Accident Analysis and Prevention 37 (2005) 902–909

Mobile phone use—Effects of handheld and

handsfree phones on driving performance
Jan E.B. Törnros ∗ , Anne K. Bolling
VTI, Olaus Magnus väg 35, S-58195 Linköping, Sweden

Received 3 March 2005; received in revised form 10 April 2005; accepted 11 April 2005

Abstract

The study was concerned with effects of handsfree and handheld mobile phone dialling and conversation in simulated driving. In the
main experiment dealing with conversation, 48 participants drove a distance of about 70 km on a route which led through urban and rural
environments. In the dialling experiment, the participants drove a distance of 15 km on a rural two-lane road. The experimental design was
mixed with phone mode as a between-subjects factor and phone use (yes/no) as a within-subjects factor.
Performance on a peripheral detection task (PDT) while driving was impaired by dialling and conversation for both phone modes, interpreted
as an increase in mental workload. Driving performance was impaired by dialling—lateral position deviation increased in a similar way for
both phone modes. Conversation had, however, opposite effects—lateral position deviation decreased in a similar way for both phone modes.
Driving speed decreased as an effect of dialling with the greatest effect for handsfree phone mode. Conversation also caused reduced speed,
but only for handheld phone mode. The effects on speed can be interpreted as a compensatory effort for the increased mental workload. In
spite of the compensatory behaviour, mental workload was still markedly increased by phone use.
Subjective effects of dialling and conversation were also analysed. Most participants reported a speed decrease but no effect on lateral
position deviation as an effect of dialling or conversation. In the conversation experiment, driving performance was rated better for handsfree
than for handheld mode. In the dialling experiment, no difference between the two phone modes appeared.
© 2005 Elsevier Ltd. All rights reserved.

Keywords: Mobile phone; Handheld; Handsfree; Driving; Simulation

1. Introduction simulator study. According to mental effort ratings by the

An issue that has been under considerable debate is
the importance of telephone mode for the effects of tele-
phoning while driving. Since the physical distraction may
be less for handsfree phones than for handheld phones,
most international legislation against the use of a mobile
phone while driving also concerns handheld phones only.
However, both phone modes can cause cognitive dis-
traction which may have negative consequences (Rospa,
2002).
Comparisons have been made between the two phone
modes. Burns et al. (2002) compared the effects of mobile
phone conversation (handheld or handsfree) in a driving
∗ Corresponding author. Tel.: +46 13 204 000; fax: +46 13 141 436.
E-mail address: jan.tornros@vti.se (J.E.B. Törnros).
participants, driving while using a handheld phone was
the most difficult. The drivers slowed down when talking
on the handheld phone compared to the control condition,
interpreted as a compensation for the increased workload.
The standard deviation of speed and speed error measures
indicated that the drivers had the poorest speed control when
using the handheld phone. Reaction times to warning signs at
the roadside were slower for drivers using a handheld phone
compared to the control condition. The drivers also missed
more target signs when they were using a phone compared to
the control condition. No effect on lane keeping performance
was detected, however. Haigney et al. (2000) also found that
conversation with a mobile phone (handheld or handsfree)
during driving was associated with reduced speed. The
lateral control of the car was impaired with a handheld phone
(increased instances of driving off the carriageway) but not

0001-4575/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2005.04.007
 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909
with a handsfree phone. Brookhuis et al. (1991) investigated
the effects of using a mobile phone while driving on a quiet
rural road, a busy ring road and in a town. The phone was
either handheld or handsfree. Conversation caused increased
mental workload according to self-reported effort ratings. It
also reduced the lateral position deviation, which the authors
interpreted as an effect of increased alertness. Dialling a
telephone number, however, caused increased lateral position
deviation with the worst result for handheld phone. Matthews
et al. (2003) compared handheld and handsfree phones with
respect to subjective workload during conversation while
driving on a rural highway. Total subjective workload was
higher for handheld than for handsfree phone. No difference
appeared for mental demand, whereas physical demand was
higher for handheld phone. Strayer et al. (2004) found that
conversation via handsfree or handheld phone mode delayed
the brake reactions of their participants in simulated driving.
No difference was noted between the two phone modes.
Using an instrumented car driven on a motorway, Patten
et al. (2004) studied the effects of using a mobile phone
while driving. Effects of conversation type (simple versus
complex) and telephone mode (handheld versus handsfree)
were compared to baseline conditions. A peripheral detec-
tion task (PDT) was employed to gauge mental workload.
The results were that the PDT reaction times increased
when conversing, but no benefit of handsfree units over
handheld units was found. The content of the conversation
was far more important: the more difficult and complex the
conversation, the greater the increase of PDT reaction time.
It was also found that the driving speed was reduced by
using a handheld mobile phone, whereas no speed reduction
was apparent for the handsfree mode. Conversation type
did not, however, seem to have any marked effects on
driving speed.
It can be concluded that where the two modes have been
compared, either no difference is apparent or there is a
tendency to suggest that the use of a handheld phone would
interfere more with the driving task than a handsfree phone.
Another importing finding is that drivers tend to slow down
during conversation, maybe more so for the handheld mode,
a result which could be interpreted as a compensation for the
increased workload. Dialling a phone number during driving
may have more serious consequences than conversation.
These issues need further study, and one of the major aims of
the present study is to compare the two modes with respect
to effects on measures related to traffic safety. Driving
behaviour as well as mental workload are important in this
respect.
It is further of interest to compare the two phone
modes with regard to subjective effects. If the experience
of the users is different from what is objectively true,
there may be a need to inform the users about the true
effects of using the different modes of phones while
driving.
The effects of conversation and the effects of dialling were
studied in two separate experiments.
903
2. Method
2.1. General design
Two separate experiments were performed, the main one
dealing with conversation and the other dealing with dialling.
The same individuals participated in both. The experimen-
tal design was mixed—half the participant group was tested
with and without handsfree phone use, and the other half was
tested with and without handheld phone use (for a complete
description of the design and other details of the study, see
Kircher et al., 2004).
2.2. Participants
The participants in the conversation experiment were 48
healthy volunteers, aged 24–54 years with mean age 34 years.
They had held a driving licence for 5–35 years. They had
driven about 1000–70,000 km last year. The participant group
was randomly divided into two groups of equal size. One
group used a handsfree phone and the other group used a
handheld phone. There were about as many male as female
participants in both groups −12/12 in the handsfree group,
and 14/10 (more males) in the handheld group.
The majority of the 48 participants in the conversation
experiment also participated in the dialling experiment (23
in the handsfree and 19 in the handheld part).
2.3. Driving simulator
The well validated VTI driving simulator II was used in the
study. It is used to create realistic sensations in a laboratory
environment, and includes a cut-off vehicle cab, a comput-
erised vehicle model, a high-fidelity moving base system, a
vibration table, a wide-angle visual system and an audio sys-
tem (Nilsson, 1993). The car body used in this experiment
was a Volvo 850 with manual gear box.
2.4. Route
Two separate routes were used in the study, a long one
of varying complexity for the conversation experiment and
a short one for the dialling experiment. The long route had
rural sections with varying speed limits (70 and 90 km/h) and
urban sections with speed limit 50 km/h and with varying
complexity. The route was driven through twice making a
total driving distance of about 70 km.
The dialling experiment used a rather straight two-lane
road with speed limit 110 km/h. There was sparse meeting
traffic on the road. The driving distance was about 15 km.
2.5. Phone task
The mobile phone used in the study was a Nokia 6310 with
a small screen showing still pictures in black and white. The
904 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909
phone task in the conversation experiment was a demand-
ing paced serial addition task (Brookhuis et al., 1991). In
the handsfree part the drivers used a handsfree mobile phone
(the telephone always remained in the holder attached to the
dashboard and the driver conversed via loudspeakers) while
driving in the simulator. Only for accepting and ending the
call did the driver have to reach over to touch the corre-
sponding buttons on the phone. In the handheld condition
the telephone had to be taken in the hand for use. While on
stand-by, the telephone was placed in the holder, and when
the phone rang the driver had to pick it up and hold it to his/her
ear while talking. After finishing the call the telephone was
put back into the holder. When receiving a phone call, the par-
ticipant had to perform the paced serial addition task. Each
phone call lasted about 1 min.
Dialling was done either in handsfree mode or in handheld
mode. The driver was requested to make a phone call when
the word “Ring” in large yellow letters was projected in the
centre of the simulator screen. The phone number was glued
to the telephone.
2.6. Experimental design
In the conversation experiment, each participant received
a total of 10 phone calls at pre-selected locations during the
test drive. Comparisons were made with data collected at the
same locations (the route was driven twice) where no phone
call was made, serving as a control condition. The effect of
presentation order was controlled through counterbalancing.
Data collected during conversation (excluding picking up the
phone, starting and finishing the call, and replacing the phone
into the holder) was used in the analyses.
In the dialling experiment only one drive was made during
which the participant dialled a phone number three times.
Comparisons were made between those occasions and the
part of the drive when the phone was not used, serving as a
control condition.
2.7. Effect measures
The following performance measures are analysed in the
present article:
1. PDT (peripheral detection task): reaction time to detected
stimuli and percentage missed PDT signals as a measure
of mental workload (van Winsum et al., 1999). The drivers
respond to visual stimuli presented off centre of their for-
ward view by pressing a micro switch placed on the finger.
If a driver reacted to the PDT stimulus within two seconds,
the reaction was scored as a “hit”, and reaction time was
measured. Otherwise it was scored as a “miss”.
2. Standard deviation of lateral position (SDLP).
3. Average driving speed.
The following subjective measures are analysed in the
present article:
1. RSME (rating scale mental effort), a self-reported mea-
sure of mental effort on a continuous unidimensional scale
with eight defined points along the scale—from “abso-
lutely no effort” to “extreme effort” (Zijlstra and Van
Dorn, 1985).
2. Influence of mobile phone use on speed and lateral posi-
tion.
3. Opinion about driving performance during phone use—a
simple scale with defined endpoints (“much worse than
normal”, “much better than normal”) and the middlepoint
(“as normal”).
It should be pointed our that the RSME ratings, made after
the test drives, referred to the whole test drive irrespective
of whether the phone was used or not. Effects of using the
phone (difference between using the phone and not using
the phone) could therefore not be analysed for the RSME
measure. Comparisons based on RSME ratings could only be
made between the two driver groups (handsfree, handheld).
2.8. Procedure
Upon arriving at the driving simulator facility the partic-
ipant was introduced to the simulator. The participant was
presented a written instruction explaining the driving and
phone tasks of the conversation experiment. The instruction
was to drive as s/he would do in real traffic under the same
conditions. The conversation task was explained by the test
leader and practised. The participant then entered the sim-
ulator for a practice drive where the phone task was also
practised. Then the main test drive took place. After the test
drive the participant left the simulator to fill in the question-
naire.
The dialling experiment followed immediately thereafter.
The participant received a written instruction, followed
by the test drive. After the test drive in the simulator the
participant filled in the questionnaire relating to the dialling
session.
2.9. Significance tests
The effect of phone task (conversation and dialling) is
analysed as two different experiments since the experimental
conditions are not comparable. Behavioural data is analysed
with variance analysis. Main effects of phone use (yes/no) and
phone mode (handsfree/handheld) are analysed. Where inter-
actions between use and mode appear simple main effects of
phone use are analysed. The χ2 -test is used to analyse subjec-
tive effects of phone use on driving speed and lateral position
(after dichotomization). The Fisher exact test for independent
samples is used to compare the two phone modes with respect
to the results for the subjective effects of phone use on driving
speed and lateral position. The Wilcoxon–Mann–Whitney
test for independent samples is used to analyse the results
for RSME and for subjective driving performance. Pearson
product-moment correlation coefficients between actual
 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909
and subjective effects of phone use, and the significance
levels attached to the correlation coefficients, are calculated.
A significance level of .05 is adopted for the significance
tests.
3. Results
3.1. Influence of phone use on performance on the
peripheral detection task (PDT)
The effect of phone use (difference between phone and
control situation) on PDT reaction time is shown in Fig. 1.
In the conversation experiment, the PDT reaction time was
analysed as the mean for the locations where phone calls
took place compared to the mean for the corresponding
control locations. A main effect of both types of tasks
appeared but there was no interaction between phone use and
phone mode in either of the tasks. It can be concluded that
PDT reaction time performance was impaired by both phone
tasks, irrespective of phone mode. The size of the effect was
159 ms for the conversation task and 270 ms for the dialling
task.
The effect of phone use on percentage missed PDT signals
is shown in Fig. 2. A main effect of phone use appeared
irrespective of task, but there was no interaction between
phone use and phone mode. It can be concluded that phone
use caused an increased number of missed PDT signals for
both phone modes. The size of the effect was 12.7% units
for the conversation task and 24.3% units for the dialling
task. Thus, both phone modes had similar effects on PDT
performance.
Fig. 1. Effect of conversation and dialling on PDT reaction time—difference
between phone use and control condition (±S.D.).
Fig. 2. Effect of conversation and dialling on percentage missed PDT
signals—difference between phone use and control condition (±S.D.).
905
Fig. 3. Rated effort on RSME for conversation and for dialling (±quartile
deviation) (end points: 0 = absolutely no effort, 120 = extreme effort).
3.2. Rated effort on RSME
The drivers rated their experienced effort for the whole
route on the RSME scale, irrespective of whether the phone
was used or not. The result is shown in Fig. 3. No differ-
ence between phone modes appeared for either of the tasks.
The average rating for the conversation task corresponded
roughly to “rather much effort”–“considerable effort” and
for the dialling task to “some effort” on the scale.
3.3. Correlation between PDT performance and RSME
Correlation coefficients between effort ratings and the
effect of phone use on PDT performance (reaction time,
missed signals) were calculated for the both experiments
(conversation, dialling). No correlation coefficient was sig-
nificant for either user group (handsfree, handheld).
3.4. Influence of phone use on lateral position deviation
The effect of phone use on the average standard deviation
of lateral position is shown in Fig. 4. A main effect of con-
versation appeared, but no interaction between conversation
and phone mode. The standard deviation of lateral position
decreased as an effect of conversation, irrespective of phone
mode. The size of the effect was 1.3 cm.
A main effect of dialling appeared, but no interaction
between dialling and phone mode. It can be concluded that
the standard deviation of lateral position increased as an effect
of dialling, irrespective of phone mode. The size of the effect
was 6.5 cm.
Fig. 4. Effect of conversation and dialling on standard deviation of lat-
eral position (SDLP)—difference between phone use and control condition
(±S.D.).
906 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909
Fig. 5. Reported change in lateral position deviation (LPD) as a result of
conversation and dialling.
Thus, the effect of phone use was the opposite in the
two experiments. No difference appeared, however, between
handsfree and handheld phone use.
3.5. Influence of phone use on subjective lateral position
deviation
The effect of phone use on subjective lateral position devi-
ation is shown in Fig. 5. In the conversation experiment 20 of
the handsfree drivers and 17 of the handheld drivers answered
the question in a way that can be classified either as increased
lateral position deviation or no effect (reduced deviation was
not included as an alternative). The most common answer
was “no effect”. The answers to the question did not differ
for the two phone modes. The result in the dialling experi-
ment was the same as in the conversation experiment (data
from 16 handsfree and 13 handheld drivers).
3.6. Influence of phone use on average speed
The effect of phone use on speed is shown in Fig. 6. A
main effect of conversation and an interaction between con-
versation and phone mode were obtained. Analysis of simple
main effects shows that average speed was not influenced
by conversation for handsfree mode whereas it was reduced
by conversation for handheld mode. The size of the speed
reduction was 2.8 km/h.
A main effect of dialling and an interaction between
dialling and phone mode appeared. Analysis of simple main
effects shows that average speed was reduced by dialling
Fig. 6. Effect of conversation and dialling on average speed—difference
between phone use and control condition (±S.D.).
Fig. 7. Reported speed change as a result of conversation and dialling.
for both phone modes. The effect was greater for hands-
free mode; 3.7 km/h compared to 2.0 km/h for handheld
mode.
Thus, the speed reduction was greater for handheld
mode than for handsfree mode in the conversation experi-
ment, whereas the opposite effect appeared in the dialling
experiment. It could be added that the speed levels were
quite normal for the traffic environments in the two experi-
ments.
3.7. Influence of phone use on subjective speed
The effect of phone use on subjective speed is shown
in Fig. 7. In the conversation experiment 22 of the hands-
free drivers and 18 of the handheld drivers answered
the question. The most common answer was “speed
reduction”. The answers to the question did not differ
for the two phone modes. In the dialling experiment 22
of the handsfree drivers and 19 of the handheld drivers
answered the question. The result was similar as in the
conversation experiment with no difference between phone
modes.
3.8. Correlation between actual and subjective effect of
phone use on speed
In the conversation experiment the correlation between
objective and subjective effect on driving speed is insignifi-
cant for both modes. Some individuals reported an effect of
phone conversation which was the opposite to the effect on
their driving behaviour. Seven of 30 persons who reported
reduced speed drove faster when using the phone, and three
of four persons who reported increased speed drove more
slowly when using the phone.
In the dialling experiment the result was similar. The cor-
relation between objective and subjective effects on driving
speed is also insignificant for both modes. Some individ-
uals reported an effect of dialling which was the opposite
to the real effect. Four of 26 persons who reported reduced
speed drove faster when dialling, and all four persons who
reported increased speed actually drove more slowly when
they dialled.
 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909 907

than for handsfree mode (Burns et al., 2002; Brookhuis et al.,

Fig. 8. Rated driving performance for conversation and dialling (±quartile
deviation) (end points: 0 = much worse than normal, 100 = much better than
normal).
3.9. Rated driving performance
Fig. 8 shows the result of the ratings regarding driving
performance. In the conversation experiment the handsfree
drivers reported better driving performance than the hand-
held drivers. The average rating for handsfree drivers was
somewhat closer to “as normal” than to “much worse than
normal”. The average rating for handheld drivers was some-
what closer to “much worse than normal” than to “as normal”.
In the dialling experiment there was no difference between
the two phone modes. The average rating was closer to “as
normal” than to “much worse than normal”.
4. Discussion
The result of the study regarding driving behaviour is sum-
marised in Table 1.
The peripheral detection task was used as a measure of
mental workload. An increase in mental workload may inter-
fere with driving tasks that also make demands on the driver’s
information processing abilities (Alm and Nilsson, 2001).
The handsfree and the handheld mode were rather equiv-
alent in terms of increased workload caused by phone use
according to the PDT results. This conclusion is valid both
for dialling and for conversation. Handsfree mobile phone use
is commonly supposed to be more “user friendly” since the
driver does not have to hold the phone in the hand. Conversely,
the results clearly show that the handsfree and the handheld
mode were rather equivalent in terms of increased workload,
a result that closely corresponds to the results found by Patten
et al. (2004) in real traffic. In other studies, the result for men-
tal effort ratings was somewhat mixed—higher for handheld
1991), or no difference between the two modes (Matthews et
al., 2003).
Lateral position deviation, used as the main driving perfor-
mance measure, was affected by dialling a phone number—it
increased for both phone modes. The result is interpreted as an
indication of reduced safety. The effect was of similar size for
the two phone modes, a result which deviates somewhat from
Brookhuis et al. (1991), who found a greater lateral position
deviation for the handheld mode as a result of dialling. The
result for the conversation experiment was, however, quite
the opposite. Lateral position deviation decreased to a small
extent for both modes. This finding deviates from Haigney
et al. (2000) (who showed impaired lateral control only for
handheld phone) and from Burns et al. (2002), who did not
demonstrate any effect of phone use on lane keeping per-
formance for either phone mode. A similar effect to what
was found in the present experiment, decreased lateral posi-
tion deviation, was however demonstrated for car following
in real traffic (Brookhuis et al., 1991). The interpretation,
suggested by Brookhuis et al. (1991), was that the decreased
lateral position deviation is an effect of increased driver alert-
ness. Alternative interpretations might be apt, however. One
such attempt might be couched in terms of safety margins
in lateral position when driving. A “virtual safety margin”
can be seen as an invisible line to the left and to the right of
the vehicle, and the driver wants to keep the vehicle inside
this “tunnel”, where he or she feels safe. The driver accepts a
certain lateral position deviation as long as the vehicle does
not cross the safety margins. In a situation with increased
workload (for example high speed, a complex traffic envi-
ronment or mobile phone conversation) the driver positions
the safety margins closer to the vehicle, which means that the
lateral position deviation will decrease since the driver wants
to keep the car inside the margins. It could, in other words,
be interpreted as a compensation for the increased workload.
Other explanations also seem plausible, however. It might be
related to the steering becoming less prioritised during the
phone conversation.
Driving speed, being related to crash risk and to the seri-
ousness of a crash, is also relevant from a traffic safety
perspective. The drivers reduced their speed when dialling.
The speed reduction was, however, greater for the hands-
free mode, which should be interpreted in terms of different
degrees of compensation. The reason for the results can be
supposed to lie in the fact that for dialling the number in

Table 1
The effects of phone use on driving behaviour (+: increase, −: decrease, 0: no effect)
Measure Dialling Conversation

Handsfree Handheld Handsfree Handheld

PDT reaction time + + + +
PDT missed signals + + + +
Lateral position deviation + + − −
Speed − − 0 −
908 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909
handsfree mode the drivers had to look away from the road
for a longer time period than in handheld mode. When ringing
up in handheld mode the drivers grabbed the mobile phone
from the holder (which hardly ever required the driver to
look away from the road), and then held the phone close to
the upper part of the steering wheel while dialling. From this
position the mobile phone was much closer to the frontal view
of the driver than in the handsfree mode. The driving speed
was also affected by phone conversation. The speed reducing
effect of mobile phone conversation was, however, restricted
to handheld mode (even though a similar tendency appeared
for handsfree mode). Speed reduction caused by talking on
the phone while driving a car has been demonstrated in other
simulator studies as well: Haigney et al. (2000) showed the
effect for both phone modes, whereas Burns et al. (2002) and
Patten et al. (2004) found a speed reducing effect for handheld
phones only. Since the speed reduction can be interpreted as
an attempt to compensate for the increased workload caused
by the phone conversation, conversation via a handheld phone
seems to trigger a larger compensatory effect than a handsfree
phone. The result is a bit intriguing since no difference in PDT
performance was apparent between the two phone modes.
The drivers using a handheld phone may to some extent have
made the driving task less mentally loading by slowing down.
This way, mental workload might have become rather simi-
lar to the workload in the handsfree condition. An alternative
explanation might be that drivers tend to underestimate the
risk associated with conversation when using a handsfree
phone, and therefore do not compensate by reduced speed
as much as when they use a handheld telephone. The latter is
more intuitively associated with increased risk.
A strict comparison between dialling and conversation
cannot be made, since the test situations were quite different,
and the experimental control of the presentation order was not
controlled, the dialling sessions always coming after the con-
versation session. However, the impression is that the effects
of dialling might be more negative for traffic safety, since the
lateral control was negatively affected only by dialling, not
by conversation. The effect for dialling is probably related to
the visual/motor component which was lacking in the conver-
sation task. The fact that compensation in terms of reduced
speed differed between the two experiments might be inter-
preted in a similar way—speed reduction was apparent for
both phone modes during dialling with the largest effect for
handsfree mode, whereas such compensation was apparent
only for handheld mode during conversation.
The participants reported driving performance as being
more influenced/impaired by conversing on the handheld
than on the handsfree phone, a result which hardly resem-
bles the results for actual driving performance. For dialling,
the rated driving performance did not differ between the two
modes—the correspondence between subjective and objec-
tive effects regarding driving performance was better in this
case.
The most obvious fact that ought to be transmitted to the
public is the fact that handsfree phones do not seem to be
any safer than handheld phones. The effects of conversation
seem to be rather similar for the two phone modes. When
dialling, the situation may even be the opposite one with the
handsfree phone less safe. These facts are not compatible to
what many national governments have legislated, not having
had sufficient knowledge and understanding of these issues.
It can be discussed whether the effects found in the exper-
iment are safety critical from a traffic perspective. The result
for dialling certainly points in that direction and a recom-
mendation to use voice-activated handsfree phones instead
of manually controlled phones seems to be in place. As
to conversation, the PDT results indicate that the conver-
sation was demanding in terms of mental workload, which
could be interpreted as the driver having less mental capacity
left to attend to traffic during the conversation. This could
have negative consequences in terms of reduced readiness
to respond if a risky situation were suddenly to appear. It
can be concluded with reasonable certainty that the drivers
using handheld phones tried to compensate for the increased
workload caused by the mobile phone conversation by speed
reduction. To what extent the reduced speed would compen-
sate for the reduced readiness is unclear.
When discussing safety effects of dialling versus con-
versation, the time aspect should be considered. Although
dialling may be associated with a higher risk than conversa-
tion, it probably makes up a much smaller part of the total
time of using a telephone compared to conversation. Differ-
ences observed for the conversation phase, e.g. the lack of a
speed reduction when using a handsfree phone, may have a
much larger impact on the number of accidents than differ-
ences during the dialling phase.
The present study concentrated on the effects of mobile
phone conversation and dialling on driving. Other aspects of
mobile phone use while driving still remain to be analysed
more in detail, such as distractions caused by starting or fin-
ishing a call, looking for a phone number to dial, mishaps like
dropping the phone, etc. A mobile phone with a screen show-
ing black-and-white still images was used in the study. There
are, however, newer and more advanced types of mobile
phone on the market with the capacity to transmit moving
images in colour. The risk of interference with the driving
task may well increase further with these new phones. This
issue would also require detailed study.
Acknowledgement
This project was funded by the Swedish Road Adminis-
tration (SRA).
References
Alm, H., Nilsson, L., 2001. The use of car phones and changes in driver
behaviour. Int. J. Vehicle Des. 26 (1), 4–11 (special issue).
 J.E.B. Törnros, A.K. Bolling / Accident Analysis and Prevention 37 (2005) 902–909
Brookhuis, K., de Vries, G., de Waard, D., 1991. The effects of
mobile telephoning on driving performance. Accident Anal. Prev. 23,
309–316.
Burns, P.C., Parkes, A., Burton, S., Smith, R.K., Burch, D., 2002. How
dangerous is driving with a mobile phone? Benchmarking the impair-
ment to alcohol. TRL Report TRL 547, TRL, Crowthorne, United
Kingdom.
Haigney, D.E., Taylor, R.G., Westerman, S.J., 2000. Concurrent mobile
(cellular) phone use and driving performance: task demand character-
istics and compensatory processes. Transport. Res., Part F, 113–121.
Kircher, A., Vogel, K., Törnros, J., Bolling, A., Nilsson, L., Patten,
C., Malmström, T., Ceci, R., 2004. Mobile Telephone Simulator
Study. VTI Meddelande 969A. Swedish National Road and Trans-
port Research Institute, Linköping, Sweden.
Matthews, R., Legg, S., Charlton, S., 2003. The effect of cell phone
type on drivers subjective workload during concurrent driving and
conversing. Accident Anal. Prev. 35 (4), 451–457.
Nilsson, L., 1993. Behavioural research in an advanced driving
simulator—experiences of the VTI system. In: Paper Presented at the
909
37th Annual Meeting of the Human Factors and Ergonomic Society,
Seattle, USA, October 11–15.
Patten, C., Kircher, A., Östlund, J., Nilsson, L., 2004. Using mobile tele-
phones: cognitive workload and attention resource allocation. Accident
Anal. Prev. 36, 341–350.
Rospa (Royal Society for the Prevention of Accidents), 2002. The risk
of using a mobile phone while driving. On-line paper. Available at
http://www.rospa.com/pdfs/road/mobiles/report.pdf.
Strayer, D.L., Drews, F.A., Crouch, D.J., 2004. A comparison of the
cell phone driver and the drunk driver. On-line paper. Available at
http://papers.ssrn.com/sol3/papers.cfm?abstract id=570222.
van Winsum, W., Martens, M.H., Herland, L., 1999. The effects of
speech versus tactile driver support messages on workload, driver
behaviour and user acceptance (TM-99-C043): TNO Human Factors,
The Netherlands.
Zijlstra, F., Van Dorn, L., 1985. The Construction of a Scale to Mea-
sure Perceived Effort. Delft University of Technology, The Nether-
lands.