Edited by: Sergio A. Useche, University of Valencia, Spain
Reviewed by: Mireia Faus, University of Valencia, Spain; Anne Lusk, Harvard T. H. Chan School of Public Health, United States
This article was submitted to Environmental Psychology, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The Norwegian authorities want to limit the extent of car use in city areas to existing levels. Such a limitation would help combat climate change, improve health of citizens, and alleviate congestion. This implies that any further increase in transport needs will have to be met by walking, cycling and use of public transport. Reaching this ambitious goal requires knowledge about cyclists' preferences concerning operation and maintenance (M&O) of roads and foot/cycle paths. Previous research suggests that M&O have great implications for travel mode choice, bicycle route/path choice, safety, security, and comfort. With the need to serve bicyclists of all ages and genders, this study additionally explores which M&O of roads and foot/cycle the different demographic groups perceive positively or negatively. This article reports results from a nationwide survey in the summer of 2019. Two thousand three hundred seventy-six cyclists across Norway (55% male; 29% <40; 17% >60) participated to determine the cyclists' perceptions about year-round M&O of roads and foot/cycle paths. Respondents, rather than being randomly selected, completed an internet-linked survey. The variables included maintenance of foot/cycle paths in terms of salt and snow plowing and operation and maintenance of roads in terms of glass, holes/bumps, and conditions. Our results suggest that female cyclists suffer more from adverse conditions than do males. We also find that males are more likely to cycle during winter, which is an additional indication that adverse conditions affect women and men differently. Surprisingly, older cyclists report to be less affected by poor conditions than younger cyclists. Self-selection to participate in the survey among older cyclists might be an important explanation for this result. Cycling conditions vary greatly between geographical areas, reflecting the large climatic variations across Norway. Most respondents have experienced a cycle accident where conditions contributed, and many sometimes forfeit cycling due to adverse conditions. Implications for future research and practice of M&O are discussed.
The Norwegian authorities have stated that they want to limit car use in city areas to existing levels (“nullvekstmålet”). They are in line with a global movement that seek to open city centers up to pedestrians and bicyclists (Wolfe,
There are basically two types of designated cycle infrastructure implemented in Norway: separated paths for bicyclists and pedestrians, and bicycle lanes in the ordinary streets. Bicycle lanes, normally with red colored asphalt or paint, are most common in the city centers, whereas bicycle (and pedestrian) paths dominate outside cities. To mix pedestrians and cyclists on the same infrastructure is common, and it is also allowed for cyclists to cycle on foot paths, sidewalks and in pedestrian streets.
Bicycle infrastructure is normally built according to a uniform standard defined by the Norwegian Public Roads Administration, but local adjustments are often applied. Still, in all the studied areas, the bicycle infrastructure is more or less similar, consisting of bicycle lanes in inner city centers and bicycle paths shared with pedestrians outside city centers. New bicycle and pedestrian paths are normally built with pavement for pedestrians, separating them from bicyclists.
In most urban areas in Norway, and in particular in Oslo, there has been a steady and huge increase in the use of bicycle lanes during later years. The Oslo city authorities have ambitious plans to build a city-wide bicycle network and 52 km of new bicycle infrastructure were built from 2016 to 2020. Today Oslo, with a population of 700 000 inhabitants, has a total bicycle network of ca. 250 km (Oslo City Counsil,
Infrastructure design could be important to achieve a transport modality shift and increase bicycle ridership. Literature reviews have suggested that building accommodations for cyclists and extending the network of cycle paths and lanes increase cycling (Buehler and Dill,
Increasing the share of transport done by cycling could increase the number of traffic injuries (Bjørnskau,
Demographic factors have been suggested as determining the influence of M&O on cyclists (Svorstøl et al.,
Norwegian roads and infrastructure face seasonally dependent challenges, and winter-related conditions impact elderly and physically disabled individuals more severely. The consequences of suboptimal M&O are far reaching, and influence route choice, choice of transportation mode, whether to travel, and both perceived and actual safety. Thus, proper and improved M&O could greatly boost security and accessibility of cycling and help reach the ambitious goals of increasing the share of cyclists to 20% in Norwegian city areas. Research has suggested that improved winter maintenance could increase bicycle use by 18% and reduce reliance on cars (Bergström,
While infrastructure is necessary for a safe and pleasant trip by bike, proper and prioritized M&O is also a prerequisite for making cycling more attractive and available to more people (Garrard,
A recent survey investigated cyclists' perceptions of M&O across Norway (Johansson and Bjørnskau,
Previous research has suggested that M&O is important to increase the cycling shares and to do so safely. It has been suggested that adverse conditions affect different demographic differently. Additionally, Norway is a diverse geographical country with distinct issues in summer and winter. In this paper, we will investigate demographic differences regarding M&O as well as considering geographical effects. In light of previous research and experiences, we have formulated the following hypotheses:
1) Female cyclists will be more affected by poor cycling conditions than men
2) Older participants will be more affected by poor cycling conditions than younger participants
3) Maintenance and operation affect whether one chooses to cycle
4) Conditions both in summer and winter impact different groups differently
We will also investigate other factors regarding demographic and geographic differences in how people experience M&O.
This paper presents results from a survey with a cross-sectional design administered during June, July and the start of August 2019. Recruitment for this study was done in multiple ways. We wanted a broad geographical representation of Norwegian cyclists. Six different recruitment strategies were used. (1) We sent out invitational emails to people who had participated in previous surveys and agreed to be contacted for future surveys. (2) The Norwegian Cyclists' Association (SLF) included a link to the survey in their newsletter. (3) We had a paid advertisement on Facebook. (4) Employees in the city of Oslo also received an invitational email. (5) Several bicycle repair shops agreed to have posters about the study put up. (6) Project partners in the Norwegian Public Roads Administration shared the link to the survey. Additionally, some people reported being recruited through other channels, such as word of mouth. At every instance, it was made clear that we wanted participants who cycled.
We ended up with a total sample of 2,556 cyclists who represented the largest urban areas across Norway. Due to the many channels of recruitment and low control of the spread of the invitation, it is impossible to calculate any response percentage (see
Independent variables in the analyses; all except “gender” and “often cycled” are categorical.
100 | 2,376 | ||
Other | 14 | 338 | 2,987,606 |
Akershus | 7 | 176 | 624,055 |
Bergen | 8 | 189 | 283,929 |
Buskerud | 6 | 137 | 283,148 |
Jæren | 13 | 312 | 201,801 |
Oslo | 16 | 386 | 681,067 |
Trondheim | 19 | 450 | 198,219 |
Tromsø | 16 | 388 | 76,975 |
100 | 2,376 | ||
Male | 55 | 1,308 | |
Female | 45 | 1,068 | |
100 | 2,376 | ||
<40 | 29 | 699 | |
40–49 | 29 | 687 | |
50–59 | 24 | 578 | |
>60 | 17 | 411 | |
100 | 2,376 | ||
Other | 16 | 369 | |
Hybrid | 28 | 665 | |
MTB | 14 | 341 | |
Racer | 15 | 364 | |
E-bike | 27 | 637 | |
100 | 2,376 | ||
<4 days per week | 38 | 908 | |
More than 4 days per week | 62 | 1,468 |
Participants were informed at multiple points that the questionnaire was going to survey them about their thoughts and actions as cyclists regarding M&O. All included variables are presented in
Dependent variables included in multiple different analyses.
Accident where road conditions contributed | 100 | 711 |
At least one = 1 | 74 | 525 |
Authorities use too much salt on foot and cycle paths | 100 | 1,418 |
Agree = 1 | 44 | 620 |
How satisfied are you with snow plowing on foot and cycle paths | 100 | 1,586 |
Well satisfied = 1 | 47 | 740 |
How problematic are shards of glass on the road | 100 | 2,331 |
Problematic = 1 | 49 | 1,136 |
How problematic are holes and bumps on the road | 100 | 2,369 |
Problematic = 1 | 51 | 1,212 |
Do you sometimes forfeit cycling during summer due to poor road conditions? | 100 | 2,376 |
Yes = 1 | 30 | 720 |
Winter Cycling | 100 | 2,376 |
Never/sometimes = 0 | 55 | 1,309 |
Often = 1 | 45 | 1,067 |
Variables included were developed on the basis of literature review, stakeholder's comments and previous interviews (Johansson and Bjørnskau,
Geographical area of the participant was determined using the postal code they reported. They entered this code, and the code was translated to area using an official table. The areas were then divided based on participant numbers and topographical similarity. The total population of Norway in 2019 was 5,356,800.
A map of Norway highlighting included areas.
Participants' age group was recoded from an original 10-year interval as presented in the questionnaire. They were recoded as “ <
Some cyclists use several bicycles. Participants were asked “
Cycling frequency or how often participants cycle was measured asking “
For those variables with a 7-point Likert scale, an eighth response option “
Participants' views of the use of salt on foot and cycle paths was measured using the statement “
Satisfaction with snow plowing on foot and cycle paths was measured asking “
Encountering shards of glass can be annoying, but also dangerous. How often cyclists experience this will affect how problematic it is to them. This question was phrased as “
Potholes and uneven roads were asked about in a similar way “
The variable was dichotomized so that those who ticked alternatives 5, 6 or 7 were coded as 1 and those who ticked alternatives 1–4 were coded as 0.
Choosing not to cycle due to poor conditions was measured using the question “
Accidents where conditions contributed was only asked to those 711 (30%) respondents who reported having experienced a cycle accident. The question was phrased as “
How often participants cycle during winter was asked “
All analyses were performed in SPSS 26. In the logistic regression models, a standardized effect size [OR/Exp(B)] is reported. Confidence intervals are omitted for brevity, and
For all analyses, we report
For regression analyses, we included our measure of cycling frequency separately. By including this variable, we may control away some of the variation that we actually are interested in. However, when controlling for this, we get a better insight into participants' evaluation of the importance of maintenance and operations to cycling frequency. Therefore, models are calculated with multiple steps to investigate this effect.
To allow for direct investigation of relationships of all included variables, we performed a bivariate correlation. This matrix is presented in
Bivariate correlations (Pearson's
Several relationships between dependent variables emerge as statistically significant. For example, those who have had an accident where conditions contributed do more often think that current snow plowing is insufficient and are more likely to experience holes as problematic. Similarly, those who sometimes forfeit cycling due to poor conditions also experience the other variables as problematic.
Geographical areas also yield multiple significant coefficients. Respondents from Oslo and Trondheim are more likely to agree with the statement that maintainers use too much salt. Shards of glass are least problematic in Tromsø. Respondents from Buskerud and Oslo are more likely to forfeit cycling due to poor conditions.
In contrast to the elderly, young people are more likely to experience problematic holes and bumps in the road. Elderly also have fewer accidents where conditions contributed but are more bothered by shards of glass on the road. Females cycle less often than males.
Cyclists using racer bikes are more likely to experience shards of glass on the roads as problematic. E-bike users are more susceptible to potholes and to experience accidents where conditions contributed.
When incorporating multiple predictor variables in the same model, one can investigate the effect of each variable while controlling for the effect of the others. Six separate hierarchical regression models are presented in
Six hierarchical logistic regression models predicting six different measures of impact suffered due to conditions or M&O. OR, Odds ratio [EXP(B)]. *
Few coefficients change a lot when cycling frequency is introduced. The effects of
Those who cycle often report experiencing an accident where conditions contributed to a larger extent than those who cycle seldom. Respondents from
Females are less critical of the amount of salt used then men, and younger participants are least likely to think that too much salt is used. Those who live in
In
Cyclists in
All bike types experience not executing a planned trip due to poor conditions less frequently than
Cycling conditions change a lot through the year in countries with snowy winters. Investigating which factors are important in determining whether one cycles during winter, may yield additional information about the importance of M&O. Whether one chooses to cycle during winter or not could be an important proxy for their evaluation of conditions for winter cycling.
Hierarchical logistic regression model predicting winter cycling in three steps.
Areas (ref: Other) | Akershus | 0.91 | 0.90 | 0.83 | |||
Bergen | 1.65 | 1.70 | 1.60 | ||||
Buskerud | 0.61 | 0.46 | 0.49 | ||||
Jæren | 1.88 | 1.75 | 1.49 | ||||
Oslo | 1.64 | 1.24 | 1.12 | ||||
Trondheim | 1.56 | 1.02 | 0.96 | ||||
Tromsø | 0.78 | 0.72 | 0.73 | ||||
Gender (ref: Male) | Female | 0.72 | 0.71 | 0.71 | |||
Age groups (ref: <40) | 40–49 | 1.34 | 1.32 | 1.30 | |||
50–59 | 1.20 | 1.24 | 1.25 | ||||
>60 | 0.98 | 0.93 | 0.93 | ||||
Bike type (ref: Other) | Hybrid | 1.87 | 1.50 | 1.55 | |||
MTB | 1.40 | 1.41 | 1.43 | ||||
Racer | 1.57 | 1.36 | 1.37 | ||||
E-bike | 1.58 | 1.14 | 1.18 | ||||
Often cycled | – | 9.135 | 9.60 | ||||
Too much salt | – | – | 0.84 | ||||
Satisfied with snow plowing | – | – | 1.65 | ||||
Nagelkerke |
0.061 | 0.288 | 0.300 |
It seems that cyclists in Buskerud experience M&O as quite poor, as they cycle markedly less during winter (see also
There are many factors to consider when authorities want to increase the use of bicycles for transport. In this paper we have investigated some factors related to maintenance and operation, that were highlighted by previous research. Some groups and demographics suffer more from poor conditions than others, and geographic variation is important to consider. For most problems women are more vulnerable than men. However, according to our survey, elderly cyclists are not particularly affected by poor conditions, the did for instance not consider holes in the surface to be such a great problem. This section will discuss the proposed hypotheses and the exploration of data. Limitations are considered.
Previous research has found that female cyclists are particularly prone to suffer from poor conditions and M&O (Svorstøl et al.,
Bivariate results show multiple significant relations. Women report cycling less than men, which may influence their experience with different kinds of conditions. It may also be that they find current road conditions too poor to increase their cycling. This finding is also in line with previous research finding that females perceive cycling as less available to them (Sahlqvist and Heesch,
As previous research has noted, the gender differences are more visible when conditions are particularly bad, such as icy roads and heavy snow (Johansson and Bjørnskau,
These results indicate that difficult conditions are a more pressing issue for female cyclists. The way current M&O is executed may thus be an important barrier for more women to start cycling. Previous research in a Norwegian context has suggested that women are more cautious in traffic than men (Bjørnskau,
Our second hypothesis stated that older participants would be more affected by adverse conditions than younger ones. The results from the current study do not support this hypothesis. In the regression models, there are only statistically significant results of age when assessing whether holes pose problems and whether they sometimes forfeit cycling. Both are also opposite of what one would expect given our hypothesis; older people find holes less problematic and are less prone to forfeit cycling due to poor conditions. This is true regardless of the control for cycling frequency, and seems contrary to other findings that younger cyclists are prone to accidents and risky behavior (i.e., Useche et al.,
The correlation matrix suggests that several relationships for the oldest participants are larger than what we would expect if the null hypothesis was true. They are less likely to have experienced an accident where conditions contributed, less bothered by holes and bumps, and less likely to forfeit cycling. This also strengthens our conviction to reject our hypothesis, as the effects are robust regardless of control for cycling frequency–these results are not due to other factors such as elders generally cycling less than the general populace.
While beyond the scope of this article, other research suggests an interaction effect between age and gender, especially for some kinds of conditions (Johansson and Bjørnskau,
Particularly two of the dependent variables in this study are key to investigate this hypothesis. About three fourths of our participants have experienced at least one accident where conditions contributed. Experiencing such an accident could inhibit people to cycle, and one third sometimes forfeit cycling due to poor conditions. Therefore, we conclude that our hypothesis is confirmed. These kinds of results have been expanded upon by previous research (Johansson and Bjørnskau,
Our selection of geographic areas covers quite large differences in climate. This is also evident when considering results and consistent with previously reported survey results (Johansson and Bjørnskau,
Participants from
Based on regression models, there is no reason to conclude that some areas are definitively worse than others. It seems they all have their issues and unique challenges, as their answers vary. However, forfeiting to cycle due to poor conditions, even while controlling for other variables, is a serious result. Based on this variable, it seems that
The study is based on a self-report questionnaire administered at one point in time. This means that results are affected by any bias that results from self-reporting one's own behavior and cognitions. This also means that respondents had to rely on memory when assessing how conditions were last winter, which for some was half a year in the past. The study sample may also include some self-selection bias; how much one chooses to cycle or whether one chooses to cycle during winter could be an important proxy for their evaluation of cycling conditions.
It is important to remember that the survey was administered during summer, so bike type may not be as relevant for winter conditions.
As Norwegians are allowed to cycle on pavements, formulations about road surface may not be unambiguous. There are differences in who chooses to cycle in car lanes and who chooses to cycle on pavements that may impact the results reported in this study. The survey administered in this study was at times aimed at investigating the surface condition of whatever surface the responding cyclist used to bike on, while sometimes clearly referring to specific kinds of infrastructure. We recommend that future research seeks to investigate these factors further. Additionally, there may be differences between cities in how much and what kind of cycling-specific infrastructure is available, which may exacerbate differences between them.
Another limitation is our limited control for cycling frequency (“
While generally not advised, it can sometimes be the best solution to dichotomize variables (MacCallum et al.,
These findings have clear implications for future M&O. In their pursuit to incentivize more cycling, authorities should prioritize certain areas more than others. As it is today, Norwegian cyclists are most bothered by holes and bumps in the road, shards of glass and insufficient snow plowing. However, not all groups of cyclists are equally affected by this, and geographic variations are important. Surprisingly, it seems that older cyclists are less bothered than younger ones by some difficult conditions. They are also less likely to forfeit cycling. There might be a self-selection bias explaining this—only those who still cycle have been included in the analysis, and the elderly participants may be particularly experienced and fit compared to the average elderly population. It is indeed possible that many elderly, who do not cycle and hence are not included in this sample, could potentially have cycled if the conditions were better. This effect could also extend to other parts of the results, as participants were quite experienced and chose to partake in the study voluntarily. Thus, it is difficult to conclude what measures the authorities should use to get more people to start cycling, but what we clearly see is that better maintenance and operations could in general increase cycling use, and in particular for women and for cyclists under 40. This could be on typical commuter stretches or close to kindergartens.
Given that female cyclists are more troubled by poor conditions and more risk aversive than men, improved maintenance and operation is particularly important if one wants to increase cycling among women.
Norwegian authorities want to increase the share of transportation done by cycling, particularly in city areas. To achieve this, and to do so safely, more information about maintenance and operation and how cyclists perceive conditions is needed. This article reports a nation-wide survey to 2,376 cyclists across Norway. Our results suggest that females suffer more from adverse conditions than do males. Furthermore, we find that females are less likely to cycle during winter, which may be an important proxy for their perception of winter conditions for cyclists. In contrast, older cyclists in our sample do not seem to be as affected as younger cyclists. Cycle type also matters; cyclists on racers and e-bikes are more bothered by holes in the road than riders of other cycle types. Geographic variation has a great impact on the effects of cycling conditions, and it is also evident that the cyclists' perceived quality of maintenance and operation varies greatly between geographical areas. Improved maintenance and operation could promote cycling especially for women and the potential seems to be much larger in some areas than others.
The datasets presented in this article are not readily available because they are anonymized. Requests to access the datasets should be directed to the corresponding author.
The studies involving human participants were reviewed and approved by NSD–Norwegian Centre for research data. The patients/participants provided their written informed consent to participate in this study.
OA was chiefly responsible for the literature review, analysis, and text production. TB was chiefly responsible for the discussion chapter and general tips and revision. All authors contributed to all parts of this study.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The research has been conducted at the Institute of Transport Economics. The final analyses and paper writing were financed by the Institute of Transport Economics. An initial version of the current article was peer-reviewed by the ICSC 2020 conference.