Journal of Transport Geography 106 (2023) 103502 Contents lists available at ScienceDirect Journal of Transport Geography journal homepage: www.elsevier.com/locate/jtrangeo Gendered travel behaviour in micromobility? Travel speed and route choice through the lens of intersecting identities Jerònia Cubells a, *, Carme Miralles-Guasch a, b, Oriol Marquet a, b a Grup d’Estudis en Mobilitat, Transport i Territori (GEMOTT), Departament de Geografia, Universitat Autònoma de Barcelona, Carrer de la Fortuna s/n, 08193 Bellaterra (Cerdanyola del Vallès), Barcelona, Spain b Institut de Ciència i Tecnologia Ambientals (ICTA), Universitat Autònoma de Barcelona – Edifici ICTA-ICP, Campus de Bellaterra, 08193 Cerdanyola del Vallès, Barcelona, Spain A R T I C L E I N F O A B S T R A C T Keywords: Evidence shows that the use of micromobility vehicles is currently skewed towards young men, who are more Bicycle likely to adopt risky behaviours, such as fast or aggressive riding. However, research on micromobility travel E-scooter behaviour founded on GPS data has repeatedly failed to disaggregate findings by gender, or to account for Travel behaviour intertwined sociodemographic identities. In this study, we investigate how bike-share cyclists and private e- GPS Gender scooter riders navigate through Barcelona’s cycling facilities, and whether intersecting identities (in terms of Intersectionality gender, age and parenthood) influence their performance. Using 911 GPS-tracked trips of 89 participants, we build a multilevel linear mixed effects model that analyses the speed at which cyclists and e-scooter riders travel. The results show that the speed gender gap is particularly salient among e-scooter users, while being almost non- existent among cyclists. The model further indicates that cycling facilities influence riding speed, and that intersecting identities significantly explain how micromobility users circulate through the city. Drawing on theories of performativity and embodiment, we argue that findings pointing to gender differences in travel behaviour might be explained by women’s unwillingness to negotiate traffic or tolerate harassment in highly masculinised spaces. These findings contribute to the limited available knowledge on the objective travel behaviour of micromobility users from a gender and intersectional perspective, i.e. recognising the complex way in which multiple forms of discrimination associated with identity combine. We also provide valuable insights into how the design of urban environments and targeted policies can have diverse effects on different micro­ mobility users. 1. Introduction disregarded by drivers (Egan and Philbin, 2021; Levels, 2020). When asked to share the space with pedestrians, they all need to learn how to The emergence of micromobility is changing mobility dynamics in interact with each other, since the faster pace of bicycles and e-scooters cities worldwide while adding complexity and creating new challenges. threatens walkers’ perception of safety (Che et al., 2021; Tuncer et al., A myriad of conventional and electric bicycles, electric scooters (e- 2020). While such conflicts are inherent to the nature of public space, scooters hereafter), and other light personal mobility vehicles weave they are highly dependent on how these new forms of transportation are their way through lanes, roads, and footpaths, blurring the traditional manoeuvred (Gibson et al., 2022). boundaries of transport spaces (Gibson et al., 2022). Nevertheless, the Conflicts for urban space also need to be framed in terms of gender. rise of micromobility as a new competitor for urban space has created Men on two wheels are more likely to perform risky behaviours, such as significant friction with drivers and pedestrians alike. Micromobility stunts, fast riding, swerving, quickly accelerating or braking or running users must continually negotiate their right to public space on roads and red lights (Arellano and Fang, 2019; Balkmar, 2018; Gioldasis et al., sidewalks, a negotiation that can lead to conflicts for urban space (Essa 2021; Saber et al., 2022). In Barcelona, men cyclists are not only more et al., 2018; Liang et al., 2021). When cyclists and e-scooter riders are prone to breach traffic rules (Lind et al., 2021), but they also constitute asked to share the space with cars they often feel insecure and the majority of micromobility users (Roig-Costa et al., 2021). Apart from * Corresponding author. E-mail addresses: jeronia.cubells@uab.cat (J. Cubells), carme.miralles@uab.cat (C. Miralles-Guasch), oriol.marquet@uab.cat (O. Marquet). https://doi.org/10.1016/j.jtrangeo.2022.103502 Received 27 June 2022; Received in revised form 2 November 2022; Accepted 30 November 2022 Available online 9 December 2022 0966-6923/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 modal choice and riding behaviour, gender also influences route pref­ Social differences determine not only the decision to choose micro­ erences. In this respect, women seem more reluctant to ride through off- mobility modes but also the riding behaviour during the trip. In this street lanes and paths shared with cars (Fitch and Handy, 2020; Gray­ respect, men are less prone to rule compliance, as well as more likely to stone et al., 2022). Hence, urban space conflicts are not only drawn as adopt risky and fast riding practices (Arellano and Fang, 2019; Flügel unbalanced power dynamics between vehicles with different speeds but et al., 2019; Gioldasis et al., 2021; Lind et al., 2021; Romanillos and also as conflicts in which there is an uneven share of riders in terms of Gutiérrez, 2020). These differences in travel behaviour have been gender, manoeuvring and using the space differently. These underlying traditionally attributed to a general attitude of risk aversion by women gendered processes of space conflicts have implications for urban policy. (Graystone et al., 2022; Prati et al., 2019). However, the dichotomous Therefore, studying the relations between riding behaviour, route conceptualisation of gender picturing men as fast riders and women as preferences and gender is fundamental to design inclusive spaces that safety concerned fails to capture the gendered processes supporting can ease conflicts between drivers, micromobility users, and pedes­ these differences. As some authors have recently emphasised, the theory trians. To date, only a few studies have approached riding behaviour and of performativity (i.e. how gender is socially constructed) and the theory route choice of micromobility users from a gender perspective using of embodiment (i.e. the way in which power dynamics influence objective data. mobility at the scale of the body) can deepen our understanding of this This study was set in Barcelona, where micromobility trips represent matter (Ravensbergen et al., 2019). 3.9% of the modal share (IERMB, 2020). We investigate how bike-share In the case of cyclists, fast riding has long been associated with an users (21% of micromobility share) and private e-scooter users (26% of embodied performance of normative masculinity and assertions of micromobility share) navigate the public space, and whether intersect­ spatial dominance (Popan, 2019), that some authors refer to as ‘careless ing identities in terms of gender, age and parenthood influence their masculinity’ (Balkmar, 2018). When facing such attitudes, women cy­ travel behaviour and performance. To operationalize riding behaviour clists have been found to adopt two types of strategies: (1) desisting to we analyse the speed, cycling infrastructure preferences, frequency of compete for their right to occupy the bicycle lane by changing their use, time of the day and distance travelled using the GPS-tracked trips of riding schedules/routes or (2) mirroring fast riding practices (Heim 89 private e-scooter and bike-share users during 283 days. We LaFrombois, 2019; Sersli et al., 2022). On the one hand, women reported hypothesise that (1) intersecting identities influence trip characteristics ‘feeling in the way’ when encountering aggressive male riding behav­ and spatiotemporal preferences and that (2) the outcome is different iours in the bicycle lane (Sersli et al., 2022). This feeling is rooted in the from the one expected from the simple accumulation of single identities. idea that public spaces have historically been framed as masculine, Drawing on the theory of intersectionality (Crenshaw, 1991), we expect productive spaces, and women are socialised to perceive that they that forms of oppression combine and overlap creating travel patterns should not be in them, or if they are then they should minimise their that are distinct from those expected if unidimensional identities were presence whenever possible (Heim LaFrombois, 2019; Sayagh and considered. In this article, we consider the age, gender and child-care Dusong, 2021). On the other hand, the second response to fast riding is responsibilities of participants to holistically understand their experi­ that women often feel pressured to cycle more rapidly, thus reproducing ence riding bicycles and e-scooters. masculine gender stereotypes (Heim LaFrombois, 2019; Sersli et al., Following this introduction, we contextualise how intersecting 2022). identities shape the riding experience, and how this diversity has been In addition to negotiating their presence in the public space, women regarded when constructing speed models. Next, we describe how GPS using micromobility also face the potential negative consequences of data were map-matched to the street network and how routes were increased visibility while riding. Increased visibility is perceived to inferred. The following sections explore the scheduling of trips, built intensify harassment against women, and the same is stated when environment preferences, and speed of micromobility. To conclude, we gender intersects with other identities, such as race (Balkmar, 2018; discuss the intertwined nature of cyclists’ and e-scooter riders’ travel Lubitow et al., 2019; Ravensbergen, 2022). Having to deal with behaviour with sociodemographic variables, and the resulting implica­ increased visibility and the prospect of harassment has a direct effect on tions in urban policies. the spatial behaviour of women using micromobility, as they develop specific strategies such as avoiding isolated or low-density urban spaces 2. Literature review during night-time hours (Pellicer-Chenoll et al., 2021; Sersli et al., 2022). People with other identities also develop specific strategies with Previous literature on micromobility has acknowledged a gender which to manoeuvre through the city. For instance, most parents gap, in which women are less likely to travel with micromobility modes acknowledge changing their behaviour when carrying children on their than men (Laa and Leth, 2020; Pellicer-Chenoll et al., 2021; Reilly et al., bicycles by riding more slowly or avoiding busy roads (Hatfield et al., 2020; Sardi et al., 2019). This gender split is more noticeable for e- 2019). Likewise, people of colour reported that concerns about police scooter riders than for cyclists (Bieliński and Ważna, 2020), a finding violence and unfair law enforcement treatment constrain their route also observed in Barcelona, where women represent a third of e-scooter schedules and choices (Lubitow et al., 2019). riders and, almost half of bike-sharing cyclists (Roig-Costa et al., 2021). Studies analysing cyclists’ and e-scooter riders’ gendered perfor­ Indeed, gender differences in cycling appear to be linked to the presence mance have used a variety of data sources, such as surveys, interviews of dedicated cycling facilities (Higuera-Mendieta et al., 2021; Wang and and videos (Essa et al., 2018; Gibson et al., 2022; Liang et al., 2021; Akar, 2019). Thus, bicycle use is close to gender parity in contexts with Tuncer et al., 2020). While each of these methodologies brings a specific extensive cycling infrastructure, while the gender gap is broader in those set of strengths and weaknesses, the use of GPS-tracked trips is partic­ places where dedicated infrastructure is missing (Braun et al., 2016; ularly appropriate to understand how micromobility users navigate the Mooney et al., 2019). Nonetheless, gender is not the only sociodemo­ public space. GPS devices log precise locations at regular intervals, graphic factor to influence cycling and riding rates (Bretones and Mar­ enabling researchers to derive riding behaviour indicators such as speed, quet, 2022). When micromobility use is studied through intersecting acceleration, and route choice. Most importantly, GPS devices also allow identities, researchers have found that it tends to be skewed towards tracks to be associated with GIS context data on the characteristics of the young, employed, highly educated, white men without children (Dill road or infrastructure. This is relevant, as elements of the built envi­ and McNeil, 2020; Reck and Axhausen, 2021; Soltani et al., 2019; Xin ronment such as slope, presence of cycling facilities, or density of in­ et al., 2018). These social differences in micromobility use permeate at tersections are known to influence riding behaviour (Almannaa et al., all levels, from travel behaviour indicators such as the number of trips, 2021; Eriksson et al., 2019; Zuniga-Garcia et al., 2021). Notwith­ distance travelled or trip purpose, to riding performance such as speed, standing, riding behaviour does not depend solely on the characteristics acceleration, or aggressiveness. of the infrastructure as sociodemographic variables have also been 2 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 found to be meaningful (Clarry et al., 2019; Hatfield et al., 2019; pedestrianised streets (Fig. 1). Bicycle lanes are physically separated Lubitow et al., 2019; Romanillos and Gutiérrez, 2020). Despite this, from motorised traffic and pedestrians, in contrast with bicycle sharrows most studies to date have failed to disaggregate data by gender, or only and pedestrianised streets. Bicycle sharrows (a portmanteau of ‘shared’ have a small sample size of people who do not identify as men (Eriksson and ‘arrow’) are one-way streets where riders share the driveway with et al., 2019; Li et al., 2020; Zuniga-Garcia et al., 2021). Similarly, the motorised vehicles while pedestrians circulate on the sidewalk. By few speed models that include multiple sociodemographic variables, contrast, the pavement is shared with both pedestrians and motorised analyse them independently rather than intertwined (Romanillos and vehicles in pedestrianised streets. These streets are usually located in Gutiérrez, 2020; Xu et al., 2015). Moreover, speed models have been areas with high pedestrian flows, and micromobility vehicles can usually frequently drawn from smartphone apps that track micromobility trips, be ridden both ways while motorised vehicles circulate one way with a such as Strava, Map My Ride, Bikemap or Endomondo. Indeed, most of maximum speed set to 10 or 30 km/h. the studies aimed at analysing speed patterns have used smartphone Bicycles and e-scooters represent up to 4% of all the trips conducted apps as the main data source: Clarry et al. (2019) explored 5713 trips in the city, and the number of cyclists and riders has grown by 14% since made by 554 cyclists, Flügel et al. (2019) investigated 48,633 cycling 2018 (Felipe-Falgas et al., 2022; IERMB, 2020) together with an trips undertaken by 328 (e-)cyclists and Romanillos and Gutiérrez expanding fleet of food delivery workers that travel through the cycling (2020) examined 6022 routes uploaded by 328 cyclists. The represen­ infrastructure (Alvarez-Palau et al., 2021). We narrowed our analysis to tativeness of tracking apps users has largely been questioned, not only conventional (non-electric) bicycles of the public bike-sharing system as because they mainly identify as men, but also because these apps well as private e-scooters. The dock-based bicycle sharing system, Bicing, exacerbate normatively masculine cycling behaviour (Barrie et al., has >100,000 registered users and holds a fleet of 7000 vehicles (Sor­ 2019; Garber et al., 2019; Schirck-Matthews et al., 2021). iguera and Jiménez-Meroño, 2020). The membership has an annual cost of €50 and offers unlimited free-of-charge trips for journeys shorter than 3. Methodology 30 min. Each additional 30 min of use has a fee of €0.7 and trips longer than 2 h are charged €5/h (Bustamante et al., 2022). In contrast, the 3.1. Study setting municipality does not currently offer an e-scooter-sharing platform nor does it allow private e-scooter enterprises to operate within the city This study was set in the municipality of Barcelona, which is located limits. Thus, all current e-scooter users in Barcelona ride their privately on the northwest coast of the Mediterranean Sea, and housing 1,636,732 owned vehicles. residents (2020). The city has a compact planning organisation and an even spatial distribution of services and amenities. (Marquet and 3.2. Data gathering Miralles-Guasch, 2018). Barcelona is equipped with over 200 km of cycling infrastructure divided into bicycle lanes, bicycle sharrows and Between September 2020 and July 2021, the NEWMOB project Fig. 1. Cycling infrastructure typologies within the study context. Source: Own elaboration. 3 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 surveyed 902 micromobility users in Barcelona streets. Within this 3.5. Sample characteristics and statistical analysis operation, a total of 326 e-scooter owners, 251 moto-sharing drivers and 325 bike-sharing users were interviewed using an intercept survey and The final data set used to study the performance of cyclists and riders following the CAPI (Computer Assisted Personal Interviewing) meth­ consisted of 35,538 GPS data points that belonged to 911 trips which odology (more information is available in the study by Roig-Costa et al., were undertaken on 283 different days by 55 unique cyclists and 34 2021). Of all the interviewees, a sample of 70 conventional bike-sharing unique e-scooter riders. In this sample, most participants identified as cyclists and 65 e-scooter riders was further asked to wear a GPS device men (56%), were younger than 35 years of age (62%), graduated (58%), (QStartz BT-Q1000XT; QStarz International Co., Ltd., Taiwan, R.O.C.) employed (79%) and childless (76%). In comparison to Barcelona’s for a week. We invited participants to wear the devices all day long population, young adults and men were overrepresented in our sample, except when sleeping. At the end of the study, participants received a 20- although participants shared similar levels of education and employ­ trip ticket for the city’s public transport. To characterise daily mobility ment with the rest of the citizens (Ajuntament de Barcelona, 2022; patterns rather than single trips, we excluded those participants who had Idescat, 2021, Idescat, 2021b). The survey handled out to micromobility not worn the GPS a minimum of 8 h in 1 of the 7 days they were given users offered multiple options to reflect their demographic diversity, the device. That resulted in a sample of 93 valid users. particularly regarding gender identity (woman, man, non-binary). Nonetheless, all the participants ended up choosing one of the first two options, and the analysis on gender therefore only evaluated in­ 3.3. GPS data and travel log dividuals who self-identified as women or men. Each user participated for an average of 3.2 days (sd = 1.7) and 10.2 trips (sd = 8.8) over the GPS devices were programmed to log the exact location of the week they were given the GPS device. participant every 15 s. Afterwards, GPS data were processed through the We used descriptive statistics and bivariate analysis to characterise Human Activity Behavior Identification Tool and data Unification Sys­ trip attributes and cycling infrastructure use. Thanks to the map- tem (HABITUS) software. HABITUS uses a heuristic algorithm to iden­ matching process, we were able to calculate the average speed at each tify trips within GPS trajectories and infer their transport mode by street segment (Fig. 2). Moreover, a multilevel linear mixed effects calculating the distance and speed between sequential GPS points model was run, modelling the speed of cyclists and e-scooter riders as (Berjisian and Bigazzi, 2022). Therefore, this software classifies trips the sum of urban attributes, sociodemographic characteristics, and with a 90th percentile speed of ≥10 km/h and < 25 km/h as micro­ temporal variables. Specifically, speed was adjusted by travel mode mobility trips. Trips with a 90th percentile speed below or above the (bicycle; e-scooter), sociodemographic variables (self-reported gender; aforementioned speed threshold were respectively classified as walking age; having a child under 18 years), time variables (type of day; time of and driving. These parameters have been used in previous research to day), as well as road characteristics (net route slope; cycling infra­ successfully distinguish micromobility journeys from walking or structure typology; traffic lights within 20 m from the observation, in­ motorised trips (Carlson et al., 2015). Given the scope of the research, tersections within 35 m from the observation). We chose the same we filtered out walking and driving trips from the analysis. distance to traffic lights and intersections as Clarry et al. (2019) to es­ HABITUS does not discern between e-scooter and bicycle trips, hence timate the effect of these urban elements. To test whether travel a travel diary was designed to assist in the task of identifying the specific behaviour was gendered and whether intertwined identities (consid­ mode of each micromobility trip. Participants were asked to fill out a ering gender, age and parenthood) moderated the riding and cycling travel diary at the end of each participated day, requiring them to input speed, we intersected sociodemographic variables within them and with daily information on the timing, travel mode, and purpose of individual the use of cycling facilities, as well as the time of day. The model, which trips. To avoid human errors associated with travel diaries, such as was built using the R package ‘lme4’ (Bates et al., 2015), included user- incomplete entries or misreported time stamps (Kang et al., 2018), we specific and trip-specific random effects to control for unobserved het­ filtered out those participants who combined e-scooter and bicycle trips erogeneity. To facilitate the interpretation of the model, we computed in a single day (n = 4). None of the participants used different travel and plotted the marginal effects employing the ‘ggeffects’ R package modes in different days. Consequently, the remaining users (n = 89) (Lüdecke, 2018). We only calculated the marginal effects of those var­ were either cyclists or e-scooter-riders. iables that tested for statistical significance at a 95% level of confidence in the model’s analysis of variance. The outcome of this process is the 3.4. GIS processing predicted speed for each factor’s category holding the rest of the vari­ ables at their average value. To analyse how cyclists and e-scooter riders navigate in the public space, GPS tracks were map-matched. GPS points were snapped to the 4. Results street network using a dynamic buffer with an initial search tolerance of 25 m. If the process failed, the search radius was gradually expanded by 4.1. Trip characteristics an interval of 25 m with the upper tolerance radius set to 150 m. In comparison with other studies, we have used a more constrained search Cyclists and e-scooter riders present distinct travel patterns when radius to avoid inaccuracies in the route generation process (Li et al., navigating public space. Bicycle trips of users of Barcelona’s bike- 2020; Romanillos and Gutiérrez, 2020). As a result, 70.4% of GPS points sharing system last significantly longer (2:32 min difference), cover a were map-matched with a search radius set to 50 m; 98.0%, with a 100 greater distance (0.7 km difference) and are less frequent (0.5 daily trips m radius and 99.4%, with a 150 m radius. Routes were then created difference) than trips made using privately-owned e-scooters (Table 1). employing the Network Analysis toolset from ArcGIS Pro, which uses Regarding how gender might influence cyclists’ and e-scooter riders’ Dijkstra’s algorithm to estimate the shortest path between consecutive performance, the descriptive results indicate that women spend more points. To remove noisy data trips that either (1) lasted <2 min or >2 h time travelling and undertake longer trips than men (30 s and 0.4 km (n = 234), (2) whose average speed was >60 km/h (n = 3), or (3) that difference for cyclists; 26 s and 0.1 km for e-scooter riders). Both contained unmatched points to the street network (n = 7) were filtered micromobility vehicles use the cycling infrastructure quite similarly. out from the analysis (Clarry et al., 2019). Similarly, trips with obser­ Almost half of all bike-share and private e-scooter trips take place using vations above 60 km/h (n = 98) were also excluded (Flügel et al., 2019). bicycle lanes (44.2%; 42.9%) and the second most used environment is As result, 27.3% of generated routes were excluded during the data pedestrianised streets (20.1%; 23.6%). It is noteworthy that bicycle cleaning process. The number of routes there were in each network lanes and pedestrianised streets are found in 14.6% and 27.8% of Bar­ segment is captured in ‘Appendix A Supplementary material’. celona streets, respectively. The main difference between the two 4 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 Fig. 2. Average speed on each street network link. Source: Own elaboration. Table 1 Trip characteristics by transport mode, gender and cycling infrastructure. Time in min (sd) Distance in km (sd) N of daily trips (sd) Cycling infrastructure use as % of trip occurring within (sd) Bicycle lane Bicycle sharrow Pedestrianised street None Shared bicycle Women 11:06 (8:52)* 3.0 (3.2)* 2.9 (1.9)* 44.8 (31.5) 10.2 (15.4) 17.4 (27.2) 27.5 (28.2)* Men 10:36 (7:59)* 2.6 (2.7)* 3.0 (1.9)* 43.7 (32.2) 12.8 (20.3) 22.1 (33.3) 21.4 (26.5)* Total 10:48 (8:22)** 2.8 (2.9)** 3.0 (1.9)** 44.2 (31.9) 11.7 (18.4) 20.1 (30.9) 24.0 (27.4) Private e-scooter Women 8:59 (7:00)* 2.1 (2.1)* 3.9 (2.6)* 46.1 (33.5) 12.2 (21.5) 23.9 (33.6) 17.8 (21.6)* Men 8:08 (6:46)* 2.0 (2.6)* 3.1 (2.6)* 39.5 (32.6) 13.9 (20.6) 23.3 (33.2) 23.3 (27.2)* Total 8:34 (6:54)** 2.1 (2.4)** 3.5 (2.6)** 42.9 (33.2) 13.0 (21.1) 23.6 (33.4) 20.5 (24.6) * Significant p-values (<0,05) in the ANOVA test between gender within each transport mode. ** Significant p-values (<0,05) in the ANOVA test between transport modes. vehicles is that e-scooter riders seem less likely to choose streets without towards mobility (IERMB, 2020). On weekdays, the volume of trips any adapted infrastructure: only 20.5% of e-scooter trips take place in taken in Barcelona (accounting for active mobility, public transport, and unprotected streets, compared to 24.0% of bike-share trips. Results also private transport) spikes at 8:00 and 11:00, and reaches its peak around indicate that route choice between men and women is unequal. Up to 18:00. Similarly, micromobility usage has its first peak at 9:00 and 27.5% of all bicycle trips undertaken by women occur within streets reaches its summit at 13:00–14:00 and 18:00–19:00. In contrast to e- with no bike infrastructure whatsoever, whereas men tend to prefer scooter travels, bike-share trips have an additional peak hour around bicycle sharrows (2.6% difference) and pedestrianised streets more 16:00. There is an even number of women and men cycling during the often (4.7% difference). In contrast, women riding an e-scooter opt for daytime. Nonetheless, most of the bike-share trips taken from 18:00 bicycle lanes (6.6% difference) in greater proportions than men. until midnight are made by men. For e-scooters, men’s use peaks at In terms of timing, bicycle and e-scooter trips share overall mobility 11:00 and 21:00, while women’s spikes are closer in time. Indeed, rush hours, although their occurrence during the day is slightly delayed. women represent most of the e-scooter users during the afternoon Fig. 3 compares our micromobility data set to the annual travel survey valley. that interviews a representative sample of residents of the Barcelona Metropolitan Region to study their daily mobility habits and attitudes 5 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 Fig. 3. (a) Hourly share of total and micromobility trips. (b) Hourly share of bike-share trips by gender. (c) Hourly share of private e-scooter trips by gender. Source: Own elaboration using travel data from IERMB (2020) for non-micromobility modes. 4.2. Speed patterns difference of 2.84 km/h). On the contrary, men and women cycle at similar speeds, with pedestrianised streets being the urban environment The results of the model indicate that cycling facilities shape how having a wider speed gap (speed difference of 1.88 km/h). The age cyclists and e-scooter riders navigate since vehicle speed significantly group that travels the fastest are young (16–24 years of age) e-scooter fluctuates depending on the urban environment its users are riding riders (predicted speed of 11.30 km/h). According to our results, e- through (Table 2). The predicted speed for each cycling facility can be scooter users tend to slow down as they age. Although, we observe that found in ‘Appendix A Supplementary material’. Bicycle lanes are the men tend to postpone the beginning of this downward trend when cycling infrastructure where the highest speed is reached (with a pre­ compared to women. Among bicycle users, we see the opposite trend: dicted speed of 10.46 km/h), followed by pedestrianised streets (9.31 young cyclists are the age group that travels the slowest (predicted speed km/h), and bicycle sharrows (9.17 km/h). Together with the absence of of 8.44 km/h) and speed slightly increases as the age group changes. cycling facilities (9.02 km/h), other urban elements that slow down the Moreover, having a child under 18 years seems to significantly decrease micromobility traffic flow are intersections (predicted difference of 0.72 travel speed (predicted speed difference of 1.22 km/h). This reduction is km/h) and traffic lights (predicted difference of 0.86 km/h). Speed has notable among young women, whereas men experience it at later age not been found to differ significantly due to slope or time of day vari­ stages. In general, micromobility users travel quicker at night, although ables. Additionally, e-scooter users ride significantly slower than cyclists these differences are more pronounced among women. However, when although it only translates into a 0.52 km/h predicted speed difference. the results are disaggregated by transport mode, opposite gender pat­ Despite being significant, speed differences are not substantially terns arise, which might be due to a small sample size of users travelling meaningful when only considering single variables. Nevertheless, when at night. transport mode is intersected with other variables (two- and three-way interactions), the model indicates that the speed at which e-scooters 5. Discussion and bicycles are ridden is significantly influenced by the sociodemo­ graphic background of the participant. Our results shed light on the influence of intersecting identities on To facilitate the interpretation of interaction terms in Table 2, we the trip schedules, spatial preferences and travel speed of bike-share calculated the marginal effects of those that had tested for statistical users and e-scooter riders navigating through Barcelona streets. In significance in the model’s analysis of deviance. Fig. 4 shows the pre­ terms of timing, we found that bike-share and e-scooter trips overlap dicted speed at which micromobility users drive, which has been with commuting trips, a pattern also observed in the micromobility adjusted for different transport modes, sociodemographic characteris­ schemes of other cities (Li et al., 2020; Talavera-Garcia et al., 2021). Our tics, cycling infrastructures and time of the day. Speed differences due to findings, however, show that these modes are also highly used during gender are more noticeable for e-scooter riders. Male e-scooter users ride the afternoon, especially by women, which might indicate a link be­ faster than their female counterparts. Fig. 4 shows that these speed tween micromobility trips and non-work-related trip purposes. This is dissimilarities are greater in bicycle sharrows (with a predicted speed consistent with the findings in the study by Roig-Costa et al. (2021), 6 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 Table 2 Table 2 (continued ) Model 1. Micromobility users’ speed according to sociodemographic attributes, Coefficient Std. p- street-network properties, and time variables. Error value* Coefficient Std. p- E-scooter:Man: Error value* Afternoon Intercept 12.51 0.74 E-scooter:Woman: 1.45 1.02 Evening E-scooter:Man: Infrastructure characteristics 0.00 0.98 Evening Cycling infrastructure None =ref =ref <0.001 E-scooter:Woman: Bicycle lane 1.48 0.34 − 6.41 3.22 Night Bicycle sharrow 0.70 0.49 E-scooter:Man:Night 2.61 1.63 Pedestrianised street Micromobility mode 0.69 0.46 (<30 km/h) Mode:Gender E-scooter:Male 4.46 1.44 0.049 Intersection at 35 m No =ref =ref <0.001 Mode:Age E-scooter:16–24 4.27 1.43 0.001 Yes − 0.65 0.15 E-scooter:35–44 − 1.51 1.34 Traffic light or stop E-scooter:>45 − 0.25 1.31 No =ref =ref <0.001 sign at 20 m E-scooter: Yes − 0.87 0.15 Mode:Gender:Age − 5.07 2.14 0.030 Man:16–24 Net route slope − 0.01 0.01 0.202 E-scooter: − 0.66 1.75 Man:35–44 E-scooter:Man:>45 − 4.07 1.85 Time variables Morning Time of day =ref =ref 0.096 (08:00–12:00) Sociodemographic variables Afternoon Gender:Age Man:16–24 1.24 1.21 0.854 − 1.26 0.71 (12:00–18:00) Man:35–44 1.32 1.16 Evening Man:>45 3.53 1.51 − 1.58 0.73 (18:00–00:00) Gender:Age:Having a Night (00:00–06:00) 2.47 1.24 Woman:25–34:Yes 0.67 1.80 0.011 child under 18 years Day of the week Weekday =ref =ref 0.569 Man:25–34:Yes 2.58 2.15 Weekend 0.19 0.34 Woman:16–24:Yes − 1.19 2.05 Man:16–24:Yes 2.85 1.68 Woman:35–44:Yes 3.76 1.74 Micromobility mode Man:35–44:Yes 1.04 1.71 Mode Bicycle =ref =ref <0.001 Woman:>45:Yes 4.76 1.71 E-scooter − 0.22 1.03 Random effects Sociodemographic variables Route-specific 8.31 2.88 Gender Woman =ref =ref 0.995 User-specific 0.46 0.68 Man − 2.04 0.99 Age 16–24 − 0.73 0.82 * Analysis of Deviance (Type II Wald chi-square tests). 25–34 =ref =ref 0.784 35–44 − 0.02 0.91 >45 − 1.60 1.24 where it was noted that there is an association between micromobility Having a child under No =ref =ref 0.031 use in Barcelona and personal purposes, and also with other similar 18 years findings in Gdansk (Bieliński and Ważna, 2020) and Vienna (Laa and Yes − 3.03 1.33 Leth, 2020). During night hours, a significant gender gap is observed among cyclists, as men are much more likely to use the bike-sharing Interaction terms system than women starting from 18:00 onwards. Significantly, this Infrastructure characteristics gender gap is not found among e-scooter riders, which might have to do Mode: Cycling E-scooter:Bicycle with differences in vehicle ownership and the docked nature of Barce­ − 0.27 0.52 0.024 infrastructure lane lona’s public bike-sharing systems. While privately owned e-scooters E-scooter:Bicycle might be ridden door to door, bicycles must be picked up and parked in − 1.93 0.74 sharrow docking stations. This lapse of parking the bicycle and walking to the E-scooter: Pedestrianised street − 0.38 0.67 destination might slow cyclists’ pace after a moment of increased visi­ Mode:Gender: Cycling Bicycle:Man:Bicycle bility that is perceived to intensify harassment (Lubitow et al., 2019; − 0.75 0.47 0.030 infrastructure lane Ravensbergen, 2022). This is also consistent with our finding that E-scooter:Man: 1.14 0.57 women cycle much faster at night. Accelerating to compensate for Bicycle lane Bicycle:Man: Bicycle increased visibility is in line with other research, which noted that − 0.49 0.65 women develop specific strategies when cycling at night (Pellicer-Che­ sharrow E-scooter:Man: 2.14 0.75 noll et al., 2021). Furthermore, it should be noted that bike-sharing trips Bicycle sharrow duration might be constrained by their fee structure, with which in­ Bicycle:Man: − 1.17 0.63 dividuals who surpass 30-min rides are charged (Bustamante et al., Pedestrianised street E-scooter:Man: 2022). 0.36 0.69 Pedestrianised street On a spatial level, there are significant differences on how bicycles Time variables and e-scooters manoeuvre through different cycling infrastructure. Gender:Time of day Man:Afternoon 1.16 0.94 0.045 Consistent with the literature, we found that bicycle lanes are the type of Man:Evening 1.07 0.95 Man:Night − 5.23 1.72 infrastructure leading to greater speeds for both shared bicycles and Mode:Gender:Time of E-scooter:Woman: private e-scooters (Arellano and Fang, 2019; Flügel et al., 2019; Zuniga- 1.06 0.97 0.014 day Afternoon Garcia et al., 2021). It is noteworthy that women subscribed to the bike- − 1.38 0.89 sharing system are the group that uses cycling infrastructure the least, particularly when it is shared with motorised vehicles. This result may 7 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 Fig. 4. Predicted travel speed (km/h) for interaction terms. Source: Own elaboration. be explained by the fact that some women cyclists might feel discour­ women mirroring normative masculine fast-driving practices when aged to compete for their right to occupy that space, even in streets with using cycling facilities, as they might feel nudged to cycle rapidly to fully dedicated cycling infrastructure (Heim LaFrombois, 2019; Rav­ avoid ‘feeling in the way’ (Heim LaFrombois, 2019; Sersli et al., 2022). ensbergen, 2022; Sersli et al., 2022). This unwillingness to assertively The tendency of women to either evade the cycling infrastructure or use negotiate traffic is more persistent in spatial contexts with strong norms it at higher speeds than expected would indicate that both strategies of of male appropriation of public space (Sayagh and Dusong, 2021) and desisting to compete for the lane and embodying fast cycling practices Barcelona’s traffic could be framed as one of such environments since are coexisting at the same time in Barcelona’s streets. the drivers’ fleet is highly masculinised (Cubells et al., 2020). Avoiding In contrast, women riding e-scooters appear to almost exclusively cycling through traffic may also go hand in hand with an unwillingness choose routes with dedicated cycling infrastructure. While they adopt to tolerate harassment from drivers (Balkmar, 2018; Graystone et al., similar speeds as cyclists, men riding e-scooters are the micromobility 2022; Lubitow et al., 2019). Indeed, half of the women and non-binary users who travel the fastest in each type of cycling facility, a finding that cyclists who were surveyed in Barcelona reported to have experienced is also reported by Arellano and Fang (2019). Reckless men using an aggression due to their gender identity or expression, and 30% of micromobility modes have been identified to be problematic for other them stated that they had changed their route choice because of these users riding at moderate speeds, as these users might feel discouraged to encounters (Col⋅lectiu Punt 6, 2021). All in all, women cyclists in our take the lane (Balkmar, 2018; Sersli et al., 2022). Risk-taking and fast- sample seemed to use dedicated bike infrastructure at a lower rate than riding practices are examples of how careless masculinity, in this case, what would be expected. performed by micromobility users, can actively dissuade other potential Nonetheless, when women choose to circulate using the dedicated users to share the same infrastructure, which would hinder the adoption bike infrastructure, they tend to travel faster than the rest of the users, potential of micromobility modes and hurt the prospects of a more only being overtaken by men riding e-scooters. Our findings indicate sustainable mobility. that there is no gender gap in terms of speed among bike-share users, as Intersecting identities in terms of gender, age and parenthood are no significant speed differences were found between men and women. observed to further influence how micromobility users travel. Age has These findings contradict previous studies on the subject (Flügel et al., not significantly explained the variance of prior speed models for bi­ 2019; Romanillos and Gutiérrez, 2020). They seem to be a result of cycles, although the highest speeds were observed among middle-aged 8 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 users (Flügel et al., 2019; Romanillos and Gutiérrez, 2020). The fact that CRediT authorship contribution statement middle-aged adults travel faster than younger adults could be associated with increased confidence, due to spending more time cycling (Sersli Jerònia Cubells: Conceptualization, Data curation, Formal analysis, et al., 2022). According to our findings, whereas young bike-share cy­ Methodology, Software, Validation, Visualization, Writing – original clists are the age group that travels the slowest, young e-scooter riders draft. Carme Miralles-Guasch: Resources, Supervision, Project are the age group that travels the fastest, especially men. This is in line administration. Oriol Marquet: Conceptualization, Methodology, with previous studies that identified young men riders as being more Validation, Resources, Writing – review & editing, Supervision, Project likely to perform risky behaviours (Gioldasis et al., 2021). Another administration. factor that has been observed to shape micromobility users’ perfor­ mance is having children. Prior research revealed that cyclists eclaration of Competing Interest acknowledged the fact that they travelled slower when cycling with children (Hatfield et al., 2019; Janke and Handy, 2019). Our results None. quantitatively show that parents of minor children ride at a slower pace, even when travelling alone. It should be noted that Barcelona’s bicycle- Data availability sharing system does not support child bike seats or trailers, and that it is forbidden by law to carry passengers when using an e-scooter (Ajunta­ The authors do not have permission to share data. ment de Barcelona, 2017). Therefore, our results suggest that having children permeates to travel behaviour, even when not travelling with minors, a finding that had been previously reported only for other Acknowledgements transport modes (Maciejewska and Miralles-Guasch, 2019). Further­ more, being a parent seems to influence the travel behaviour of men and The authors wish to express their gratitude to the participants who women at different age stages. shared their time for this research. We are grateful to Oriol Roig-Costa for managing the data-gathering process. Our research colleagues Zey­ 6. Conclusions nep Akinci, Laia Mojica, Samuel Nello-Deakin and Guillem Vich have also shared insightful comments and suggestions during the design and This article has explored the travel behaviour of bike-share cyclists development of this project. In addition, the authors would like to thank and e-scooter riders using GPS-tracked trips. We have studied how the anonymous reviewers and the editor for providing valuable feedback micromobility users navigate multiple cycling facilities along the day to that helped to improve the manuscript. This work was supported by the test whether (1) intersecting identities (in terms of gender, age and Spanish Ministry of Science and Innovation (Electric, light and Shared: the parenthood) influenced trip characteristics and spatiotemporal prefer­ rise of micromobility in Spain and its environmental, social and health con­ ences and whether (2) axes of identity resulted in combined and over­ sequences. a multi method study using GIS, tracking and accelerometry lapped travel patterns different than the simple accumulation of (MICROMOV), PID2019-104344RB-I00); the Generalitat de Catalunya unidimensional identities. Indeed, our results have demonstrated that (Inclusiva, Sostenible, Saludable i Resilient. La Mobilitat i la Ciutat en l’Es­ intertwined identities deeply influence riding behaviour among new cenari Postpandemia, PANDE 00023, 2020); and the Institut de Cultura, micromobility users in ways that could not be explained by the sum of Ajuntament de Barcelona (New mobility in the city: NEWMOB, 19S01360 single sociodemographic characteristics. In particular, we have found -006). J. Cubells is supported by a PhD grant by the Agency for Man­ that men and women have distinct travel behaviour, which is further agement of University and Research Grants (AGAUR, FI_B 00063, 2022) defined by their age and whether they are parents. For instance, we have and O. Marquet is funded by a Ramón y Cajal fellowship (RYC-2020- observed that child-care has opposite effects on riding speed for men and 029441-I), awarded by the Spanish Ministry of Science and Innovation. women depending on their age. This study is not without limitations. We note that this analysis is Appendix A. Supplementary data specific to the urban context and mobility schemes of Barcelona. Further research is required to comprehend the applicability of the findings in Supplementary data to this article can be found online at https://doi. other contexts and transport modes, such as private bicycles. In com­ org/10.1016/j.jtrangeo.2022.103502. parison to studies using smartphone GPS data, we gathered a smaller sample while overcoming some of their limitations concerning repre­ References sentativeness, at least in terms of gender parity. In addition, GPS data accuracy is not precise enough to capture the exact location of trips Ajuntament de Barcelona, 2017. Ordenança de circulació de vianants i de vehicles [WWW Document]. Normativa Local. URL. https://cido.diba.cat/normativa_local/ occurring on streets with different infrastructures, such as sidewalks, 3925/ordenanca-de-circulacio-de-vianants-i-de-vehicles-ajuntament-de-barcelona bicycle lanes, and roads. In these cases, it was assumed that cycling (accessed 5.27.22). infrastructure was used. If cycling facilities were lacking, we could not Ajuntament de Barcelona, 2022. Evolució de la població ocupada [WWW Document]. Treball, Població activa i ocupada. URL. https://ajuntament.barcelona.cat/barcelo determine whether users were circulating on the road or the pavement. naeconomia/ca/mercat-de-treball/poblacio-activa-i-ocupada/evolucio-de-la- Also, machine learning methods, in contrast to heuristic algorithms, poblacio-ocupada (accessed 9.2.22). promise to outperform trip detection in GPS trajectories in the future, Almannaa, M.H., Ashqar, H.I., Elhenawy, M., Masoud, M., Rakotonirainy, A., Rakha, H., 2021. A comparative analysis of e-scooter and e-bike usage patterns: findings from which would increase the accuracy of trip characteristics and mode the City of Austin, TX. Int. J. Sustain. Transp. 15, 571–579. https://doi.org/10.1080/ inference (Berjisian and Bigazzi, 2022). Finally, we acknowledge that 15568318.2020.1833117. this analysis could not account for the variation in gendered identities, Alvarez-Palau, E.J., Calvet-Liñán, L., Viu-Roig, M., Gandouz, M., Juan, A.A., 2021. Economic profitability of last-mile food delivery services: Lessons from Barcelona. and future research should be aimed at exploring the travel behaviour of https://doi.org/10.1016/j.rtbm.2021.100659. other social groups of cyclists and riders, such as non-gender normative Arellano, J.F. (Frank), Fang, K., 2019. Sunday drivers, or too fast and too furious? people or people with racialised experiences. Transp. Find. https://doi.org/10.32866/001c.11210. Balkmar, D., 2018. Violent mobilities: men, masculinities and road conflicts in Sweden. In conclusion, our study has provided a framework to acknowledge Mobilities 1–16. https://doi.org/10.1080/17450101.2018.1500096. the intertwined nature of cyclists’ and e-scooter riders’ travel behaviour Barrie, L., Waitt, G., Brennan-Horley, C., 2019. Cycling assemblages, self-tracking digital with axes of identity using objective spatial data. Our findings highlight technologies and negotiating gendered subjectivities of road cyclists on-the-move. the need to take an intersectional approach to individual characteristics Leis. Sci. 41, 108–126. https://doi.org/10.1080/01490400.2018.1539682. Bates, D., Mächler, M., Bolker, B.M., Walker, S.C., 2015. Fitting linear mixed-effects when seeking to design policies aimed at regulating micromobility and models using lme4. J. Stat. Softw. 67, 1–48. https://doi.org/10.18637/JSS.V067. create inclusive urban environments. I01. 9 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 Berjisian, E., Bigazzi, A., 2022. Evaluation of methods to distinguish trips from activities IERMB, 2020. Enquesta Mobilitat en dia Feiner (EMEF) [WWW Document]. Enquestes de in walking and cycling GPS data. Transp. Res. Part C Emerg. Technol. 137, 103588 Mobilitat. URL. https://iermb.uab.cat/ca/enquestes/enquestes-de-mobilitat/#144 https://doi.org/10.1016/j.trc.2022.103588. 7843451840-2-0 (accessed 4.5.22). Bieliński, T., Ważna, A., 2020. Electric scooter sharing and bike sharing user behaviour Janke, J., Handy, S., 2019. How life course events trigger changes in bicycling attitudes and characteristics. Sustainability 12, 9640. https://doi.org/10.3390/su12229640. and behavior: insights into causality. Travel Behav. Soc. 16, 31–41. https://doi.org/ Braun, L.M., Rodriguez, D.A., Cole-Hunter, T., Ambros, A., Donaire-Gonzalez, D., 10.1016/j.tbs.2019.03.004. Jerrett, M., Mendez, M.A., Nieuwenhuijsen, M.J., de Nazelle, A., 2016. Short-term Kang, M., Moudon, A.V., Hurvitz, P.M., Saelens, B.E., 2018. Capturing fine-scale travel planning and policy interventions to promote cycling in urban centers: findings from behaviors: a comparative analysis between personal activity location measurement a commute mode choice analysis in Barcelona, Spain. Transp. Res. Part A Pol. Pract. system (PALMS) and travel diary. Int. J. Health Geogr. 17, 40. https://doi.org/ 89, 164–183. https://doi.org/10.1016/j.tra.2016.05.007. 10.1186/s12942-018-0161-9. Bretones, A., Marquet, O., 2022. Sociopsychological factors associated with the adoption Laa, B., Leth, U., 2020. Survey of E -scooter users in Vienna : Who they are and how they and usage of electric micromobility. A literature review. Transp Policy (Oxf). ride. J. Transp. Geogr. 89, 102874 https://doi.org/10.1016/j.jtrangeo.2020.102874. https://doi.org/10.1016/J.TRANPOL.2022.09.008. Levels, A., 2020. (re-) claiming urban streets: the conflicting (auto)mobilities of cycling Bustamante, X., Federo, R., Fernández-i-Marin, X., 2022. Riding the wave: predicting the and driving in Berlin and New York. J. Transp. Hist. 41, 381–401. https://doi.org/ use of the bike-sharing system in Barcelona before and during COVID-19. Sustain. 10.1177/0022526620941200. Cities Soc. 83, 103929 https://doi.org/10.1016/j.scs.2022.103929. Li, W., Wang, S., Zhang, X., Jia, Q., Tian, Y., 2020. Understanding intra-urban human Carlson, J.A., Jankowska, M.M., Meseck, K., Godbole, S., Natarajan, L., Raab, F., mobility through an exploratory spatiotemporal analysis of bike-sharing trajectories. Demchak, B., Patrick, K., Kerr, J., 2015. Validity of PALMS GPS scoring of active and Int. J. Geogr. Inf. Sci. 34, 2451–2474. https://doi.org/10.1080/ passive travel compared with SenseCam. Med. Sci. Sports Exerc. 47, 662–667. 13658816.2020.1712401. https://doi.org/10.1249/MSS.0000000000000446. Liang, X., Meng, X., Zheng, L., 2021. Investigating conflict behaviours and characteristics Che, M., Wong, Y.D., Lum, K.M., Wang, X., 2021. Interaction behaviour of active in shared space for pedestrians, conventional bicycles and e-bikes. Accid. Anal. Prev. mobility users in shared space. Transp. Res. Part A Pol. Pract. 153, 52–65. https:// 158, 106167 https://doi.org/10.1016/j.aap.2021.106167. doi.org/10.1016/j.tra.2021.08.017. Lind, A., Honey-Rosés, J., Corbera, E., 2021. Rule compliance and desire lines in Clarry, A., Faghih Imani, A., Miller, E.J., 2019. Where we ride faster? Examining cycling Barcelona’s cycling network. Transp. Lett. 13, 728–737. https://doi.org/10.1080/ speed using smartphone GPS data. Sustain. Cities Soc. 49, 101594 https://doi.org/ 19427867.2020.1803542. 10.1016/j.scs.2019.101594. Lubitow, A., Tompkins, K., Feldman, M., 2019. Sustainable cycling for all? race and Col⋅lectiu Punt 6, 2021. Dones i persones no binàries en bici [WWW Document]. Estudi gender–based bicycling inequalities in Portland, Oregon. City Community 18, de mobilitat ciclista a Barcelona des d’una perspectiva feminista. URL. https://www. 1181–1202. https://doi.org/10.1111/cico.12470. punt6.org/wp-content/uploads/2022/07/informe-bicis-2021_catala.pdf. Lüdecke, D., 2018. Ggeffects: tidy data frames of marginal effects from regression Crenshaw, K., 1991. Mapping the margins: intersectionality, identity politics, and models. J Open Source Softw 3, 772. https://doi.org/10.21105/JOSS.00772. violence against women of color. Stanford Law Rev 43, 1241. https://doi.org/ Maciejewska, M., Miralles-Guasch, C., 2019. “I have children and thus i drive”: 10.2307/1229039. perceptions and motivations of modal choice among suburban commuting mothers. Cubells, J., Marquet, O., Miralles-Guasch, C., 2020. Gender and age differences in Finisterra 54, 55–74. https://doi.org/10.18055/finis16035. metropolitan car use. Recent gender gap trends in private transport. Sustainability Marquet, O., Miralles-Guasch, C., 2018. Resilient territories and mobility adaptation (Switzerland) 12. https://doi.org/10.3390/SU12187286. strategies in times of economic recession: evidence from the metropolitan region of Dill, J., McNeil, N., 2020. Are shared vehicles shared by all? A review of equity and Barcelona, Spain 2004–2012. Eur. Urban Reg. Stud. 25, 345–359. https://doi.org/ vehicle sharing. J. Plan Lit. https://doi.org/10.1177/0885412220966732. 10.1177/0969776417703158. Egan, R., Philbin, M., 2021. Precarious entitlement to public space & utility cycling in Mooney, S.J., Hosford, K., Howe, B., Yan, A., Winters, M., Bassok, A., Hirsch, J.A., 2019. Dublin. Mobilities 16, 509–523. https://doi.org/10.1080/17450101.2021.1913067. Freedom from the station: spatial equity in access to dockless bike share. J. Transp. Eriksson, J., Forsman, Å., Niska, A., Gustafsson, S., Sörensen, G., 2019. An analysis of Geogr. 74, 91–96. https://doi.org/10.1016/j.jtrangeo.2018.11.009. cyclists’ speed at combined pedestrian and cycle paths. Traffic Inj. Prev. 20, 56–61. Pellicer-Chenoll, M., Pans, M., Seifert, R., López-Cañada, E., García-Massó, X., Devís- https://doi.org/10.1080/15389588.2019.1658083. Devís, J., González, L.-M., 2021. Gender differences in bicycle sharing system usage Essa, M., Hussein, M., Sayed, T., 2018. Road users’ behavior and safety analysis of in the city of Valencia. Sustain. Cities Soc. 65, 102556 https://doi.org/10.1016/j. pedestrian–bike shared space: case study of Robson street in Vancouver. Can. J. Civ. scs.2020.102556. Eng. 45, 1053–1064. https://doi.org/10.1139/cjce-2017-0683. Popan, C., 2019. Slowness. In: Büscher, M., Adey, P. (Eds.), Bicycle Utopias: Imagining Felipe-Falgas, P., Madrid-Lopez, C., Marquet, O., 2022. Assessing environmental Fast and Slow Cycling Futures. Routledge, New York, pp. 172–188. performance of micromobility using LCA and self-reported modal change: the case of Prati, G., Fraboni, F., de Angelis, M., Pietrantoni, L., Johnson, D., Shires, J., 2019. Gender shared E-bikes, E-scooters, and E-mopeds in Barcelona. Sustainability 14, 4139. differences in cycling patterns and attitudes towards cycling in a sample of European https://doi.org/10.3390/su14074139. regular cyclists. J. Transp. Geogr. 78, 1–7. https://doi.org/10.1016/j. Fitch, D.T., Handy, S.L., 2020. Road environments and bicyclist route choice: the cases of jtrangeo.2019.05.006. Davis and San Francisco, CA. J. Transp. Geogr. 85, 102705 https://doi.org/10.1016/ Ravensbergen, L., 2022. ‘I wouldn’t take the risk of the attention, you know? Just a lone j.jtrangeo.2020.102705. girl biking’: examining the gendered and classed embodied experiences of cycling. Flügel, S., Hulleberg, N., Fyhri, A., Weber, C., Ævarsson, G., 2019. Empirical speed Soc. Cult. Geogr. 23, 678–696. https://doi.org/10.1080/14649365.2020.1806344. models for cycling in the Oslo road network. Transportation (Amst) 46, 1395–1419. Ravensbergen, L., Buliung, R., Laliberté, N., 2019. Toward feminist geographies of https://doi.org/10.1007/s11116-017-9841-8. cycling. Geogr. Compass 13. https://doi.org/10.1111/gec3.12461. Garber, M.D., Watkins, K.E., Kramer, M.R., 2019. Comparing bicyclists who use Reck, D.J., Axhausen, K.W., 2021. Who uses shared micro-mobility services? Empirical smartphone apps to record rides with those who do not: implications for evidence from Zurich. Switzerland. Transp. Res. D Transp. Environ. 94, 102803 representativeness and selection bias. J. Transp. Health 15, 100661. https://doi.org/ https://doi.org/10.1016/j.trd.2021.102803. 10.1016/J.JTH.2019.100661. Reilly, K.H., Noyes, P., Crossa, A., 2020. From non-cyclists to frequent cyclists: factors Gibson, H., Curl, A., Thompson, L., 2022. Blurred boundaries: E-scooter riders’ and associated with frequent bike share use in new York City. J. Transp. Health 16, pedestrians’ experiences of sharing space. Mobilities 17, 69–84. https://doi.org/ 100790. https://doi.org/10.1016/j.jth.2019.100790. 10.1080/17450101.2021.1967097. Roig-Costa, O., Gómez-Varo, I., Cubells, J., Marquet, O., 2021. La movilidad post Gioldasis, C., Christoforou, Z., Seidowsky, R., 2021. Risk-taking behaviors of e-scooter pandemia: perfiles y usos de la micromovilidad en Barcelona. Revista Transporte y users: a survey in Paris. Accid. Anal. Prev. 163, 106427 https://doi.org/10.1016/j. Territorio. https://doi.org/10.34096/rtt.i25.10958. aap.2021.106427. Romanillos, G., Gutiérrez, J., 2020. Cyclists do better. Analyzing urban cycling operating Graystone, M., Mitra, R., Hess, P.M., 2022. Gendered perceptions of cycling safety and speeds and accessibility. Int. J. Sustain. Transp. 14, 448–464. https://doi.org/ on-street bicycle infrastructure: bridging the gap. Transp. Res. D Transp. Environ. 10.1080/15568318.2019.1575493. 105, 103237 https://doi.org/10.1016/j.trd.2022.103237. Saber, F., Mirzaei-Alavijeh, M., Mostafavi-Darani, F., Zamani-Alavijeh, F., 2022. Why Hatfield, J., Poulos, R.G., Murphy, S.M., Flack, L.K., Rissel, C., Grzebieta, R., McIntosh, A. male adolescent bicyclists perform risky stunts? A qualitative study. Transp. Res. S., 2019. Safety aspects of riding with children: descriptive analysis of adult riders’ Part F Traffic Psychol. Behav. 88, 1–12. https://doi.org/10.1016/j.trf.2022.05.003. self-report. Accid. Anal. Prev. 131, 33–44. https://doi.org/10.1016/j. Sardi, P., Dutra, R., Macke, J., Sarate, R., 2019. “Green, but not as green as that”: an aap.2019.06.012. analysis of a Brazilian bike-sharing system. J. Clean. Prod. 217 https://doi.org/ Heim LaFrombois, M.E., 2019. (re)producing and challenging gender in and through 10.1016/j.jclepro.2019.01.240. urban space: women bicyclists’ experiences in Chicago. Gend. Place Cult. 26, Sayagh, D., Dusong, C., 2021. What forms of socialization lead women to stop cycling 659–679. https://doi.org/10.1080/0966369X.2018.1555142. during adolescence? Int. Rev. Sociol. Sport. https://doi.org/10.1177/ Higuera-Mendieta, D., Uriza, P.A., Cabrales, S.A., Medaglia, A.L., Guzman, L.A., 10126902211029622, 101269022110296. Sarmiento, O.L., 2021. Is the built-environment at origin, on route, and at Schirck-Matthews, A., Hochmair, H.H., Paulus, G., Strelnikova, D., 2021. Comparison of destination associated with bicycle commuting? A gender-informed approach. cycling path characteristics in South Florida and North Holland among three GPS J Transp Geogr 94, 103120. https://doi.org/10.1016/j.jtrangeo.2021.103120. fitness tracker apps. Int. J. Sustain. Transp. 1–16 https://doi.org/10.1080/ Idescat, 2021. Estadística dels estudis de la població [WWW Document]. In: Indicadors 15568318.2021.1943073. dels estudis de la població de 25 a 64 anys. Barcelona. URL. https://www.idescat. Sersli, S., Gislason, M., Scott, N., Winters, M., 2022. Easy as riding a bike? Bicycling cat/pub/?geo=mun%3A080193&id=eep&n=14468#Plegable=geo (accessed competence as (re)learning to negotiate space. Qual. Res. Sport Exerc. Health 14, 9.2.22). 268–288. https://doi.org/10.1080/2159676X.2021.1888153. Idescat, 2021b. Estimacions de població [WWW Document]. Sexe i edat quinquennal. Soltani, A., Allan, A., Anh Nguyen, H., Berry, S., 2019. Bikesharing experience in the city Barcelona. URL. https://www.idescat.cat/pub/?id=ep&n=9124&ge of Adelaide: insight from a preliminary study. Case Stud. Transp. Pol. 7, 250–260. o=mun:080193,prov:08#Plegable=geo (accessed 9.2.22). https://doi.org/10.1016/j.cstp.2019.01.001. 10 J. Cubells et al. Journal of Transport Geography 106 (2023) 103502 Soriguera, F., Jiménez-Meroño, E., 2020. A continuous approximation model for the Xin, F., Chen, Y., Wang, X., Chen, X., 2018. Cyclist satisfaction evaluation model for free- optimal design of public bike-sharing systems. Sustain. Cities Soc. 52, 101826 floating bike-sharing system: a case study of Shanghai. Transp. Res. Rec. 2672, https://doi.org/10.1016/j.scs.2019.101826. 21–32. https://doi.org/10.1177/0361198118770193. Talavera-Garcia, R., Romanillos, G., Arias-Molinares, D., 2021. Examining spatio- Xu, C., Li, Q., Qu, Z., Jin, S., 2015. Predicting free flow speed and crash risk of bicycle temporal mobility patterns of bike-sharing systems: the case of BiciMAD (Madrid). traffic flow using artificial neural network models. Math. Probl. Eng. 2015, 1–11. J. Maps 17, 7–13. https://doi.org/10.1080/17445647.2020.1866697. https://doi.org/10.1155/2015/212050. Tuncer, S., Laurier, E., Brown, B., Licoppe, C., 2020. Notes on the practices and Zuniga-Garcia, N., Ruiz Juri, N., Perrine, K.A., Machemehl, R.B., 2021. E-scooters in appearances of e-scooter users in public space. J. Transp. Geogr. 85, 102702 https:// urban infrastructure: understanding sidewalk, bike lane, and roadway usage from doi.org/10.1016/j.jtrangeo.2020.102702. trajectory data. Case Stud. Transp. Pol. 9, 983–994. https://doi.org/10.1016/j. Wang, K., Akar, G., 2019. Gender gap generators for bike share ridership: evidence from cstp.2021.04.004. Citi bike system in new York City. J. Transp. Geogr. 76, 1–9. https://doi.org/ 10.1016/j.jtrangeo.2019.02.003. 11