Accident Analysis and Prevention 105 (2017) 38–43

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

Can cycling safety be improved by opening all unidirectional cycle

paths for cycle traffic in both directions? A theoretical examination of
available literature and data
Rob Methorst a,∗ , Paul Schepers a,b , Jaap Kamminga c , Theo Zeegers c , Elliot Fishman d
a
Ministry of Infrastructure and the Environment, The Netherlands
b
Utrecht University, Faculty of Geosciences, The Netherlands
c
Dutch Cyclist Union, The Netherlands
d
Institute for Sensible Transport, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Many studies have found bicycle-motor vehicle crashes to be more likely on bidirectional cycle paths
Received 8 December 2015 than on unidirectional cycle paths because drivers do not expect cyclists riding at the right side of the
Received in revised form 13 May 2016 road. In this paper we discuss the hypothesis that opening all unidirectional cycle paths for cycle traffic in
Accepted 18 May 2016
both directions prevent this lack of expectancy and accordingly improves cycling safety. A new national
Available online 1 June 2016
standard requires careful consideration because a reversal is difficult once cyclists are used to their
new freedom of route choice. We therefore explored the hypothesis using available data, research, and
Keywords:
theories. The results show that of the length of cycle paths along distributor roads in the Netherlands,
Cycling
Cycling safety 72% is bidirectional. If drivers would become used to cyclists riding at the left side of the road, this result
Bidirectional cycle paths raises the question of why bidirectional cycle paths in the Netherlands still have a poor safety record
Contra-flow cycling compared to unidirectional cycle paths. Moreover, our exploration suggested that bidirectional cycle
paths have additional safety problems. It increases the complexity of unsignalized intersections because
drivers have to scan more directions in a short period of time. Moreover, there are some indications that
the likelihood of frontal crashes between cyclists increases. We reject the hypothesis that opening all
unidirectional cycle paths for cycle traffic in both directions will improve cycle safety. We recommend
more attention for mitigating measures given the widespread application of bidirectional cycle paths in
the Netherlands.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction Bicyclists travelling against the direction of traffic on unidirec-

Over recent decades, the Netherlands has established a repu-
tation as one of the safest countries in which to cycle, on a per
distance travelled basis (Schepers et al., 2015). Although unrivalled
in both cycling participation and safety, bicycle use on bidirectional,
segregated paths has been shown to carry an elevated risk when
compared to cycling on unidirectional paths. This paper sets out to
determine whether the safety of cycling in the Netherlands would
be improved by adopting a policy of making all cycle paths bidirec-
tional. Thereby, this study focuses on a system level while previous
research, described in the following, was done at a location level
such as a crossing.
∗ Corresponding author.
E-mail address: rob.methorst@telfort.nl (R. Methorst).
tional cycle paths are 4–6 times as likely to sustain bicycle-motor
vehicle crashes at intersections as compared to those travelling
with the direction of traffic (Schepers and Voorham, 2010; Wachtel
and Lewiston, 1994). Even though legal and indicated to drivers by
traffic signs, the likelihood of bicycle-motor vehicle crashes at inter-
sections with bidirectional cycle paths is found to be higher than
at intersections with unidirectional paths (Schepers et al., 2011;
Vandenbulcke et al., 2014) (see Fig. 1). These elevated risks have
been explained by drivers’ expectations and related visual scan-
ning strategies at intersections. Drivers entering a distributor road
from a minor road have difficulties in detecting cyclists from the
right (in case of right-hand driving) (Räsänen and Summala, 1998),
see the sketch in Fig. 2. Summala et al. (Summala et al., 1996) found
that drivers turning right scanned the right leg of the T-intersection
less frequently and later than those turning left. Their explanation
is that drivers turning right focus their attention on cars from the
left because those coming from the right pose no threat to them.

http://dx.doi.org/10.1016/j.aap.2016.05.018
0001-4575/© 2016 Elsevier Ltd. All rights reserved.
 R. Methorst et al. / Accident Analysis and Prevention 105 (2017) 38–43
Fig. 1. Unsignalized intersection with bidirectional cycle paths at both sides of the
distributor road.
Fig. 2. Sketch of an encounter between a driver entering a distributor road from a
minor road and a cyclist coming from the right (riding on the left side of the road).
Van Haeften (2010) found similar results in Groningen which has a
bicycle modal share as high as 50%, one of the highest in the world
(Ligtermoet, 2009).
Because of the elevated risks at intersections, the Dutch Design
Manual for Bicycle Traffic (CROW, 2007) cautions against applica-
tion of bidirectional bicycle paths unless it strongly reduces cyclists’
need to cross distributor roads and avoid potential large shares of
illegal contraflow cycling. Most bicycle-motor vehicle crashes in
urban areas in the Netherlands occur at unsignalised intersections
of distributor roads (Schepers and Voorham, 2010). Per passing
cyclist, the number of bicycle-motor vehicle crashes is about two
times as high while crossing a distributor road compared to while
crossing a minor road at an unsignalised intersection. However,
since the number of minor roads is high, there are about two times
as many bicycle-motor vehicle crashes while crossing a minor road
as compared to while crossing a distributor road (Schepers et al.,
2011; Schepers and Voorham, 2010). Because of the high frequency
of the former crash type and its increased likelihood at bidirectional
cycle paths, a net safety improvement through bidirectional cycle
paths is expected to be rare (CROW, 2007).
The lack of expectancy of cyclists from the right for cars entering
from a minor road seems to be the main problem of bidirectional
cycle paths (Summala et al., 1996). It has been suggested that this
expectancy problem is related to a lack of uniformity. Some paths
are unidirectional and others are bidirectional. It can therefore be
hypothesized that cycling safety can be improved by allowing con-
traflow cycling on all unidirectional cycle paths. A few Dutch cities
like Breda already decided to convert unidirectional cycle paths into
bidirectional bicycle paths (Gemeente Breda, 2007).
Uniform allowance of contraflow cycling on all unidirec-
tional cycle paths in a number sufficiently large jurisdictions (i.e.
provinces in the Netherlands) with other jurisdictions serving as
39
control areas would be ideal from a research perspective to inform
a before-after study. However, such research and or new national
standard containing this new policy requires careful consideration
because a reversal is difficult once cyclists are used to their new
freedom of route choice. This paper explores the hypothesis that
cycling safety can be improved by converting all unidirectional
cycle paths into bidirectional cycle paths in the Netherlands using
available data, research, and theories. We did so by considering the
development of illegal contraflow cycling on unidirectional cycle
paths (Section 2), the presence of bidirectional and unidirectional
cycle paths (Section 3), a discussion of alternative theories for the
aforementioned expectancy theory to judge the safety of bidirec-
tional cycle paths (Section 4), and consideration of other crash
types that may be affected next to the aforementioned bicycle-
motor vehicle crashes at intersections (Section 5). The results are
summarised and discussed in Section 6.
2. Is contraflow cycling increasing on unidirectional cycle
paths?
Currently available data suggest that contraflow cycling on uni-
directional paths is substantial but constant. This is important
because increased contraflow cycling could potentially cause safety
levels at unidirectional cycle paths to decrease to a level compara-
ble to bidirectional paths because of the high risk of cyclists riding
against the direction of traffic on unidirectional paths (Schepers and
Voorham, 2010; Wachtel and Lewiston, 1994). The earliest study
we found about contra-flow cycling was conducted in 1994 by Van
Minnen and Braimaister (1994). Of the 1314 cyclists counted at
unidirectional cycle paths around roundabouts, 4% were cycling
against the direction of traffic. A count of 23,652 cyclists completed
in 2014 yielded 2% (Methorst and Schepers, 2015). Three locations
in the study by Methorst and Schepers (2015) were also examined
using conflict observations in 2012 (De Goede et al., 2013). The
share at these locations remained stable at 3%.
The share of contraflow varied between locations and over
time, e.g. between 2% and 13% in the study by Van Minnen and
Braimaister (1994) and between 0% and 9% at locations reported
by Methorst and Schepers (2015). The lowest share was found at
the location with the highest volume of cyclists (over 2000 cyclists
per hour during the countings). Similarly, Methorst and Schepers
(2015) found a share below 2% on weekdays and above 3% in the
weekend (with countings at all locations taken together). High
volumes of cyclists appear to correlate with reduced contra-flow
cycling, probably due to less space remaining available.
3. To what degree are bidirectional cycle paths applied in
the Netherlands?
The Dutch Cyclists’ Union recently finalized an online route
planner covering the whole country (Fietsersbond, 2014). Whether
cycle paths are open for cycle traffic in one or two directions is
essential information for route planning and is therefore recorded
in the underlying database. Based on this database, Zeegers and
Kamminga (2014) estimated that the Netherlands:
• Has some 15,900 km of bicycle paths along roads (standalone
bicycle paths are excluded) of which 11,400 km (72%) are bidi-
rectional paths
• The share of bidirectional bicycle paths amounts to 62% within
city limits and 79% outside city limits
• Road sections (between intersections) of unidirectional cycle
paths within city limits have an average length of 52 m versus
107 m for bidirectional cycle paths
40 R. Methorst et al. / Accident Analysis and Prevention 105 (2017) 38–43
For a sample of 250 bidirectional cycle paths within city lim-
its, Zeegers and Kamminga (2014) judged visually whether there
was a route alternative for cyclists at the other side of the road. In
about 65% of the cases there was no alternative cycle route at the
other side of the road (Zeegers and Kamminga, 2014). Moreover, for
a sample of 20 municipalities, the Cyclist Union has recorded the
width of cycle paths. Bidirectional cycle paths appear to be wider
than one-way bicycle paths, but the difference is less than a factor
of two (Zeegers and Kamminga, 2014).
There appeared to be large variations in the share of bidirec-
tional cycle paths among municipalities with some nearing 100%.
The highest shares were found in new towns like Lelystad, Almere,
Houten, and Zoetermeer, where only a small share of the length
of distributor roads have cycle facilities (Zeegers and Kamminga,
2014). In these cities cyclists are mostly separated from the coarse
distributor road network and instead have a fine grained network
of standalone bicycle paths and access roads through traffic calmed
areas at their disposal (Schepers et al., 2013). To fit logically within
a larger network of standalone cycle paths, which are bidirectional
by standard, contraflow cycling is allowed on the few cycle paths
along distributor roads.
The aforementioned finding that road sections between inter-
sections with minor roads are about two times as long for
bidirectional paths as compared to unidirectional paths suggests
that road authorities indeed account for the amount of crossings as
advised by CROW (2007). The lack of a route alternative at the other
side of the road for the majority of bidirectional cycle paths suggests
that cost savings play a role as well. To conclude, road authorities
seem to account for the recommendations in the Design Manual
for Bicycle Traffic on the application of bidirectional cycle paths
but their application in real life is much more frequent than would
be expected given CROW’s advice (CROW, 2007).
4. Human factors theories and cyclist safety
Human factors theories or ergonomics theories help explain
how roads can be designed to fit road users’ needs and capabili-
ties (Birth et al., 2009; Schepers et al., 2014). Expectancy theory
was already discussed in the Introduction in relation to drivers’
scanning strategies at intersections. Drivers insufficiently look for
cyclists from the right when entering from a minor road (Summala
et al., 1996; Van Haeften, 2010). Similar problems have been found
at roundabouts (Sakshaug et al., 2010). The safety in numbers the-
ory posits that cyclist safety is improved when cycle use is more
prevalent and this is explained by drivers adapting their expec-
tations and behaviour in response to the increased numbers of
cyclists (Jacobsen, 2003). Similarly, it can be hypothesized that
drivers learn to expect cyclists from the right once they meet suf-
ficiently large numbers of cyclists from this direction. Moreover,
a uniform application of bidirectional cycle paths would enable
a national campaign to inform drivers. We do not expect large
benefits from the latter since visual scanning strategies is largely
habitual and automatized by driving experience (Summala, 1988)
and therefore difficult to influence by communication.
Next to expectations, the elevated risk of bicycle-motor vehicle
crashes at intersections has also been related to vision. Contrary
to what one might expect, Henson and Whelan (1992) found good
visibility at T-junctions to be associated with a greater probability
of bicycle crashes when a cyclist was riding among cars. However,
when cyclists are entering from the right on a bidirectional cycle
path (from the unexpected direction for drivers), sight obstacles at
the right side appear to increase the risk of bicycle-motor vehicle
crashes compared to a situation with better visibility (Räsäsen et al.,
1998; Schepers and Voorham, 2010). Fig. 3 exemplifies such a sight
obstruction.
Fig. 3. A bidirectional cycle path with a sight obstruction at the right side of the
road.
Fig. 1 presents an example with only a 1 m clearance resulting
in a demanding task for drivers entering or leaving the distributor
road. As people have limited information processing capacity (De
Waard, 1996; De Waard et al., 2010), road users are more likely to
be overloaded if road design is complex (Birth et al., 2009; Elvik,
2006). A driver entering or leaving a distributor road has to scan
a lot of directions if he or she has to cross a bidirectional cycle
path. Some distance between the cycle path and the main carriage-
way (sometimes called ‘bent-out crossing’) gives drivers extra time
because they can first scan for cycle traffic on the bicycle path and
after that for motor traffic on the road (Elvik et al., 2009; Schepers
et al., 2011).
Unfortunately, a clearance between a bicycle path and main
carriageway of a least a few meters is not always possible and
similarly the possibilities to remove sight obstacles can be lim-
ited. Even when a lack of expectations among drivers would be
undone by a uniform application of bidirectional cycle paths, high
work load (in terms of directions to scan for drivers) in combina-
tion with restricted visibility (due to sign obstacles) will remain in
a number of instances. The hypothesis that a uniform application
of bidirectional cycle paths would improve cycling safety cannot be
underpinned by human factors theory.
5. Non-intersection (mid-block) crashes
All previous sections were concerned with bicycle-motor vehi-
cle crashes at intersections. There is to our best knowledge no
quantitative research linking the presence of bidirectional cycle
paths to collisions between bicycles, even though cyclists sustain
substantial numbers of serious injuries in such crashes (Schepers,
2010). Similarly, the potential for frontal conflicts with mopeds is
increased where they have to use the bicycle path, which, contrary
to the Dutch National policy not to allow moped on bicycle paths
within city limits, is the case on 42% of the length of bidirectional
cycle paths within city limits in the Netherlands (Zeegers and Kam-
minga, 2014). Furthermore, so the called ‘snorfiets’ (max 25 km/h
‘mopeds’, slow mopeds), comprising over three-quarters of all sold
mopeds in 2012 (Ewalds et al., 2013), ride as fast as ‘real’ mopeds
on cycle paths before the National policy was enacted in 1999, i.e.
some 34 km/h (Wijlhuizen et al., 2013). Moreover, slow mopeds
have the same mass and width, are allowed on the bicycle paths
and form potential risk of serious frontal conflicts (Methorst et al.,
2011). A small-scale study with 17 frontal crashes between cyclists
found a high frequency on bidirectional cycle paths (Schepers,
2010). Similar results were found in an in-depth study of crashes
with cyclists older than 50 years which led SWOV to advice bidi-
rectional cycle paths be sufficiently wide (in accordance with the
Design manual which also accounts for the volume of cycle traf-
fic; CROW, 2007) or even have a physical separation between both
directions of cycle traffic (Davidse et al., 2014), although we think
a separator might introduce single bicycle crashes. The numbers
in these studies are too small to draw firm conclusions but they
 R. Methorst et al. / Accident Analysis and Prevention 105 (2017) 38–43
suggest that, without additional measures, converting one-way
into two-way cycle paths is likely to increase the number of col-
lisions between cyclists and with mopeds and/or slow mopeds.
Additionally, it should be noted that bidirectional cycle paths have
to be wider to allow safe passing and overtaking, especially if
mopeds are allowed (CROW, 2007). Road authorities sometimes
lack space to provide the required width. Therefore, a uniform
application of bidirectional paths will result in an increased length
of insufficiently wide paths and accordingly increased likelihood of
collisions between cyclists and falls resulting from manoeuvres to
avoid a crash (Davidse et al., 2014).
6. Discussion: bringing different lines of research together
We hypothesized that opening all unidirectional cycle paths
for cycle traffic in both directions would improve cycling safety.
Based on the outcomes of this study we reject the hypothesis. The
hypothesis is based on the belief that drivers will adapt their scan-
ning strategies once they meet large numbers of cyclists from the
right when entering a distributor road from a minor road. How-
ever, Dutch motorists encounter cyclists at a higher rate than in
any other country, given that some 27% of all trips are by bike
(Statistics Netherlands, 2015). Additionally, some 72% of Dutch
bicycle paths are already bidirectional, meaning that motorists
could be expected to be accustomed to scanning strategies that
include cyclists coming from the contra-direction. Still, recent
research found an elevated risk of bicycle-motor vehicle crashes
at unsignalised intersections in the Netherlands (Schepers et al.,
2011).
We do not expect the safety levels at unidirectional cycle paths
to decrease to a level comparable to bidirectional paths because
of the high risk of cyclists riding against the direction of traffic
on unidirectional paths (Schepers and Voorham, 2010; Wachtel
and Lewiston, 1994). The share of contraflow cycling amounted to
between 2% and 4% of all cyclists at unidirectional cycle paths, but
there are no indications of an increase overtime.
Even without expectation problems, sufficient space is needed
to safely construct bidirectional cycle paths. A clearance between
a bidirectional bicycle path and distributor road of a least a few
meters is needed to reduce a too complex design where drivers have
to scan too many directions in a short moment of time. This problem
is aggravated at some locations by sight obstructions. Also, bidirec-
tional cycle paths need to be sufficiently wide to prevent frontal
crashes between cyclists and with mopeds where they are allowed.
Road authorities often lack space to meet these requirements. Con-
verting all unidirectional cycle paths into bidirectional bicycle paths
will cause new risks at locations with insufficient space.
6.1. International transferability of the results
This paper is to a large extent based on the Netherlands where
drivers, due to the high model share of cycling and widespread
application of bidirectional cycle paths, have ample opportuni-
ties to grow accustomed to cyclists riding along the right side
of the road. The fact that Dutch drivers coming from a side road
still insufficiently scan for these cyclists (Van Haeften, 2010) sug-
gests that more widespread application will not solve expectancy
problems in countries with lower volumes of cyclists and accord-
ingly fewer encounters with cyclists from the right. The potential
human factors issues of complexity and sight obstructions are not
restricted to cycling countries either. For instance, an additional
risk increase due to sight obstacles at the right side of a minor
road was first discovered in Helsinki (Räsäsen et al., 1998). The
risk of frontal crashes between cyclists and with (slow) mopeds
may be less elevated at bidirectional cycle paths in countries with
lower volumes of cyclists and accordingly where there are less
41
encounters with oncoming cyclists. On the other hand, in coun-
tries with lower volumes of cyclists, paths may be narrow (in the
Netherlands a minimum of 2.5 m or more for high volumes of
cyclists is advised by CROW, 2007), of poor quality and shared with
pedestrians thereby introducing new risks. For instance, Van der
Voet-Kurbatsch (2014) describes that the German Cyclist’s Associ-
ation ADFC has argued against the obligation to cycle on cycle paths
(the so-called Benutzungsplicht) because these are often insuffi-
ciently wide, of poor quality or shared with pedestrians, etc. This
position is to a large extent adopted in design guidelines and leg-
islation. There is a risk of head-on crashes with mopeds or slow
mopeds only in countries where these two-wheelers are frequently
used and where these vehicles are allowed on bicycle paths. As this
is only one of the many risks associated with bidirectional cycle
paths, there is little reason to believe that there are jurisdictions
where cycling safety can be improved by opening all unidirectional
cycle paths for cycle traffic in both directions.
6.2. Limitations
Unfortunately we could not empirically test our hypothesis
which limits our possibilities to draw firm conclusions. Technically
it is possible to simulate a city having only bidirectional cycle paths
in a driving simulator. Participants could drive multiple sessions to
monitor whether their visual scanning strategies changes. How-
ever, visual search pattern are highly automatized based on many
hours or even years of driving experience which explains the differ-
ences between experienced and novice drivers (Patten and Kircher,
2006). Practically, it is almost impossible to achieve sufficient hours
of simulator driving to resemble real world learning processes. A
higher level of external validity can be achieved by studying visual
search patterns among drivers living a city with a share of bidi-
rectional cycle paths nearing 100%. Zeegers and Kamminga (2014)
found the share of bidirectional cycle paths in some Dutch cities to
approach 100%, but these cities were primarily new towns. These
cities have only a limited length of distributor roads with bicycle
infrastructure which could prevent drivers from developing the
right expectations. A few older cities like Breda decided to make
bidirectional cycle paths their standard (Gemeente Breda, 2007)
so they may also achieve a share near 100% in the future. It is not
yet clear whether this would allow drivers to develop the right
expectations. Along with converting unidirectional cycle paths into
bidirectional cycle paths, Breda equips intersections with new traf-
fic lights (Methorst and Schepers, 2015). As visual search patterns at
signalized intersections are to a large degree based on traffic lights,
it is unlikely that it will affect drivers’ expectations at the decreased
number of unsignalised intersections and roundabouts.
6.3. Recommendations for practitioners
The analysis presented in this paper found an insufficient basis
for changing the Dutch Design Manual for Bicycle Traffic to con-
vert all cycle paths to bidirectional (CROW, 2007). However, as
more than half of the cycle path length is open in two directions
with a strongly increased risk of bicycle-motor vehicle crashes at
unsignalized intersections and roundabouts (Schepers et al., 2011;
Zeegers and Kamminga, 2014), we advise to expand the discus-
sion of appropriate mitigation measures and present a start in the
remainder of this section. For instance, a recent publication by the
Dutch Fietsberaad (Bicycle Council), advises speed reducing mea-
sures and removal of sight obstructions (Van Boggelen et al., 2011).
Very few studies specifically focussed on crossings with bidirec-
tional cycle paths. An exception is a before-after study by Summala
et al. (1996). Positive results were found for a stop sign and speed
reducing measures, but the researchers did not find visual search
behaviour to be affected by markings (Summala et al., 1996). The
42 R. Methorst et al. / Accident Analysis and Prevention 105 (2017) 38–43
Fig. 4. Traffic sign with flashing lights to warn drivers for approaching cyclists.
University of Groningen tested the traffic sign depicted in Fig. 4
which has flashing lights to warn for approaching cyclists. Visual
search behaviour improved in that drivers from a minor road more
often scanned for cyclists from the right, but it was only a small
scale test with indicative results (Van Haeften, 2010). In addition
to LED-lights on the traffic sign, there are other solutions to add
LED-lights on the surface along the crossing. Both kinds of flashing
lights are turned on after an approaching pedestrian or cyclist is
detected (Sysconnect, 2016).
7. Conclusion
Bicycle-motor vehicle crashes are more likely on bidirectional
cycle paths than on unidirectional cycle paths. In this paper we dis-
cuss the hypothesis that opening all unidirectional cycle paths for
cycle traffic in both directions would solve this problem because
drivers get used to cyclists riding at the left side of the road. Based
on an exploration of available research, data and theories, we reject
the hypothesis. Bidirectional cycle paths are far more risky than
unidirectional cycle paths. A uniform application of bidirectional
cycle paths would probably cause new risks due to sight obstruc-
tions at intersections and a lack of space to construct sufficiently
wide paths.
Acknowledgement
This work was supported by Dutch Ministry of Infrastructure
and the Environment.
References
Birth, S., Pflaumbaum, M., Potzel, D., Sieber, G., 2009. Human Factors Guideline for
Safer Road Infrastructure. World Road Association, Paris.
CROW, 2007. Design Manual for Bicycle Traffic. CROW, Ede.
Davidse, R., Van Duijvenvoorde, K., Boele, M.J., Doumen, M.J.A., Duivenvoorden,
C.W.A.E., Louwerse, W.J.R., 2014. Letselongevallen van Fietsende 50-Plussers
(Injury Crashes of Cyclists Older than 50 years). SWOV Institute for Road Safety
Research, The Hague.
De Goede, M., Obdeijn, C., Van der Horst, A.R.A., 2013. Conflicten Op
Fietspaden—Fase 2 (Conflicts at Bicycle Paths—Phase 2). TNO, Soesterberg.
De Waard, D., Schepers, J.P., Ormel, W., Brookhuis, K.A., 2010. Mobile phone use
while cycling: incidence and effects on behaviour and safety. Ergonomics 53,
30–42.
De Waard, D., 1996. The Measurement of Drivers’ Mental Workload. Groningen
University, Groningen.
Elvik, R., Høye, A., Vaa, T., Sørensen, M., 2009. The Handbook of Road Safety
Measures. Emerald, Bingley.
Elvik, R., 2006. Laws of accident causation. Accid. Anal. Prev. 38, 742–747.
Ewalds, D., Moritz, G., Sijstermans, 2013. Bromfietsen in Nederland (Mopeds in the
Netherlands). Statistics Netherlands, The Hague.
Fietsersbond, 2014. Fietsrouteplanner (Bicycle Route Planner). Fietsersbond,
Utrecht.
Gemeente Breda, 2007. Fietsen, net zo gemakkelijk, Breda.
Henson, R., Whelan, N., 1992. Layout and design factors affecting cycle safety at
T-junctions. Traffic Eng. Control 33, 548–551.
Jacobsen, P.L., 2003. Safety in numbers: more walkers and bicyclists, safer walking
and bicycling. Inj. Prev. 9, 205–209.
Ligtermoet, D., 2009. Bicycle Policies of the European Principals. Fietsberaad,
Utrecht.
Methorst, R., Schepers, J.P., 2015. Tweerichtingsfietspaden en Spookrijden
(Bidirectional Cycle Paths and Riding Against Traffic at Unidirectional Cycle
Paths). Delft, Rijkswaterstaat.
Methorst, R., Schepers, J.P., Vermeulen, W., 2011. Snorfiets op het Fietspad (Slow
Moped on the Bicycle Path). Delft, Rijkswaterstaat.
Patten, C.J.D., Kircher, A., Östlund, J., Nilsson, L., Svensson, O., 2006. Driver
experience and cognitive workload in different traffic environments. Accid.
Anal. Prev. 38, 887–894.
Räsänen, M., Summala, H., 1998. Attention and expectation problems in
bicycle–car collisions: an in-depth study. Accid. Anal. Prev. 30, 657–666.
Räsäsen, M., Summala, H., Pasanen, E., 1998. The safety effect of sight obstacles and
road-markings at bicycle crossings. Traffic Eng. Control 1998 (February),
98–102.
Sakshaug, L., Laureshyn, A., Svensson, Å., Hydén, C., 2010. Cyclists in
roundabouts—different design solutions. Accid. Anal. Prev. 42, 1338–1351.
Schepers, J.P., Voorham, J., 2010. Oversteekongevallen met Fietsers (Crossing
Accidents with Cyclists). Delft, Rijkswaterstaat.
Schepers, J.P., Kroeze, P.A., Sweers, W., Wust, J.C., 2011. Road factors and
bicycle-motor vehicle crashes at unsignalized priority intersections. Accid.
Anal. Prev. 43, 853–861.
Schepers, J.P., Heinen, E., Methorst, R., Wegman, F.C.M., 2013. Road safety and
bicycle usage impacts of unbundling vehicular and cycle traffic in Dutch urban
networks. Eur. J. Transp. Infrastruct. Res. 13, 221–238.
Schepers, J.P., Hagenzieker, M.P., Methorst, R., Van Wee, G.P., Wegman, F., 2014. A
conceptual framework for road safety and mobility applied to cycling safety.
Accid. Anal. Prev. 62, 331–340.
Schepers, P., Twisk, D., Fishman, E., Fyhri, A., Jensen, A., 2015. The Dutch road to a
high level of cycling safety. Saf. Sci., in press.
Schepers, J.P., 2010. Fiets-Fietsongevallen (Bicycle-Bicycle crashes). Delft,
Rijkswaterstaat.
Statistics Netherlands, 2015. Statline.
Summala, H., Pasanen, E., Räsänen, M., Sievänen, J., 1996. Bicycle accidents and
drivers’ visual search at left and right turns. Accid. Anal. Prev. 28, 147–153.
Summala, H., 1988. Risk control is not risk adjustment: the zero-risk theory of
driver behaviour and its implications. Ergonomics 31, 491–506.
Sysconnect, 2016. Interactieve oversteekplaats verbetert de veiligheid (Interactive
Crossing Improves Safety). Downloaded 6.3.2016 from http://www.
sysconnect.nl/interactieve-dynamische-oversteekplaats-zebrapad.html.
Van Boggelen, O., Schepers, J.P., Kroeze, P.A., Van der Voet, M., 2011.
Fietsberaadpublicatie 19: Samen Werken aan een Veilige Fietsomgeving
(Collaborating to Improve Cycling Safety). Fietsberaad, Utrecht.
Van Haeften, M., 2010. Het Kijkgedrag van Automobilisten en Fietsers bij
Kruispunten met een Tweerichtingsfietspad (Viewing Behaviour of Drivers and
Cyclists at Intersections with a Two-Way Cycle Track). University of
Groningen, Groningen.
Van Minnen, J., Braimaister, L., 1994. De Voorrangsregeling voor Fietsers op
Rotondes met Fietspaden (The Priority Rule for Cyclists at Roundabouts with
Cycle Paths). SWOV Institute for Road Safety Research, Leidschendam.
Van der Voet-Kurbatsch, M., 2014. Ervaringen met Speed-pedelcs in Duitsland
(Experiences with Speed-Pedecs in Germany). In: Schepers, P., Van der
Voet-Kurbatsch, M. Effecten van mogelijke gedragsregels voor speed-pedelecs.
Rijkswaterstaat.
Vandenbulcke, G., Thomas, I., Panis, L.I., 2014. Predicting cycling accident risk in
Brussels: a spatial case–control approach. Accid. Anal. Prev. 62, 341–357.
Wachtel, A., Lewiston, D., 1994. Risk factors for bicycle-motor vehicle collisions at
intersections. ITE J. 64, 30–35.
 R. Methorst et al. / Accident Analysis and Prevention 105 (2017) 38–43 43

Wijlhuizen, G.J., Dijkstra, A., Bos, N.M., Goldenbeld, C., Stipdonk, H.L., 2013.
Educated Guess About the Consequences for Casualties as a Result of
Introduction of the Measure Light Moped in the Carriageway (SOR) in
Amsterdam; A First Estimate of the Effects Related to Road Safety. SWOV
Institute for Road Safety Research, The Hague.
Zeegers, T., Kamminga, J., 2014. Een Inventarisatie van het Voorkomen van
Tweerichtingsfietspaden in Nederland (An Inventory of Bidirectional Cycle
Paths in the Netherlands). Fietsersbond, Utrecht.