The "suddenly appearing event" is a needed element in the study of the impact of cellular telephone use on driver behavior because it comes closest to simulating an emergency situation, an instance when drivers' responses may be most crucial. Future studies should strive to increase the realism of the event, i.e., to use a stimulus other than a red square, and should necessitate simultaneous tracking and brake reaction maneuvers.
  
  
• The authors did not report reaction time differences for one second and 30 seconds after initiation of the cellular telephone call. Workload may be higher at one second, the time of the ring, as opposed to 30 seconds into the task, which may be reflected in differential reaction times depending on the timing of the cellular telephone task.
  
  
• The driving task used was an "easy" road with no interaction from other traffic. This may not be a realistic representation of actual traffic conditions.
  
  
• There was no analysis of easy versus hard workload. Cellular telephone use may have a stimulating or alerting effect for boring roadways.
  
  
• No analysis of intense versus light conversation was reported, which in other studies has shown a significant effect on driver performance.
  
  
• No analysis was made of placing a cellular telephone call, especially dialing, a task which may be more critical or detrimental than receiving a cellular telephone call or holding a conversation.
  
• The findings related to speed are opposite to findings in other similar studies. Usually, older drivers drive slower than younger drivers. Although the finding in the present study may indeed be due to "speed blindness" as proposed by the authors, this unique finding deserves closer analysis.
  
  

Pachiaudi, G., & Chapon, A. (1994). Car phone and road safety. XIVth International Technical Conference on the Enhanced Safety of Vehicles, No. 94-S2-0-09. Munich, Germany.

Type of Study: Driving simulator.

Keywords: hands-free cellular telephones, speed variability, self-report

Author's Abstract:

The aim of the present study was to experimentally assess the potential risk on road safety when using a hands-free cellular telephone while driving and maintaining interactive conversation. This exploratory research was carried out on an interactive driving simulator and was based on about 40 routes performed by 17 subjects. In a first step, the observation of speed variations induced by the use of a cellular telephone brought two types of reaction into prominence: 1) no effect, and 2) a more rigid driving behaviour. The latter is shown by a speed increase or decrease, or by a longer period of oscillations around the required speed, or even by a total loss of speed control.

These reactions were compared to answers to a questionnaire asking for the causes of difficulties encountered and how the driving task was disturbed by phoning while driving, and vice versa. This comparison allowed us to see how subjects managed the dual task; in most cases, they used a time sharing strategy during which the main task, i.e., driving, was often perturbed by the second one, i.e., phoning.

Sample and Methods:

• Seventeen drivers participated.

  9 were 18-35 years old.

  8 were 45 years or older.
  
  

• Participants received a call and held a cellular telephone conversation while driving on a simple route in a driving simulator. Subjects were required maintain vehicle speed at either 90 or 130 kph. The nature of the conversation was not described.
  
  
• A hands-free cellular telephone was used although the model type was not specified.
  
  
• Variation in vehicle speed was used as the dependent measure.
  
  
• Subjective measures were also collected through post-study questionnaires. These questionnaires focused on the possible causes of difficulties encountered during the conversation, the tendency to alter driving behavior while telephoning, and the difficulty in telephoning while driving.

Major Findings:

• In 18 out of 39 instances of concurrent telephoning and driving, no speed modification could be detected when comparing the speed variation before the call and during conversation. Among these instances:
  

  Some drivers, after initial stabilization [the time period allowed for stabilization was not specified by the authors], kept a constant speed.

  Others had some difficulty stabilizing the speed as seen by symmetric and steady oscillations of about 15 to 20 kph, but these findings were not statistically significant with regard to cellular telephone use or driver age.

  A few others had significant difficulty stabilizing the speed as seen by oscillations of more than 50 kph, though again, no differences were found with regard to cellular telephone use or driver age.
  
  

• In 21 out of 39 cases, it seemed that there was a conflict between the two tasks of driving and telephoning, and driving often lost its status of prime task. This principle of "behaviour stiffening", i.e., doing one task or the other alternately, in relation to speed was observed in several ways:

  A steady increase of speed (about 30 kph) without any attempts to correct it.

  An increase of the period of speed oscillation (e.g., from 60 seconds to 100 seconds or from 90 seconds to 160 seconds)

  A loss of speed control, leading to oscillations whose amplitude could be over 80 kph.
  
  

• These findings, however, did not reach statistical significance.

  Multiple choice questionnaires gauged participants' impressions of the dual-task experience.

  Younger participants reported feeling more difficulties than the older participants.

  The main causes of difficulty that participants reported were "noise" and "dual task."

  50% of participants reported that they would slow down from a speed of 130 kph while telephoning. 33% of participants reported that they would slow down from a speed of 90 kph while telephoning.

  Older participants reported NOT being disturbed by telephoning while driving much more than the younger participants. However, no significant age differences in performance were found.

Author's Conclusions:

• Of the seventeen participants in the sample, only two showed no change in their speed and did not report feeling any trouble while telephoning.
  
  
• Nine participants decreased their speed while telephoning and the remaining six participants experienced mental overload which would have resulted in a degradation of performance.
  
  
• In view of these results, more systematic investigations of the various elements found in the driving situation, such as lane keeping, speed, reaction time, eye direction should be performed. Age and practical experience of cellular telephone use should be taken into account in this further stage.

Critical Assessment:

• This study would have been strengthened by using validated cognitive tasks within cellular telephone conversations to establish a quantifiable mental workload level, and to reduce learning effects.
  
  
• Additional driving performance measures besides speed variation could have shown task tradeoffs and changes in driving performance better than speed variation alone.
  
  
• While the authors' interpretation of "behaviour stiffening" is interesting, they also note that no significant differences were found between the cellular telephone and control conditions, or between age groups. Additionally, percentages of responses are given for the subjective measures without determining significance. In the absence of significant results, the authors' firm conclusions that drivers experienced "mental overload" or a "conflict provoked by the increase of mental load" seem premature.
  
  
• The 40 routes driven and used to analyze the data were done by 17 subjects. The authors do not describe which of the 17 subjects drove what number of routes. 40 is not a multiple of 17; therefore, each subject did not drive the same number of routes.
  
  
• The nature of the cellular telephone conversation was not described.

Parkes, A. M. (1991). Drivers business decision making ability whilst using carphones. In Lovessey, E. (Ed.), Contemporary Ergonomics, Proceedings of the Ergonomic Society Annual Conference (pp. 427-432). London: Taylor & Francis.

Type of Study: On-road research, Laboratory research

Keywords: cellular telephones, dual tasks, cognitive interference

Author's Abstract:

This paper reports the first stage of analysis of drivers abilities to make decisions while using cellular telephones in a moving vehicle. Results show that subjects have difficulty remembering and interpreting complex information. It is concluded that there is a need for improved customer support and greater functionality of future cellular telephone systems, if full system potential is to be achieved.

Sample and Methods:

• 24 drivers (15 male, 9 female), ranging in age from 18-50.
  
  
• Cellular telephone conversations consisted of a hybrid test whose development was influenced by the Watson-Glazer Critical Thinking, Wechsler Adult Intelligence Scale, Wechsler Memory Scale and Hodgkins Differential Aptitude tests. The resultant test consisted of seven sections: numerical and verbal memory, simple arithmetic, numerical reasoning, inference, deduction and interpretation. Participants were introduced to a sample of the test material prior to actual testing.
  
  
• All participants completed four experimental conditions over two testing sessions.

  Two conversation conditions were conducted while driving: driver to passenger (DP); and cellular telephone to office (CO). Two other conversation conditions were conducted in a laboratory: stationary phone to office (TT); and face-to-face (FF).

  The experimenter took the role of the person on the other end of the telephone. During the DP condition, this experimenter was in the front seat of the car. In the CO condition a second experimenter sat in the back seat to give directions while the conversation was held over the cellular telephone with the first, remote experimenter.
  
  

• · Before starting either driving condition, the participant was shown a map of the route to be followed, and completed a short familiarization drive in the car. The route involved a mixture of suburban and rural roads.
  
  
• Participants received calls only and the design of the cellular telephone was not reported.

Major Findings:

• Participants' scores on the critical thinking test were highest for face-to-face conversations, followed by stationary telephone to office conversations, followed by driver-to-passenger conversations, and finally, cellular telephone to office conversations.
  
  
• Differences were found between the driver to passenger (DP) condition and cellular telephone to office (CO) condition for the following sub-scores:

  Verbal memory decreased by 25%.

  Numerical memory decreased by 21%.

  Interpretation decreased by 23%.
  
  

• Differences between the stationary telephone to office (TT) condition and the cellular telephone to office (CO) condition were found for the following sub-scores:

  Verbal memory decreased by 21%.

  Numerical memory decreased by 20%.

  Interpretation decreased by 19%.

Author's Conclusions:

• It seems that the combination of the driving task with the cellular telephone results in difficulty in remembering verbal or numerical data, and in making correct interpretations from background information.
  
  
• The lower scores cannot be solely attributed to the dual demands of the driving task, as scores in the driver to passenger condition were not significantly different to those produced in the laboratory in single task conditions. Videotape analysis suggests that the experimenter-as-passenger naturally made allowances for traffic movements and maneuvers when administering the test.
  
  
• Some reservation should be placed on notions of the development of the future "office on the move."
  
  
• Users of cellular telephones need to be realistic in terms of what they can and should be used for.
  
  
• Suppliers could help the situation by providing explicit advice to customers about the potential difficulties of holding, or being drawn into, complex business negotiations while on the move.
  
  
• Suppliers should further investigate the possibility of increasing the functional capability of cellular telephones. The "hands-free" facility is a major advance, but other features such as the proposed "intelligent answerphone" should be pursued. Systems that can divert, record, and interrupt messages appropriately may prove cost effective for many business people of tomorrow.

Critical Assessment:

• This study assumes people are single channel information processors. Further studies should examine the mutual interference between driving and telephoning within a broader, more parallel framework.
  
  
• The present study would have benefited from simultaneous measures of driving performance. This addition may have better illustrated drivers' attentional tradeoffs between the driving and cellular telephone tasks.
  
  
• The experimenter-as-passenger arrangement used in the DP condition allows for the possibility of experimenter bias and/or Hawthorne effects which may skew the results.

Parkes, A. M. (1993). Voice communications in vehicles. In Franzer, S. and Parkes, A. (Eds.), Driving future vehicles (pp. 219-228). London: Taylor and Francis.

Type of Study: Review of experimental findings.

Keywords: cellular telephones, driver performance, technology development

Abstract:

A literature review was performed to explore the questions: a) Does talking on a cellular telephone affect driving performance? and b) Does driving make it difficult to use a cellular telephone in the same way as a standard fixed telephone? A limited capacity model of information processing is used to highlight tradeoffs in both tasks, as shown by the experimental findings reviewed. Problems associated with cellular telephone conversations are emphasized, both in terms of the effect on the driving task and on the conversation itself. Implications for driver adaptation as well as human factors implications for the development of future cellular telephone technology are discussed.

Sample and Methods:

• Types of studies reviewed include the following:

  Driving simulator studies.

  On-road driving studies.

  Experimental social psychology studies of cellular telephone behavior.

  Perceptual and cognitive studies relevant to the dual task of mobile telephoning while driving.

Author's Conclusions:

• Though it has been less than comprehensive so far, research into the effects of in-vehicle cellular telephone conversations has important implications not only for that application but also for the nature of future interface design.
  
  
• Priority can be given to the primary task of driving while talking on a cellular telephone, without observable decrements in performance, so long as some threshold point is not reached. Cumulatively, the literature has shown performance decrements, mainly swerving, increased reaction time in braking, decreased following distance, and inaccurate spatial perception, that become observable in association with variations in surrounding traffic, road type, driver's age, cellular telephone design, and nature of telephone task.
  
  
• Driving and communication system usage can be considered within the context of a limited capacity model of human information processing. In this model, a task may take up a proportion of available capacity, leaving a certain "spare capacity." The introduction of a second task will make demands on this limited channel and take up spare capacity of the system.
  
  
• The operational skill level component of the driving task is reasonably robust and only likely to show deterioration at times of high primary (i.e., driving) task difficulty. It might be hoped that the driver might take appropriate action in such situations and either not accept incoming calls or close down current ones.
  
  
• While drivers seem aware that holding cellular telephone conversations while driving involves them in increased workload and a certain amount of stress, drivers also seem largely unaware of the increases in their response times. Generally, drivers' lack of awareness or adjustment of their driving impairment while using a cellular telephone is analogous to the public perception of the level of impairment produced at low levels of blood alcohol.
  
  
• Many cellular telephones have sophisticated voice recognition facilities, and a large number of commands for call set-up, number dialing and call termination can be activated by voice alone. Such voice input, allowing truly hands-free and more importantly, eyes-free operation, is to be encouraged if it can be demonstrated that a superior manual alternative does not exist.
  
• Future cellular telephones may benefit from features that would be able to intervene with appropriate messages to the second party, reducing pressure on the driver to keep up the flow of conversation when encountering immediate driving task demands. This technology could help control for the fact that the remote person does not "punctuate" their speaking the same as an in-car passenger would. Cellular telephones could be linked to route guidance or collision avoidance systems to provide such an intervention. System integration is a challenge for future designers.
  
• The temptation to design a "mobile office" in a car should be tempered by the reality of a person's limited ability to simultaneously respond to conversation demands and driving demands.

Critical Assessment:

• This article puts safety research on cellular telephone use into a practical context by exploring mobile communication in both directions: the effects of driving on telephone behaviors and the effects of telephoning on driving. The limited capacity model of human information processing suggests that aspects of both tasks, driving and communicating, may be compromised by performing them simultaneously.
  
  
• It should be remembered that a limited capacity model of attention is just one framework in which phenomena such as driving and telephoning can be discussed. For example, improved performance at the operational/control level of driving (Brookhuis, et al., 1991) suggests that neurocognitive arousal and levels of processing views of attention are needed in the study of mobile communication.
  
  
• Parkes acknowledges that during times of high driving difficulty, drivers should not use their cellular telephones. He envisions a time when advances in technology allow the cellular telephone itself to shut off at such times. Until such advances are realized, driving conditions sometimes may change so suddenly as to not permit drivers an effective return of all attention to the driving task. Except for one study in which car-sized boxes moved into the drivers' path, requiring an emergency lane change (Stein et al., 1987; or Department of the California Highway Patrol, 1987), the literature appears to be missing comparisons of drivers' emergency response performance between cellular telephone use and non-use. Future research may explore this issue in more depth.

Petica, Stefan (1993). Risks of cellular phone usage in the car and its impact on road safety. Recherche-Transports-Securite, 37, 45-56.

Type of Study: Review of literature and actions taken by various countries.

Keywords: cellular phones, traffic accident risk, driver attention

Abstract:

The effect of telephone communications on driving and hence on road safety have not been very clear until recently and experimental studies of the subject (which have been fairly contradictory) have not received much of a response. But the rapid increase in this type of equipment which is expected in the years to come has refocused attention onto the subject and the risks of accidents resulting from the implications of this task on the driver are starting to be considered. In order to better locate the true risk the author has analyzed the most significant variable in the situation. The major directions which this research is taking are as follows:

• examination of the socio-technical factors arising from the considerable increase in the number of carphones (four million in France by the year 2000),
  
• examination of several typical experimental studies of the psychological and behavioral effects this task has on driving,
  
• presentation of the activities of the public authorities abroad with respect to this potential hazard and the position as regards regulations in several countries,
  
• summary of a few facts which have emerged from the analysis of the state of the art, conclusions on risk assessment and proposals for educational, ergonomic and regulatory measures.

Sample and Methods:

• Types of studies reviewed include the following:

  Driving simulator studies.

  On-road driving studies.

  Experimental social psychology studies of cellular telephone behavior.

  Perceptual and cognitive studies relevant to the dual task of mobile telephoning.
  

• Actions taken by public entities reviewed include the following:

  International survey of experts and policy-makers conducted by INRETS.

  United States experience including the California Highway Patrol study.

  European experience including Switzerland, Germany, Spain, Sweden, Great Britain, and France.

Author's Conclusions:

• Due to the interference that might arise when the tasks of driving and telephone communication are performed simultaneously, it should be taken into account that the probability of an accident in such cases rises significantly. The extent of risk, if principally linked to circumstances on the road, may be negligible or rise in probability two to four times.
  
  
• The increase in probability of an accident has been established for all types of users and for control, guidance and navigation activities; but its distribution varies, according to the age of the user, traffic conditions, placement and type of phone. This increase has been demonstrated from the experimental point of view by most of the studies analyzed, even though from the point of view of accident statistics there does not exist much proof. This is largely due to the difficulty involved in the actual observation of these types of accidents.
  
  
• Dialing of telephone numbers is shown to be the most dangerous, considering its characteristics: visual distraction and splitting of attention while driving with one hand. This is true even for the hands-off type of phone.
  
  
• The probability of accidents is much higher for all types of phones when problems, emergencies or conflicts arise.
  
  
• It is very difficult to discover any detrimental and secondary effects of phone usage that might be due to driving style and to the influence that this can have on behavior of other users. More research is called for.
  
  
• From the safety point of view, the presence of a car phone has certain advantages, but these advantages are due to the presence of the phone in the car, not to its utilization during driving.
  
  
• The risk of accidents is considerably decreased if the task of communication is simplified by:

  - the presence of a memory function which shortcuts the dialing procedure

  - the presence of an automatic answering system

  - the presence of a hands-off or similar system and,

  - in the future, voice recognition.
  
  

• For the mid-term, the growth in the number of car phones, as well as the actual practice of phone usage under traffic conditions might increase the number of accidents.
  
  
• A preventative and regulatory attitude on the part of the public authorities imposes itself, especially since new communication products will soon be introduced in cars. the combination and interference of the above potentially negative factors, perceived as negligible, will certainly pose new problems for road safety if certain measures are not taken. Based on the fact that other countries have already regulated car phone usage in one way or another, a similar attitude would be desirable in Europe.

Critical Assessment:

This is an early review of cellular phone issues and foreign public and governmental approaches to dealing with the potential risk of cellular phone use.

Redelmeier, D. A., & Tibshirani, R. J. (1997). Association Between Cellular Telephone Calls and Motor Vehicle Collisions. The New England Journal of Medicine, Vol. 336, Number 7, 453-458

Type of Study: Epidemiologic, case-crossover design..

Keywords: cellular telephones, motor vehicle collisions, crash risk

Author's Abstract:

Because of a belief that the use of cellular telephones while driving may cause collisions, several countries have restricted their use in motor vehicles, and others are cionsidering such regulations. We used an epidemiologic method, the case-crossover design to study whether using a cellular telephone while driving increases the risk of a motor vehicle collision.

Sample and Methods:

 
• The authors studied 699 Toronto drivers who had cellular telephones and who were involved in motor vehicle collisions resulting in substantial property damage but no personal injury.
  
  
• Each person's cellular telephone calls on the day of the collision and during the previous week were analyzed through the use of detailed billing records.
  
  
• The time of each collision was estimated from each subject's statement, police records, and telephone listings made to emergency services.

Major Findings:

• A total of 26,798 cellular telephone calls were made during the 14 month study period. The risk of collision when using a cellular telephone was four times higher than the risk when a cellular telephone was not being used (relative risk, 4.3; 95 percent confidence interval, 3.0 to 6.5)
  
  
• The relative risk was similar for drivers who differed in personal characteristics such as age and driving experience; calls close to the time of collision were particularly hazardous (relative risk, 4.8 for calls placed within 5 minutes of the collision, as compared with 1.3 for calls placed more than 15 minutes before the collision; P<0.001); and units that allowed the hands to be free (relative risk, 5.9) offered no safety advantage over hand-held units (relative risk, 3.9; P not significant). .
  
  
• Thirty nine percent of the drivers called emergency services after the colision, suggesting that having a cellular telephone may have had advantages in the aftermath of an event. .
  
  
• The authors also found that the relative risk of having a crash while using a cellular phone was estimated to be similar to the hazard associated with driving with a blood alcohol level at the legal limit.

Authors' Conclusions:

• The use of cellular telephones in motor vehicles is associated with a quadrupling of the risk of a collision during the brief period of a call. .
  
  
• Decisions about regulation of such telephones, however, need to take into account the benefits of the technology and the role of individual responsibility.

Critical Assessment

This study is suggestive of a relationship between cellular phone use and crashes that merits further experimental inquiry, but it has several limitations as well.

Redelmeier and Tibshirani themselves point out several limitations to their study. They note that causality cannot be inferred from such a study. By way of example, they mention that emotional stress might lead to both increased cellular phone use and decreased driving ability, so that individual calls may have nothing to do with increased crash risk. They also list four weaknesses in their study.

First, only volunteer drivers participated, perhaps leading to underestimates of risk caused by riskier drivers opting out. Second, they point out that people vary in their driving behavior from day to day, though Redelmeier and Tibshirani consider the findings hard to explain in terms of such variations because of consistent findings between the whole sample and a subset of 72 subjects who remembered (up to a year later) having driven during both the hazard period and the control period.

Third, case-crossover analysis does not eliminate all forms of confounding, particularly in regard to temporary conditions, though again the article's authors believe such factors are unlikely to account for the magnitude of association observed. Finally, they point out that collision involvement did not mean the cellular phone owner was judged "at fault". This was left unspecified in the article and the authors indicate that perhaps cellular phone use merely decreases a driver's ability to avoid a collision caused by someone else.

Maclure and Mittleman (1997) point out additional limitations to the study and qualifications to its results. While they applaud the use of the case-crossover design (Maclure was the originator of this approach), they indicate that the use of pilot subject data to adjust for the "intermittency of driving" was not convincing because of possible unmeasured differences between the pilot subjects and the study subjects. They have more faith in the analysis of the 72
people who recollected driving during both periods, though they acknowledge that a relative risk result from this group may be an overestimate due to incomplete participation and faulty memory. Maclure and Mittleman indicate that the lack of a safety advantage for hands-free phones may simply be the result of too little statistical power to test for this effect. The risks associated with placing a call, the risk extinction curve over time after a call ends, and the kinds of collisions that are most likely to increase are all in need of future research, they point out.

To these caveats and critiques, the present authors add the following. While Redelmeier and Tibshirani distinguish between exact and inexact collision time estimates, no separate analysis of the 231 exact cases is reported. The distinction between exact and inexact, once made in the report, is not considered further. Determining the exact time of a collision is difficult.

Contamination across sources (e.g., driver statement is also used in a police record to indicate crash time) may have occurred. The analysis of the 72 people who remembered up to a year or so later that they were driving in both periods is similarly susceptible to memory errors. By any reckoning, the time of collision is subject to numerous sources of error.

Average call length (based on calls placed the week before the collision by this group of subjects) was 2.3 minutes, with 76% lasting 2.0 minutes or less. This suggests a positively skewed distribution with a long right tail, a distribution of mostly short (i.e., less than 2 minute) calls with some calls lasting substantially longer. The importance of this data relates to the fact that the investigators focussed their analysis on 5-minute and 10-minute-long hazard intervals prior to the collision. It is not known if the subject was actually on the cellular phone at the time of the collision.

The study contrasted a time period on the day of the collision with a comparison period on a day preceding the collision. The authors assert that this approach would identify an increase in risk if there were more telephone calls immediately before a collision than would be expected solely by chance. The key measure that was analyzed is termed "relative risk." In words, relative risk was defined as "the probability of having a collision when using a cellular telephone at any time during a 10-minute interval as compared with the probability of having a collision when not using a cellular telephone at any time during a 10-minute interval." (p. 456).

Quantitatively, relative risk is calculated as follows. The following example is an explanation of the "crude" relative risk assessment given on p.455 of the article, as explained by Redelmeier in a phone interview with one of the present report's authors. For that example, the relative risk assessment is based on the following 2 x 2 table:

Presumably, the interpretation is that the baseline risk (not observed or estimated) was the same on the crash day and the preceding day. Therefore, by this line of reasoning, the baseline risk was raised by some multiplier equal to the ratio of the observed cellular phone uses on the crash days and cellular phone uses on the preceding days. Because of the many variables that can affect crash hazard probabilities but that cannot be equated with the case-crossover study design, the authors point out that a causal relationship between cellular phone use and crashes cannot be drawn.

The implication of causality based on relative risk metrics would require very strong assumptions about the equality of baseline risk for each matched-pair in the study on all accounts except cell phone use. Such assumptions may not be plausible unless it can be assured that the situational characteristics (traffic situations, driver states, nature of cell phone use, etc.) were the same across the two days. The implausibility of this is reflected in the fact that an adjustment factor of 35% was used because the subject may not have even been driving during the control period.

The comparison of relative risk of a crash associated with cellular telephones and that associated with drivers with blood alcohol levels at the legal limit deserves special mention. While such a comparison may be appealing from the standpoint of emphasizing the potential adverse consequences of using a cellular phone while driving, it overlooks some important distinctions between the two categories of crashes. First, no causal link has been established between cellular telephone use and crashes. In contrast, the link between driving while intoxicated and crashes is far more clearly established, both in terms of the nature of the influence on driving and the magnitude of the problem.

Second, it must be recognized that cellular telephone use is a transient behavior, lasting, on the average (in this study) 2.3 minutes, with the majority of calls lasting 2 minutes or less. Intoxicated drivers, however, are impaired throughout a trip and thus exposure is likely to be considerably greater. The comparison given in the article would suggest that cellular phone use, per unit time, is actually much more hazardous than driving in an intoxicated state. This finding does not accord with what one might reasonably expect. Thus, the comparison in crash hazard exposure between cellular phone use and driving while intoxicated is specious unless more data than provided in the article are brought forth.

With regard to the lack of an apparent safety advantage of hands-free phones, we would add to Maclure and Mittleman's comment the fact that having such a feature does not mean it (i.e., the hands free feature) was in use at the time of a call. This issue is compounded by the fact that the specific "hands-free" features for a phone can vary considerably, requiring varying levels of interaction on the part of the driver for both dialing and conversation. Thus, the distinction between the two groups in this study may not be that clearcut.

Finally, a threat to the validity of any conclusions suggested by the Redelmeier and Tibshirani study resides in the nature of the study participants themselves. All 699 subjects were cellular phone owners who had a crash. But three other groups of drivers might be logically identified for comparison: cellular phone owners who did not have a crash, non-cellular phone owners who did have a crash, and non-cellular phone owners who did not have a crash.

None of these three other groups were considered in the analysis. It is possible that the study participants represent members who are in some sense atypical of the driving population or of cellular phone owners in general. They may be extreme in the nature of their phone use (e.g., greater frequency of calls, longer calls, more intense dialogue), in their driving style (e.g., more aggressive driving with less margin for error), or even in their human abilities (e.g., less capacity to time-share the driving task and telephoning task). Thus, caution is urged in using the Redelmeier and Tibshirani study results to infer that cellular phone use, in general, is hazardous.

In summary, Redelmeier and Tibshirani's study represents a unique and suggestive investigation of the relationship between cellular telephone use and highway safety. Increasing the current level of understanding of the nature of this relationship awaits future research that helps untangle the many threads of potentially influential factors present in this study.

Serafin, C., Wen, C., Paelke, G., & Green, P. (1993). Car phone usability: a human factors laboratory test. Proceedings of the Human Factors and Ergonomics Society 37th Annual Meeting, 220-224.

Type of Study: Driving simulator.

Keywords: hand-held cellular telephones, hands-free cellular telephones, voice-activated cellular telephones, lane deviation, age differences, head-up display

This paper describes an experiment that examined the effect of cellular telephone design on simulated driving and dialing performance. The results were used to help develop an easy to use cellular telephone interface and to provide task times as input for a human performance model. Twelve drivers (six under 35 years, six over 60 years) participated in a laboratory experiment in which they operated a simple driving simulator and used a cellular telephone. The cellular telephone was either manually dialed or voice-operated and the associated display was either mounted on the instrument panel (IP) or a simulated head-up display (HUD).

The cellular telephone numbers dialed were either local (7 digits) or long distance (11 digits), and could be familiar (memorized before the experiment) or unfamiliar to the subject. Four tasks were performed after dialing a telephone number; two of the tasks were fairly ordinary (listening, talking) and two required some mental processing (loose ends, listing). In terms of driving performance, dialing while driving resulted in greater lane deviation (16.8 cm) than performing a task while driving (13.2 cm).

In addition, the voice-operated cellular telephone resulted in better driving performance (14.5 cm) than the manual cellular telephone (15.5 cm) using either the IP display or HUD. In terms of dialing performance, older drivers dialed 11-digit numbers faster using the voice-activated cellular telephone (12.8 seconds) than the manual cellular telephone (19.6 seconds).

Dialing performance was also affected by the familiarity of numbers. Dialing unfamiliar numbers using the voice-activated cellular telephone was faster (9.7 seconds) than using the manual cellular telephone (13.0 seconds) and 7-digit unfamiliar numbers were dialed faster (8.2 seconds) than 11-digit unfamiliar numbers (14.5 seconds). Thus, the voice-operated design appears to be an effective way of improving the safety and performance of car cellular telephone use, but the location of the display is not important.

Sample and Methods:

 
• Twelve licensed drivers (six men and six women), ranging in age from 20-76 years old.
  

  Participants were divided into two subgroups, six younger (20-35 years old, mean = 24 years) and six older (over 60 years old, mean = 70 years).

  Participants had never used a cellular telephone before.

  Driving conditions in the simulator included nighttime driving on a single-lane, slightly curved road.

  Cellular telephone designs tested were hand-held, hands-free, and voice-activated models mounted on either the instrument panel or displayed in a head-up display on the windshield.

  Participants practiced dialing each type of telephone number once and practiced driving and using the cellular telephone at the same time until comfortable with the task. The tasks to perform in place of telephone conversations were also practiced, in the following order: loose ends, listing, talking, and listening. The communications tasks included "loose ends" (how many unconnected ends are there in a capital letter), "listing" (name as many items in a category as possible in a fixed time period, e.g., "type of furniture), talking (answer the question " What did you do last weekend?"), and listening (i.e., listen to a hypothetical situation and answer multiple choice questions about it). Each task lasted about 30 seconds and all test participants were given the same materials.

• The study examined two types of cellular telephones, manual and voice-operated, combined with two types of displays, instrument panel (IP) display and head-up display (HUD). The IP display was located in the center console of the dashboard, and the HUD was positioned to the left of the driver's view.

Major Findings:

• The main disturbance in driving performance was found during periods of manual dialing while driving, as measured by standard deviation of lane position (16.8 cm versus 13.2 cm for performing a cellular telephone task while driving and 14.2 cm for driving only).

  Voice activated dialing led to better driving performance than the manual handset dialing (14.5 cm versus 15.5 cm).

  The driving performance of older drivers was significantly worse (15.2 cm) than younger drivers (14.2 cm).
  
  

• Dialing performance, as measured by the time it took to enter the digits of the telephone number, was affected by cellular telephone type, length of telephone number, type of telephone number (familiar versus unfamiliar), and age. Only display type did not significantly affect dialing performance.

  Older drivers dialed 11-digit numbers faster using the voice-activated cellular telephone than the manual cellular telephone.

  Participants dialed unfamiliar numbers faster using the voice-activated cellular telephone than the manual cellular telephone, but there was no difference for familiar numbers.

  Participants dialed unfamiliar 7-digit numbers faster than 11-digit numbers, but there was no difference between familiar numbers.

  Older men dialed faster than older women, while younger women dialed faster than younger men.
  
  

• For the loose ends task (i.e., identifying structural features of letters, so that for the letter A, the answer would be "two," since two lines do not connect to other lines at the bottom of the letter), the response time for the letter G was higher than every letter except K. While older women had longer response times than younger women, older and younger men did not differ.
  
  
• 10 of the 12 drivers preferred the voice/head-up-display combination; the other two preferred the voice/instrument panel combination. The least preferred (8 of 12 participants ranked it "worst") was the manual/instrument panel combination.

Author's Conclusions:

• Compared with a manual handset, voice input with a HUD or IP display resulted in less lane position deviation for all drivers and faster dialing times for older drivers dialing unfamiliar numbers.
  
  
• Drivers preferred the voice-operated cellular telephone.
  
  
• Voice-operated designs appear to be an effective way of improving safety and performance of cellular telephone use.
  
  
• Age influenced both driving performance and dialing times, indicating that the older driver should be taken into account in the design of cellular telephones.
  
  
• In the manual/head-up-display combination, there may be location effects between the head-up-display and the handset display location, but the study of this location was restricted by hardware limitations on the simulator, so that the HUD was always positioned to the left of the driver's view.

Critical Assessment:

• Dialing time is a less relevant operational definition of safety for voice-activated cellular telephones than for manual cellular telephones, since visual attention does not need to be diverted from the road. In fact, voice dialing added only 0.3 cm of standard deviation of lane position to baseline driving, compared to 1.0 cm added for manual dialing. It is conceivable that for some drivers, saying a telephone number slowly would be less distracting than trying to say it as quickly as possible.
  
  
• Further operational definitions of driving performance besides lane position need to be studied in order to understand the effects of manual versus voice-activated dialing on driver performance.
  
  
• The practical significance of an age-related 1 cm increase in standard deviation of lane position is unclear.
  
  
• The study was conducted in a driving simulator, with its attendant interpretive problems.
  
  
• Some of the conversational materials (e.g., loose ends) have no apparent analogues in real-world cellular telephone communications.

Serafin, C., Wen, C., Paelke, G., & Green, P. (1993). Development and human factors tests of car phones. (UMTRI Technical Report No. 93-17).

Study Types: Study I &endash; Pilot for Usability Study , Study II &endash; Usability study , Study III &endash; Summary and assessment of all differences between the two versions of the report.

Keywords: cellular telephone interface, cellular telephone functions, button labels, head-up display, hand-held cellular telephones, hands-free cellular telephones, voice-activated cellular telephones, lane deviation, age differences, dual tasks, cognitive interference

Author's Abstract:

This report describes three experiments to develop an easy to use car phone interface. In the first experiment, 19 people at two local secretary of state driver licensing offices gave their preferences for button labels and abbreviations. The following labels (and abbreviations) are recommended: power (Pwr), Call, End, delete (Del), memory (Mem), and recall (Rcl). Twelve drivers (six under 35 years, six over 60 years) participated in the third experiment, a laboratory study, in which they operated a simple driving simulator and used a car phone.

The phone was either manually dialed or voice-operated, and the associated display was either mounted on the instrument panel (IP) or was a simulated head-up display (HUD). Phone numbers dialed were either local (7 digits) or long distance (11 digits), and could be familiar or unfamiliar. In addition, there were four conversational tasks, two of which were fairly ordinary (listening, talking) and two of which required some mental processing (loose ends, listing).

Driver performance (voice&endash;5.7 inches; manual--6.1 inches) and dialing times (voice&endash;9.2 seconds; manual&endash;10.7 seconds) were better with the voice-operated phone than the manual phone using either the IP display or the HUD. In addition, younger drivers outperformed older drivers with regard to both driving (younger&endash;5.6 inches; older 6.0 inches) and dialing performance (younger&endash;7.4 seconds; older&endash;12.6 seconds). Thus, voice appears to be an effective way of improving the safety and performance of car phone use, but the location of the display is not important. The benefits of voice are particularly noticeable for older drivers.

STUDY I

Type of Study: Pilot for Usability Study

Sample and Methods:

• Nineteen licensed drivers (fourteen men and five women), ranging in age from 20-71 years old.

  Participants were fluent in English.

  Participants were recruited while waiting in line at two Michigan secretary of state offices.

  Participants had never used a cellular telephone before.

• A computer interface was used to display a car phone keypad. Function buttons were labelled with question marks. After each function was demonstrated to the participant, the participant was asked to label the associated function key based on the demonstration. This task involved both naming the function and providing an abbreviation for that name that could be applied to the appropriate button.
  
  
• Participants devised labels for the following demonstrated cellular phone functions: power, delete, dial, end, answer, memory, and recall.

Major Findings:

• Across participants, the same labels were chosen most frequently for dial (DIAL), answer (ANSWER), memory (MEMORY), and recall (RECALL) functions.
  
  
• The most common abbreviation strategies that participants employed for labeling functions were truncation (e.g., POW for power) and vowel deletion (e.g., PWR for power).
  
  
• For the functions power, delete, and end, no consensus emerged as to the preferred label choice.

Author's Conclusions:

• Data on drivers' car phone label and abbreviation preferences was collected successfully in the secretary of state driver licensing office setting.
  
  
• No preferred label or abbreviation choice emerged for several functions.
  
  
• Results were used to develop stimulus materials for Study II.

Critical Assessment:

• While driver preferences for cellular phone button labels are an important usability issue, Study I falls outside the major focus of this literature review since no driving performance measures were taken.

STUDY II

Type of Study: Usability study

Sample and Methods:

• Twelve licensed drivers (seven men and five women), ranging in age from 22 to 53 years old.

  All but one participant had never used a cellular telephone before.

  Drivers were employees of the University of Michigan Transportation Research Institute (UMTRI).
  
  

• Drawings of cellular phone interfaces were presented to participants.

  Button labels in the drawing were varied for the following functions: power, call, end, delete, memory, and recall.

  Button functions were abbreviated according to three methods: 1., vowel deletion, 2., truncation, and 3., mixed vowel deletion and truncation. These labeling patterns were presented in counterbalanced order.
  

• Participants were instructed to tell what each button label abbreviation stood for and to state the function of each button. Participants were also asked to explain the sequence of buttons they would push to place a call. Participants were asked to choose a their preference for a button abbreviation method.

Major Findings:

• Among participants who were presented the interface labeled by the vowel deletion method first, two out of twelve participants mistook "DLT,"an abbreviation of "delete," for "dial tone."
  
  
• Among participants who were presented the interface labeled by the truncation method first, three participants mistook the meaning of "REC," an abbreviation of "recall," for "record" or "receive." In addition, two participants misinterpreted "POW," an abbreviation of "power," as a function to retrieve answering machine messages.
  
  
• Eleven of the twelve participants preferred the mixed vowel deletion and truncation method.
  
  
• In the "place a call" task, only one participants named the correct sequence of buttons. The most common error involved pressing "CALL" before dialing, rather than after dialing.

Author's Conclusions:

• People are able to decode cellular phone function abbreviations better when they are encoded according to a mix of abbreviation techniques.
  
  
• Participants' performance in abbreviation decoding and function naming was best when button labels were abbreviated with a combination of truncation and vowel deletion methods.
  
  
• The dual function of the "call" button, sending and receiving calls, was not well understood by novice cellular phone users.

Critical Assessment:

• While driver comprehension of cellular phone button labels is an important usability issue, Study I falls outside the major focus of this literature review since no driving performance measures were taken.

STUDY III

Type of Study: See Serafin, Wen, Paelke, & Green (1993) for a full review.

• Information from the larger report is provided here to supplement the above referenced review. The following points summarize and assess all differences between the two versions of the report.

Sample and Methods:

• Participants were screened for visual acuity at the time of recruitment.
  
  
• Lane position data is presented in standard measurements, while the same data is presented in metric units in the Proceedings of the Human Factors and Ergonomics Society version of the report.

Major Findings:

• Drivers' glance behavior while using the phone in the simulator was analyzed. Only one participant of each age by gender cell was included in the glance analysis, and only data from the instrument panel (IP) condition using long distance telephone numbers was analyzed. This section of the report should be regarded as exploratory.
  
  
• Similarly, data on participants' dialing patterns was included in the larger report. Again, videos of only four participants were used in the analysis, and these were only in the manual dialing condition. Pauses in dialing were counted. Among the four participants, older telephone users inserted more pauses into their dialing, while younger users tended to chunk numbers more often while dialing.
  
  
• Errors in dialing were also reported in the larger report. Since errors were rare events, they cannot be considered statistically meaningful. Counts are given of error types (13 commissions and 3 omissions), errors within each phone configuration (manual IP panel, 7 errors; voice HUD, 4 errors; manual HUD, 3 errors; and voice IP, 2 errors), and whether the participant corrected the error or not.

Stein, A.C., Parseghian, Z., & Allen, R.W. (1987). A simulator study of the safety implications of cellular mobile phone use. (Paper No. 405). Hawthorne, CA: Systems Technology, Inc.

Type of Study: Driving simulator.

Keywords: hand-held cellular telephones, hands-free cellular telephones, voice-activated hand-held cellular telephones, cellular telephone mounting location, lane position, radio tuning, age differences

Author's Abstract:

An interactive driving simulator developed by STI was used to investigate the impacts of different types of cellular telephone design, use, and mounting locations on driver performance in negotiating a simulated route. Along the route drivers were required to negotiate curves, avoid obstacles, and respond to various road signs. Data on lane position and speed were collected along the routes, making it possible to compare driver performance when no task was required (the baseline case) with performance over similar segments while dialing a cellular telephone, receiving a call, or tuning a radio. Seventy-two subjects were segregated by sex, age, and past experience with cellular telephone use in analyzing driver performance.

Sample and Methods:

• 72 participants.

  36 males and 36 females. 24 drivers were under 25 years old; 24 were between 25-55 years old; and 24 were over the age of 55.

  In each young and middle age group, three of the drivers were current cellular telephone users. In the over 55 male group, only one driver was a cellular telephone user. In the over 55 female group, no current cellular telephone users were included.
  
  

• An off the shelf cellular telephone was used, but the model was not specified by the author.
  
  
• The cellular telephones were mounted on the center console and on the dashboard.
  
  
• Participants were trained to use the cellular telephone in a similar way that cellular telephone retailers train their customers, i.e., they were given an operation manual and observed a demonstration of cellular telephone operations.
  
  
• Participants were trained in the driving simulator until they reported feeling comfortable and were able to perform all required tasks without difficulty in all road conditions. A practice run was given before testing began.
  
  
• In order to create a workload representative of urban driving during peak traffic conditions, over the course of the testing session the driver was presented with 12 curves, 20 obstacles, and approximately 50 highway signs, 30 of which required driver response (depressing the horn switch or using the headlight dimmer foot switch). Whenever drivers observed the "airport" symbol sign, they were required to originate a cellular telephone call. All other cellular telephone calls were randomly originated and received on a straight section of road, a straight section of road when an unexpected obstacle was presented, and a curve.
  
  
• · Each driver completed two driving scenarios in which the cellular telephone mounting location varied, dash or center console. Type of cellular telephone also varied (i.e., manual dialing/hand-held handset; memory dialing/hands-free; or voice-activated/hand-held handset).
  
• Subjects dialed by manually keying in the 10-digit phone number plus an enter key, by recalling a number from memory (i.e., pressing "RCL 1"), or by a voice command (i.e., lift handset and say "TRAVEL AGENT"). Placing a call, the driver heard and was required to memorize specific flight information given by a "travel agent"; this information included airline, flight number, originating airport, and destination. On an incoming phone call, the driver had to convey the memorized information.
  
  
• During a drive scenario, the subject was required to originate three cellular telephone calls, receive three cellular telephone calls, and perform the radio-tuning task three times.
  
  
• Dependent measures were as follows: speed, number of "accidents," lane position, lane position variability, and responses to road signs.

Major Findings:

• Secondary tasks did not significantly affect drivers' speed control, increase "accidents" (defined as running off the simulator road or colliding with an object), or increase "speeding tickets" (defined as travelling at least 3 mph above limit during the randomly-determined 30% of the time when a "policeman" was present). Therefore, steering control data was most useful concerning the effects of cellular telephone use on driver performance.
  
  
• On straight roadways without obstacles, the following results were found:

  Driver's tracking ability (standard deviation of average lane position) deteriorated with manual dialing; the console-mounted cellular telephone deteriorated tracking ability more than the dash-mounted cellular telephone when manually dialing.

  While manually dialing with a console-mounted cellular telephone, older drivers' lane tracking variability was 6 inches greater than younger drivers. This variability translates into a 1.63% probability of exceeding the lane boundary, and traffic safety may begin to be adversely affected when this probability exceeds 0.50%.

  When manually dialing a console-mounted cellular telephone, middle aged drivers had a 1.60% probability of exceeding the lane edge, while older drivers had a 7.23% probability of exceeding it.

  When manually dialing a dash-mounted cellular telephone, mid-aged drivers had a 0.68% probability of exceeding the lane edge, while older drivers had a 1.63% probability of exceeding it.

  Across ages and mounting locations, lane tracking variability was significantly greater for radio tuning than for memory dialing a cellular telephone.

  Across task and mounting conditions, older drivers had a greater tracking variability than young drivers, but this difference translated into probabilities of exceeding the lane edges that were not significant in practical terms.

  The voice-activated dialing task was no more difficult, i.e., had no greater lane tracking variability, for one age group than another.

  Voice-activated dialing resulted in less lane tracking variability than the radio tuning. A driver's probability of exceeding lane boundaries while dialing with voice-activation dialing was less than 0.10%.

  When receiving calls, older drivers' ability to maintain lane position was worse than middle aged drivers, and middle age drivers were worse than younger drivers. While this difference was significant, the practical probability of exceeding the lane boundary was consistently less than 0.10%.

  Similarly, when receiving calls, drivers' ability to maintain lane position was best with no secondary task, worse when tuning the radio, and worst when receiving a call with either a hand-held or hands-free cellular telephone. Again, while these differences were significant, the practical probability of exceeding the lane boundary was consistently less than 0.10%.

  For old-age drivers, the effect of radio tuning on tracking variability reached the threshold of concern (0.50% probability of crossing lane boundary) from a traffic safety standpoint.

• · On curvy roadways without obstacles, the following results were found:

  Manual dialing resulted in greater lane tracking variability than did radio tuning. All other dialing and answering tasks caused less lane tracking variability than the radio task.

  Old-age drivers had significantly greater lane-tracking variability than mid-age or young drivers during manual dialing and memory dialing, but not for voice-activated dialing.
  
  

• On straight road segments with obstacles, the probability of colliding with obstacles or running off the road was calculated based on lane position variability. Actual collisions were rare. The following results were found:

  Manual dialing performance was consistent with straight-segment-without-obstacle performance.

  Mid-age and old-age drivers had a five-fold increase in the probability of striking the obstacle when they dialed with center console mounted cellular telephones. However, the probability of striking the obstacle only increased by one half when using the dash-mounted cellular telephone.

  Dialing a memory-dial cellular telephone did not increase the accident probability for any age group.

  Dialing with voice-activated cellular telephones increased mid-aged drivers' probability of running off the road.

  When answering the cellular telephone, old-age drivers were two to five times more likely to have an accident when answering a call with the hands-free option than with hand-held.

  Old-age drivers were 5 to 7 times more likely to be involved in an obstacle accident while tuning the radio.

Author's Conclusions:

• Consistent with other research, as a driver's age increases, any task competing for attention interacts with age to impair driving ability.
  
  
• Although manually dialing a ten-digit number can substantially increase the risk of accident involvement over that imposed by radio tuning, this risk may be reduced by appropriate cellular telephone location. Cellular telephones should be mounted in locations that are as close to the driver's line of sight as practical.
  
  
• Both voice-activated dialing and memory dialing are less hazardous than tuning a radio. In fact, for younger drivers, keying a single-digit memory location to enter a number, and then pushing the send button, was not appreciably more hazardous than the baseline condition of unencumbered driving.
  
  
• Dash-mounted cellular telephones may reduce the probability of accident involvement, compared to center console mountings. Further studies are needed to investigate the effects of other locations, such as the transmission hump, which is forward of the console and more in the driver's line of sight [but could be a further reach].
  
  
• Although hands-free operation showed no advantage over handset use when answering the cellular telephone, it could be helpful in emergency situations, since both hands could be free to respond to the emergency.
  
  
• Minimal training allows adequate carry-over from home telephone systems to the operation of cellular telephones.
  
  
• Use of voice-recognition dialing technology should be encouraged.
  
  
• The memory dialing capability of the cellular telephone should be used, but drivers should be instructed not to refer to written lists of memory locations while driving.

Citical Assessment:

• This study includes some useful descriptions of the available cellular telephone technology.
  
  
• Although part of the drivers' task involved responding to traffic signs by honking the horn, dimming the headlights, or originating a call, results are not reported as to whether drivers followed these directions. Although this part of the task may have diminished participants' sense of the study's realism, results may still have been useful in the sense that they could have shown whether drivers actually paid attention to roadway conditions in the simulator.
  
  
• As a secondary workload during cellular telephone calls, drivers were asked to memorize a script during the first cellular telephone call of a sequence and to repeat the information during the next call. The inter-trial interval was not reported, which could have a significant effect on the amount of information the individual is able to recall. Additionally, the explicit demand to learn the task allowed participants free reign in choosing encoding and retrieval strategies, whereas other studies have placed more quantifiable cognitive demands (e.g., serial addition or listing) upon participants. Therefore, the degree of interference imposed by telephoning is hard to gauge.
  
  
• No differentiation was made between data collected while dialing/answering and engaging in the ensuing conversation. Collapsing data across both tasks makes it hard to distinguish whether the dialing/answering method or the cellular telephone type had the observed effect on lane position. The effect of conversing is also obscured. Separate measurements should be made, and a fuller range of cellular telephone feature combinations should be utilized, e.g., voice-activated hands-free and memory-dialed hand-held.
  
  
• The simulator scene is described as "a rural roadway, at dusk, under somewhat reduced visibility." However, in combination with secondary tasks, the simulation task is described as "create[ing] a workload representative of urban driving during peak traffic conditions." Urban driving has been associated with a higher workload than rural driving (Brookhuis et al., 1991). However, without physiological or secondary task measures, driving task workload is hard to quantify. Caution should be also be taken in generalizing results from computer-generated simulator displays to specific, (and different), real-world environments.
  
  
• All conclusions related to radio tuning must be viewed in light of the fact that participants were required to visually search for a particular radio frequency on the dial. This task may be comparable to tuning to a frequency to obtain airport traffic information without the benefit of search and scan features commonly found on car radios today. This methodology differs from other studies in which drivers simply matched a station that another radio played by listening and moving the dial. The radio results may be best considered as time in which visual attention was diverted, but should not be taken as representative of radio use.
  
• The author asserts but does not support a number of statements such as: "if a driver uses his telephone...as an electronic 'scratch pad' to record new telephone numbers, the activity is just as dangerous as the manual dialing task;" or, "in practice, however, on any given trip more time would be devoted to telephone use than radio tuning."

Tijerina, L., Kiger, S. M., Rockwell, T.H., & Tornow, C. (1995). Workload assessment of in-cab cellular phone use by heavy vehicle drivers on the road. (NHTSA Contract DTNH22-91-C-07003).

Type of Study: On-road research with an instrumented vehicle.

Keywords: cellular telephones, commercial vehicle operation, on-road measures

Author's Abstract:

This study assessed the driver workload imposed by a cellular phone on heavy vehicle drivers under various driving conditions. Sixteen (16) professional commercial vehicle operation (CVO) licensed drivers drove an instrumented heavy truck over a 4-hour period on public roads under various conditions of ambient lighting (day or night), traffic density (light or heavy), and road type (divided or undivided). Within driving conditions combinations, cellular phone dialing, radio tuning, and communications dialogue were completed by the driver. Continuous measures were taken of visual allocation, steering and accelerator activity, speed maintenance and lane-keeping performance. Results of in-vehicle device use are presented and provide insights into useful workload measures and methods, as well as a contribution to the literature on cellular telephone system ergomatics.

Sample Studied:

• 16 professional heavy vehicle drivers. Participant ages ranged from 32-60 years old, with a mean age of 47.2 years. None were regular cellular telephone users.

Apparatus

• The test vehicle was a 1992 Volvo/White GMC conventional tractor with a sleeper and a 53 ft. 1993 Fruehauf dry freight van semi-trailer loaded to bring the gross vehicle weight to approximately 76,300 pounds.
  
  
• The test vehicle was equipped was standard engine gauges, CB radio, and AM/FM stereo radio. The vehicle was also instrumented with an array of sensors, including an automatic lane tracker, speed sensor, steering sensor, accelerator pedal angle sensor, and a storage device to capture all sensor data.
  
  
• Four video cameras and two VCR's were used to capture data on driver visual allocation, manual allocation, and the road scene ahead.
  
  
• A text message display consisting of a 7-inch diagonal VGA-compatible green-phosphor CRT was mounted on the top of the instrument panel to the right of the seated driver.
  
  
• A Motorola Model No. 19017NAABB black cellular telephone was installed to the right of the seated driver.
  
  
• The telephone included a recall feature to dial stored numbers.
  
  
• Manual dialing was completed by pressing the "send" button.
  
  
• An on-board experimenter was present at all times.
  
  
• The test route was arranged to encompass variations in lighting (dark vs. light), road type (divided vs. undivided), and traffic density (high vs. low).
  
  
• Each driving condition phase occurred over roughly a 10-15 minute period, about one each hour of a 4-hour run.
  
  
• Drivers were prompted to use the cellular telephone by text messages on the CRT display.
  
  
• During each of the four driving condition phases, eight est messages and two dialogues were presented.
  
• One dialogue had an anticipated low workload (driver required to answer biographical questions such as name, address, birth date) while the second dialogue had an anticipated high workload (questions requiring single math calculations to determine fuel needed, time required to reach a destination, etc.).
  
  
• Manual task data (auto-dial, 7-digit, and 10-digit manual cellular telephone dialing, plus radio tuning as a control condition) and cognitive task data (the two question-and-answer dialogue tasks with open-road driving as a control condition) served as measures of driver performance.

Major Findings

• 7-digit and 10-digit dialing took more glances, on average, for completion than radio tuning, with greater total time the eyes are off the road and on the device.
  
  
• Mean number of steering holds was greater during radio tuning and 10-digit dialing then for auto-dialing and 7-digit dialing.
  
  
• Smallest number of steering reversal occurred during auto-dialing compared to radio tuning, -7-digit and 10-digit dialing.
  
  
• The impact of road conditions (divided vs. undivided) on driving performance was found to be secondary to the manual task effects, but did have some impact on driving measures during cognitive task execution.
  
  
• Manual dialing did not have an effect on speed variance or lane-keeping measures.
  
  
• Mirror sampling reduced from a little over 12 percent while engaged in open road driving only to about 6 percent while drivers were engaged in a dialogue.
  
  
• No degradation of driving performance in terms of speed maintenance and lane keeping measures were noted during execution of the cognitive tasks.

Author's Conclusions

• Driver visual allocation measures can be sensitive to variations in in-vehicle workload with heavy vehicle drivers.
  
  
• Simple biographical question-and-answer dialogue reduced the number of mirror glances professional drivers took.
  
  
• Reduced mirror glances resulted in reduced situation awareness of driving conditions to a degree.
  
  
• Drivers generally showed excellent speed control over the vehicle during in-cab use of the cellular telephones, but lane-keeping, in terms of lane exceedances and related measures, was degraded.
  
  
• This suggests that cellular telephone dialing can be disruptive to lane-keeping as a measure of drivers' performance.
  
  
• In-vehicle device workload can degrade highly over learned vehicle control skills of lane-keeping, as well as decrease drive monitoring for crash hazards ahead.

Critical Assessment

As the authors point out, "high" traffic density for this study may have been relatively low as compared with other locations (e.g. metropolitan Washington, D.C.). Therefore, even though traffic density did not have an effect of workload measures in this particular study, traffic density may effect workload measure in other geographical locations.

This study benefitted from the use of several measures of driving performance, such as steering and accelerator activity, speed maintenance, and lane-keeping performance.

Trinkaus, J. (1990). Usage of cellular telephones: an informal look. Perceptual and Motor Skills, 71, 1375-1376.

Type of Study: Brief report, citing one systematic field observation study (The cellular marketplace: 1990. Washington, DC: Economics and Management Consultants International, 1990.)

Keywords: hand-held cellular telephones, systematic field observation

Author's Abstract:

The growth in the usage of cellular telephones by motor vehicle operators appears to be happening in the absence of any companion governmental operating directives. The result is a state of the world wherein drivers are mandated to be securely fastened to seats of defect-free vehicles, yet are permitted to drive with one hand on the steering wheel, while dividing their attention between road conditions and an interactive conversation.

Sample and Methods:

• Naturalistic observation of drivers in a northeastern city having about 210,000 cellular subscribers. This estimate is from The Cellular Marketplace, (1990).
  
  
• 50 one-hour convenience citings were conducted during morning and evening rush hours, Monday through Friday, of vehicles moving north on a four-lane one-way road. Traffic flowed at a rate of about 1600 vehicles per hour.
  
  
• Drivers observed holding a cellular telephone were counted.

Major Findings:

• During the 50 hours of observations, two drivers were noted to be using what appeared to be cellular telephones.
  
  
• Both operators, in moving private vehicles, were holding the instrument in the left hand while the right hand was on the steering wheel.

Author's Conclusions:

• Assuming industry projections for growth of cellular subscribers to be accurate, all other things being equal, a replication of this study done five years later should yield 6 drivers using cellular telephones. While seemingly not a large number, when viewed as an indicator of the magnitude of a count of all such drivers, it assumes meaning.
  
  
• It is difficult to understand accident abatement efforts that mandate seat belts but permit drivers to drive with only one hand on the wheel.
  
• Cellular telephone industry growth should be accompanied by a suitable ancillary program to evolve means for the safe application of its products.

Critical Assessment:

• Counting the number of drivers holding cellular telephones does not give a true measure of cellular telephone use, given the availability of hands-free technology (although we are not sure how prevalent this technology was in or before 1990). Since hands-free designs are becoming more prevalent, the author's five-year projection is not valid.
  
  
• Two drivers were reported to hold "what appeared to be cellular telephones." In fact, did observers see anyone using a cellular telephone? Could it have been the same driver on two occasions?
  
  
• Given that two drivers were observed with hand-held cellular phones over 50 hours on a road where traffic progressed at a rate of 1600 vehicles per hour, these drivers were 2 out of 80,000 drivers. This is not significant.
  
  
• Other people besides these two hand-held cellular telephone users undoubtedly were driving with only one hand on the wheel--the author's measure of safety. In fact, one-handed driving may not be a valid measure of driving performance and/or safety. Drivers shifting in manual transmission cars come to mind in this regard, though two-handed steering for emergency swerve maneuvers is almost certainly a necessity.
  
  
• Reviewing the cited study may clarify a number of methodological points that were not included in this brief report, e.g.:

  &endash; Was the traffic videotaped, or did observers simply record what they saw in real time?

• No measure of inter-observer reliability is given. In published field observation studies, it is generally accepted that two observers should report the same observations at least 80% of the time.
  
  
• What does this sample represent? Business people travel during rush hour and tend to use cellular telephones more. In fact, it was noted that half the vehicles were commercial.
  
  
• This article appears to rely most heavily on scare tactics.

Violanti, J. M., and Marshall, J. R. (1996). Cellular phones and traffic accidents: An epidemiological approach. Accident Analysis and Prevention, 28.

Type of Study: Epidemiological case-control accident study.

Keywords: cellular phones, traffic accident risk, epidemiology, driver inattention

Author's Abstract:

Using epidemiological case-control design and logistic regression techniques, this study examined the association of cellular phone use in motor vehicles and traffic accident risk. The amount of time per month spent talking on a cellular phone and eighteen other driver inattention factors were examined. Data were obtained from: (1) a case group of 100 randomly selected drivers involved in accidents within the past two years, and (2) a control group of 100 randomly selected licensed drivers not involved in accidents with the past ten years.

Groups were matched on geographic residence. Approximately 13% (N=7) of the accident and 9% (N=7) of the non-accident group reported use of cellular phones while driving. Data was obtained from Department of Motor Vehicles accident reports and survey information from study subjects. We hypothesized that increased used of cellular phones while driving was associated with increased odds of a traffic accident. Results indicated that talking more than fifty minutes per month on cellular phones in a vehicle was associated with a 5.59-fold increased risk of a traffic accident.

The combined use of cellular phones and motor and cognitive activities while driving were also associated with increased traffic accident risk. Readers should be cautioned that this study: (1) consists of a small sample, (2) reveals statistical associations and not causal relationships, and (3) does not conclude that talking on cellular phones while driving is inherently dangerous.

Sample and Methods:

• A case group of 100 randomly selected drivers involved in accidents within past two years (1992 -93).
  
  
• A control group of 100 randomly selected drivers not involved in accidents within past two years (1992-93).
  
  
• Groups were match on geographic residence.
  
  
• Method employed an epidemiological case-control design which focused on the presence or absence of risk factors rather than outcomes. In the design randomly assigned persons who had accidents were assigned as "cases" and those who had no accidents as "controls." Being a case or control represented the independent variable. To assess risk factors, the frequency of attention diverting driver behaviors and other factors which might affect the association between cellular phone use and accidents were measured. The presence or absence of these factors were represented as dependent variables.
  
  
• A survey was sent to each case and control subject and accident information was obtained from the Department of Motor Vehicles accident reports. Demographic information as well as 18 possible driver inattention behaviors was requested, including drinking beverages, smoking, talking with others in the vehicle, adjusting seats or mirrors, and cellular phone use. Information of phone use was measured from monthly cellular phone bills. Phone time use was categorized based on the median time per month (i.e., casual, business, intense). Sixty percent (60%) of case and 77% of control subjects responded to the survey.

Major Findings:

 
• Results indicated that talking more than fifty minutes per month on cellular phones in a vehicle was associated with a 5.58-fold increased risk in a traffic accident.
  
  
• Descriptive analysis indicated that cellular phone users who had accidents on average were younger, had less driving experience, and more previous accidents than non-accident subjects.
  
  
• Accident subjects spent approximately twice the number of minutes per month talking on phones than did non-accident subjects and appeared to engage in considerably more business and intense business calls.
  
  
• Talking with others in vehicle, watching scenery, and drinking beverages appeared to be the most often reported driver inattention behaviors by both accident and non-accident subjects.
  
  
• Use of a cellular phone in a vehicle while simultaneously performing other behaviors also was associated with increased odds of a traffic accident. Most significant was the combination of talking on the phone while drinking a beverage, lighting a cigarette, or taking one's hand off the steering wheel. Combined driver behaviors had lower significant odds for an accident than the use of a cellular phone alone.

Author's Conclusions:

• The combined use of cellular phones and motor and cognitive activities while driving were associated with increased traffic risk.

Critical Assessment:

• Study consists of a small sample size.
  
  
• The epidemiological case-control method is prone to potential sources of bias, although efforts were made to minimize selection bias.
  
  
• The Department of Motor Vehicles did not provide evidence that persons were actually using a cellular phone at the time of the accident and the question was not included in the survey.
  
  
• Findings suggest statistical, not causal relationship between cellular phone use and accidents.
  
  
• Reported differences may reflect driver personality or cognitive differences rather than cellular telephone effects.

Zwahlen, H. T., Adams, Jr. C. C., & Schwartz, P. J. (1988). Safety aspects of cellular telephones in automobiles. Proceedings of the ISATA Conference, Florence, Italy.

Type of Study: Closed-course test track with an instrumented vehicle.

Keywords: manual dialing, cellular telephone mounting position, lane deviatio

Abstract:

Two independent studies consisting of ten young and healthy drivers each were used to investigate lateral path deviations when driving in a straight path while dialing in a long distance telephone number on a cellular telephone under four different experimental conditions (cellular telephone mounted in a high position or low position inside car, driver permitted or not permitted to look at road while dialing the cellular telephone). Findings of increased standard deviation of lateral position while dialing suggest that dialing while driving is unacceptable from a driver safety point of view. Design enhancements should be investigated to afford safer and more efficient cellular telephone use.

Sample and Methods:

• Each of the two studies used 10 participants, for a total of 20 drivers.

  Study 1 used 10 males, mean age 23.6 years. These participants had an average of 5.9 years driving experience, and averaged driving 8,900 miles per year.

  Study 2 used 5 males and 5 females, mean age 20.8 years. These participants had an average of 4.4 years driving experience, and averaged driving 5,400 miles per year.

• The test track was located on an unused airport runway.
  
  
• Drivers were to align the longitudinal centerline of the vehicle with the centerline of the runway and drive in the straightest path possible from a standstill to 40 MPH while dialing a cellular telephone.
  
  
• A standard push button telephone was used for the telephone tasks, however, model type was not specified by the authors. This was not a cellular telephone.
  
  
• The telephone task consisted of dialing an 11-digit long distance telephone number. Telephones were mounted in a high position (keypad vertical on the top portion of the dashboard face) or in the low position (keypad horizontal to the car seat). Subjects were to either look continuously and directly at the telephone while dialing the long distance number and concurrently maintaining vehicle position over the centerline on the roadway, or to look at the roadway as often as needed in order to maintain vehicle position while dialing the telephone number.
  
  
• The telephone number (11 digits) was read by the driver from a piece of paper located near the cellular telephone at the beginning of each run. One experimenter sat beside the driver in order to monitor the looking behavior, while a second experimenter seated in the back seat recorded the task completion time.
  
  
• Each participant completed each of the four conditions (2 mounting locations and 2 looking conditions) five times in a randomized order, for a total of 20 runs per person.
  
• ·An experimenter was present inside the vehicle with the participant during the testing session.
  
  
• Lateral path deviation was measured in inches for a distance of 675 feet. A device was attached to the rear bumper at the car's midline. This device dripped liquid dye to indicate the car's path. Every 15 feet, a measurement was taken from the centerline of the runway.
  
  
• Differences in methodology between the first and second study included use of different test vehicles (a Pontiac and a Plymouth) and different telephone manufacturers (although the telephones had identical dimensions).

Major Findings:

• Drivers were able to most closely maintain a straight path along the centerline when the telephone was mounted in the high vertical position and the test drivers were permitted to look at the runway while dialing.
  
  
• The greatest lateral deviation occurred when the telephone was mounted in the low horizontal position and the drivers were not permitted to look at the runway during dialing.
  
  
• Maximum lateral path standard deviation (38.13 inches in Study 1 and 40.69 inches in Study 2) occurred during dialing, when drivers were not permitted to look at the runway and the cellular telephone was mounted in the low position.
  
  
• The lateral path standard deviation at the dialing completion time for the four experimental conditions ranged from a low of 7.01 inches for the looking high condition to a high of 25.24 inches for the drivers not looking at the runway, low position condition.
  
  
• The position of the telephone in the vehicle as well as the visual restraints (permitted or not permitted to look at the runway) had an effect on the lateral deviation of the vehicle.
  
  
• Vehicle also appears to influence the lateral mean position. The Pontiac in Study 1 had mean values that were all on the left of the runway centerline while the Plymouth in Study 2 had mean values that were all on the right of the runway centerline.
  
  
• Cumulative distributions for the dialing completion distances in both studies appear to indicate that the completion distances are fairly normally distributed with respect to the distance required to complete the dialing of the long distance telephone number. Mean completion time (distance) for all the conditions combined in Study 1 is 8.7 seconds (510.4 feet) with a standard deviation of 1.67 seconds (97.9 feet). The mean completion time (distance) in Study 2 for all conditions combined is 9.5 seconds (557.6 feet) with a standard deviation of 1.77 seconds (103.7 feet).
  
  
• Drivers looked at the road ahead an average 2.2 times while dialing the push-button telephone when the telephone was mounted in the low position. When the telephone was mounted in the high position, drivers looked at the road ahead an average 2.9 times while dialing. When drivers were permitted to look at the road, they chose not to on 13 of the 50 runs when the telephone was in the high position. However, when the telephone was in the low position and drivers were permitted to look at the road, they chose not to on only 3 out of 50 runs.
  
  
• A combined, overall standard deviation of the lateral lane position for both studies combined was calculated to be 15.41 inches over 675 feet.

Author's Conclusions:

• Based upon the combined, overall standard deviation of lane position, one out of every 52 cars (1.9%) would laterally deviate out of the driving lane at any point in time while dialing the telephone (for a six foot wide car travelling in a 12 foot wide lane, under nearly ideal conditions). Narrower lanes or worse conditions undoubtedly would increase swerving. This estimate may be conservative, since the maximum standard deviation may be as high as 40.69 inches, which occurred when drivers were not permitted to look at the runway while dialing and the cellular telephone was mounted in the low position.
  
  
• Based on the results obtained regarding number and seconds of looks inside the car to dial the telephone number, a model to be used in the design of keypads and telephone placement is presented. This model shows that if more than four looks are required to dial the telephone number, then it is considered unacceptable from a driver safety and performance point of view, as it relates to lane exceedance probability. Also, the length of time required per look should not exceed 2.0, 1.8, 1.6, and 1.4 seconds, for 1, 2, 3, and 4 looks respectively.
  
  
• Regardless of telephone mounting position, the risk of laterally deviating out of lane while dialing the cellular telephone appears too great. Design enhancements such as different input keys near or on the steering wheel, interlock devices to prevent the use of the telephones in curves and/or heavy traffic, or voice recognition input devices, need to be investigated to afford safer and more efficient telephone use.

Critical Assessment:

• The manipulation of instructions so that drivers are not permitted to look at the road may tend to inflate results, e.g., the finding of a 40.69 inch standard deviation of lane position for the low-mounted position. While a manually-dialed telephone may compete with the roadway for a driver's attention, it does not forbid the driver from looking at the road ahead. Therefore, manipulation of instructions does not bear much similarity to events that occur in the real world of driving and telephone use.
  
  
• Additionally, how did the experimenters insure that drivers didn't even peek when they were not supposed to? It is assumed that drivers were videotaped; however, the method by which videotapes were analyzed is not explained. This information would also help to understand the results pertaining to number of looks.
  
  
• This study concentrated on effects of dialing on performance, but did not include a subsequent conversation.
  
  
• It would be interesting to learn whether the liquid dye method of recording lane position was ever validated against another measure of lane position, e.g., a videotape or an instrumented vehicle that automatically detects lane tracking. Dye that evaporates in a few minutes, though convenient, does not appear particularly reliable.