Active travel infrastructure design and implementation: Insights from behavioral science

Replacing car travel with walking and cycling is at the core of the shift to healthier and more sustainable societies. Implementing dedicated infrastructure is a common measure to achieve this aim. But policymakers in multiple countries regularly contend with two obstacles: designing infrastructure that people will make use of and securing public support for implementation. We review and synthesize relevant research from behavioral science that sheds light on how to overcome these two obstacles. Given available literature, we focus on cycling infrastructure. We find that research on moderators of the success of active travel initiatives points to the importance of proximity, connectivity, and safety perceptions, particularly among women, older adults and children. We review empirical findings on which design elements make infrastructure both safe to use and perceived as safe. With respect to public support, we summarize common concerns and review research from behavioral economics and psychology that may help to counter misperceptions of the effects of active travel infrastructure. We also draw on evidence regarding support for climate policy and opinion formation more generally. The paper offers an evidence‐based guide for policymakers to design and implement active travel infrastructure, seen through the lens of behavioral science. It also highlights fruitful avenues for future research.

Motorized transport remains mostly dependent on combustion engines that run on fossil fuels.As a result, the transport sector is one of the largest contributors to climate change, with one of the fastest growing greenhouse gas emission rates since 1990 (Bleviss, 2021).It is largely accepted that modal shifts to non-motorized transport are required to reduce emissions; technological substitution via electrification will not be sufficient nor occur fast enough (Brand, Götschi, et al., 2021;Creutzig et al., 2018;Cuenot et al., 2012).Promoting active travel (e.g., cycling) for daily journeys in urban areas in particular holds substantial potential for curbing sectoral emissions (Brand, Dons, et al., 2021), with knock on benefits for air quality and public health (Barr, 2018;Nocera & Attard, 2021).
Infrastructural elements such as the design of the built environment explain a substantial amount of variation in travel behavior (Brand, Götschi, et al., 2021).However, the decision to travel by car is additionally influenced by multiple individual and social factors (Ferreira et al., 2022).Techniques from behavioral science can be used to study these factors and improve understanding of the relationship between infrastructural change and behavior change.Thus, policies that aim to maximize the modal shift toward active travel for a given investment in infrastructure can be informed by evidence from behavioral science.
The application of behavioral science to policy has expanded dramatically since the early 2010s (Sanders et al., 2018).Its scope is now far beyond the mere description of cognitive biases that might be at play in a specific domain, or the application of "nudges" to influence choices in a desired direction (e.g., Ewert, 2020).Instead, there is wider recognition of the need to use methods and theory from behavioral economics and psychology to first diagnose the psychological mechanisms that drive behavior, before developing interventions that might seek to exploit any mechanisms identified (Lunn, 2019).For example, applied behavioral scientists will consider not only how the built environment influences travel mode choice through its physical features, but also by how it might affect psychological mechanisms, such as perceptions of convenience, safety and accessibility (Ferreira et al., 2022).The first aim of this paper is to provide insight into the types of design decisions that lead to more positive perceptions of infrastructure, thereby increasing the likelihood for behavioral change.
However, simply designing infrastructure in a way that encourages use is not sufficient for meeting emission reduction targets.Well-designed plans need public support to be implemented.Because of historical prioritization of motor travel, change to infrastructure is required in most industrialized urban areas (Brand, Dons, et al., 2021).The instinctive response to change, however, is resistance (Samuelson & Zeckhauser, 1988).This instinct may be further exacerbated by the dominance of automobility (Sheller & Urry, 2000;Walks, 2015).Opposition toward bicycle lanes in particular can lead to project abandonment and the removal of new lanes (Wild et al., 2018).Behavioral science can help to understand why people oppose new initiatives.For example, there is some evidence, reviewed in more detail below, that justifications often cited by opponents of active travel infrastructure-such as negative effects on local businesses or increases in traffic congestion-are misperceptions that are unsupported by empirical evaluations of such schemes (Cairns et al., 2014;Volker & Handy, 2021).By integrating behavioral science into infrastructure plans, more nuanced strategies can be developed that help to correct misperceptions, supporting the shift toward sustainable travel choices and reducing the carbon footprint associated with car travel (Sharma & Jain, 2023).This paper reviews existing evidence on (1) how to design and (2) how to implement active travel infrastructure (hereafter "ATI") from the perspective of behavioral science, with a view to assist policymakers and academics interested in implementing well-designed ATI.We consider both aspects of ATI policy together, because neither alone will be sufficient to match the scale of necessary modal shift.
Our review is narrative in nature, such that it gathers and analyses evidence relating to a broad scope of research questions relevant to policymakers interested in implementing ATI.Systematic reviews are typically favored for their methodological rigor, which has many obvious strengths (Booth et al., 2021).However, these strengths restrict the kinds of questions that can be answered to those that can be highly specified-for example, what effect do physical barriers on cycle lanes have on crashes?Narrative reviews, on the other hand, are better suited to providing perspective on problems where the scope is broad; they allow for more comprehensive coverage of evidence for those who need to balance multiple considerations (Collins & Fauser, 2005;Rother, 2007).When planning and implementing active travel initiatives, policymakers must integrate evidence that relates to multiple research questions.The perspective we take is one of a team of behavioral scientists who work in close collaboration with policymakers.The issues addressed in this review were informed by discussions with policymakers at a national transport authority and at a local council authority.As such, the review centers around two of the main challenges faced by those working at the coalface of ATI implementation: how to design infrastructure that encourages behavior change and how to secure public support for infrastructural changes.
As the review is narrative, we employed a flexible approach to identify relevant papers.Keywords and phrases (e.g., "cyclist" AND "perception" AND "infrastructure"; "opposition" AND "active travel"; "bikelash," etc.) were chosen to reflect interest in both design and implementation of active travel infrastructure and searched on Scopus, Web of Science and Google Scholar.We restricted our search to publications in English but did not impose restrictions on publication year or geographic location.We also conducted backward citation search of references in relevant articles and forward search of papers that cited articles we judged of most importance.We included gray literature reports and working papers in our search and when initially drafting the review, but non-peer-reviewed papers are excluded here.The bulk of the papers we cite were located through this process, but others, particularly those related to psychological tendencies more generally, were included based on our expertise as an applied behavioral science team; their relevance to ATI had not been highlighted in the literature we sourced.Throughout the search process, emphasis was placed on maintaining a balance between breadth and depth.The aim was not to exhaustively gather every available piece of literature, as in a systematic review, but rather to identify seminal works, key debates, and emerging themes that would contribute to a comprehensive yet focused narrative synthesis.Hence, we reference individual academic studies, reviews and meta-reviews, spanning topics from the physical construction of cycle lanes to the psychology of opinion formation.The review is hence intended to be used as a guide for designing ATI that is successful in promoting active travel, and implementing it in a way that avoids misperceptions of its effects, given current evidence offered by behavioral science.
At the outset, it is important to understand that our aim is not to suggest that behavioral science should be used to persuade people to support schemes that they may legitimately oppose.Democratic authorities act on behalf of the public they represent and consultation with local communities is vital (Willis et al., 2022).Importantly, individuals and communities have a right to object to and challenge schemes that they fear will harm their neighborhoods.Nevertheless, the existing literature in behavioral science that we describe shows that opinion formation can be subject to strong biases and misperceptions.Our aim, therefore, is to elucidate these effects in the context of active travel so that policymakers and stakeholders can recognize them, perhaps counter them and, in general, promote informed decisionmaking about active travel initiatives.
We note also that this review is not intended as a definitive "how-to" for designing infrastructure, which naturally relies on engineering and urban planning expertise.Instead, the review may help relevant stakeholders prioritize design features that have implications for behavior.Moreover, while active travel relates to all forms of purposeful travel in which the individual exerts physical effort in order to move, we focus on increasing rates of cycling.This focus is mainly the result of greater attention paid to cycling relative to walking (and other forms of active travel) in the international literature to date (Cook et al., 2022).
The remainder of the review is structured as follows.Section 2 outlines evidence concerning behavior change following the implementation of ATI and covers environmental, psychological, and sociodemographic predictors of infrastructure use.Section 3 explores design elements that encourage usage, in light of findings from Section 2. Section 4 tackles the problem of public opposition to infrastructure change, with a focus on identifying sources of potential misperception and bias.This section synthesizes theory and evidence from behavioral economics and psychology.In Section 5, we discuss the implications of the available evidence and avenues for future research.

| USE OF ACTIVE TRAVEL INFRASTRUCTURE
The primary question for those seeking to implement ATI is inherently behavioral: does changing infrastructure lead to a change in behavior?In this section we review relevant research on predictors of use of ATI.In doing so, we present multiple behavioral factors that policymakers, urban planners, architects, engineers and others working at the coalface of implementation would want to consider if their aim is to increase use of ATI.We first summarize the evidence on aggregate effects of infrastructural change (i.e., whether cycling rates in a community change overall once infrastructure is implemented).We then focus on the environmental and individual-level characteristics that are associated with stronger effects on uptake.These predictors can be thought of as variables associated with greater behavior change.

| Active travel infrastructure and behavior change
Although the success of any individual initiative is not guaranteed, there is strong evidence that changes to active travel infrastructure, in general, achieve their aim of increasing rates of active travel.The highest quality evidence is produced by longitudinal assessments of travel behavior before and after schemes are implemented, with changes contrasted against measures taken at the same timepoints in comparable areas (e.g., nearby towns of a similar size with similar levels of active travel prior to scheme implementation).Individual studies using this "pre-post quasi-experimental" (hereafter, PPQE) approach have recorded increases in the number of people cycling of about 30% in multiple countries, although there is considerable heterogeneity in effect sizes (Aldred et al., 2019;Heinen et al., 2015;Song et al., 2017).For example, an evaluation of the "Cycle City Ambition Programme" in the United Kingdom showed increases in cycling levels of between 14% and 40% in five locations that implemented cycling infrastructure relative to comparable control areas.
In addition to individual studies, we identified six systematic reviews on the effects of changes to ATI.Each of the reviews supports the broad conclusion that infrastructural change leads to increases in cycling (Buehler & Dill, 2016;Pucher et al., 2010;Smith et al., 2017;Stappers et al., 2018;Yang et al., 2010).For example, Stappers et al. (2018) identified 19 studies that used a PPQE design to test the effect of infrastructural change on active travel and transport-related physical activity.The review showed strong effects of built environment changes on rates of cycling, particularly among individuals living close to the new infrastructure.However, the review also highlighted the need to evaluate such initiatives using robust methods, as many studies were classified as having moderate-to-high risk of bias and as a result effect sizes might be over-estimated.

| Environmental predictors of use
ATI in general is successful in promoting active forms of transportation, but certain environmental factors are especially important for use.
One such factor is individual proximity to ATI, with evidence also from meta-level reviews (Brüchert et al., 2022;Fraser & Lock, 2011;Götschi et al., 2017).PPQE evaluation of the "mini-Holland" program in London (n = 1712) suggested a potential "dose-response" to ATI: individuals living in neighborhoods with higher exposure to different schemes showed a greater propensity toward active travel (Aldred et al., 2019).Relatedly, other PPQE research in the United Kingdom (n = 470) showed that boosting proximity to public transport can increase the share of commutes that involve any form of active travel, with corresponding declines in the share of trips made entirely by private vehicle (Heinen et al., 2015).
Aside from proximity, there are other environmental factors that can affect active travel.There is some evidence that steep inclines have negative effects on infrastructure use (Fraser & Lock, 2011), although this effect may be weakened by the growing availability of e-bikes (Jenkins et al., 2022).There is some evidence from Ireland that fewer people engage in active commuting during poor weather (Deenihan et al., 2013).Broadly speaking, however, weather effects tend to be small in temperate climates (Böcker et al., 2013;Clark et al., 2014;Saneinejad et al., 2012).We could find no research systematically investigating whether beliefs about weather effects act as barrier to ATI use among the public; beliefs could play a greater role in deterring active travel than the weather itself.

| Psychological predictors of use
Turning to individual-level predictors of usage, perception of safety is perhaps the most widely cited (Fishman, 2016;Koh & Wong, 2013;Krizek et al., 2009).A review of cycling infrastructure in six cities in the United Kingdom and the Netherlands concluded that safety perceptions are the deciding factor for whether potential cyclists choose to cycle (Hull & O'Holleran, 2014).Perceptions of safety and fear of traffic is especially important in communities with an "emerging" population of new cyclists (Chataway et al., 2014).
It is also important to note that the literature points to safety perceptions rather than objective safety indicators, such as accident statistics.Near-misses and conflict with motorists are deterrents to many cyclists, particularly those with lower levels of confidence, and are unlikely to be recorded in official statistics (e.g., Abadi & Hurwitz, 2018;Branion-Calles et al., 2017).However, this leads to difficulties in defining what constitutes "safe" infrastructure (Götschi et al., 2016).While priority lights for cyclists at intersections, limited discontinuities and traffic-slowing policies all make subjectively safer infrastructure (Ferrari et al., 2020;Hull & O'Holleran, 2014), there are infrastructural considerations that can help perceptions of safety that may not necessarily be linked to reductions in crash statistics.We discuss empirical evidence for design features that influence objective safety and perceptions of safety in Section 3.
A related psychological concept is an individual's self-efficacy, defined as their belief in their own capabilities in specific domains (Bandura et al., 1999).A longitudinal survey of over 3000 individuals in the UK assessed psychological predictors of change in active travel over a 1-and 2-year period.Results showed that higher baseline levels of active travel-related self-efficacy predicted greater amounts of cycling for transport and recreation up to 2 years later (Bird et al., 2018).Similar results were observed in Australia (Badland et al., 2013).Importantly, it is possible to increase an individual's self-efficacy, for example by prompting recall of occasions with high personal control of travel choice.Interventions targeted at boosting active travel self-efficacy have been shown to increase objectively-measured active travel behaviors when tested via randomized controlled trials (Bélanger-Gravel & Janezic, 2021).To the best of our knowledge, however, the interaction between self-efficacy interventions and measures to improve safety perceptions has not yet been tested.Relative effect sizes in the literature suggest it is better to prioritize safety perceptions.

| Sociodemographic predictors of use
It is well-established that gender and age predict active travel, with men and younger adults far more likely to cycle in most countries (Moudon et al., 2005;Pucher et al., 2011;Tilahun et al., 2007).However, large-scale analyses of national travel surveys from multiple countries show that the relationship between gender and cycling exists only in low-cycling societies (Buehler & Goel, 2022).Communities where cycling has greater than a 7% share of travel modes tend to reach parity between men and women cyclists (and sometimes a gender bias in favor of women).Moreover, while older people are underrepresented in all jurisdictions, they have relatively better representation in high-cycling societies (Goel et al., 2022).Hence increasing rates of cycling among women and older adults is not an impossibility but may depend on design considerations.
Analyses of large-scale travel survey data across Asia and Europe have shown that older adults are especially sensitive to features of the built environment, such as connectivity of the infrastructure-typically measured by the density of intersections and scarcity of "discontinuities" (Brüchert et al., 2022;Cheng et al., 2019;Portegijs et al., 2020;Yang, Sasaki, Cheng, & Liu, 2022;Yang, Sasaki, Cheng, & Tao, 2022).Importantly, these analyses consider within-city differences in active travel and show that better facilities are associated with higher rates of older people cycling.Since cultural differences within cities are smaller than between countries, the results suggest that providing accessible and safe ATI to older individuals is likely to boost active travel.Recent evidence further suggests that older adults in particular may benefit from e-bikes (Jenkins et al., 2022).
Although young people are typically more likely to cycle than older adults, cycling rates among children and adolescents have been declining for decades (Adepoyibi et al., 2022;Schmassmann et al., 2023).Reviews of active travel among youth suggest that increasing rates of car ownership partly explains this trend (e.g., Pont et al., 2009).However, similar to older adults, urban planning decisions and ATI also play a role.Across multiple countries, children and young people are less likely cycle where traffic is heavy, speed limits are higher and intersections are not signalized (Ding et al., 2011;Mitra & Buliung, 2012;Panter et al., 2008).Perceived safety matters for youth too, although perhaps more important are perceptions held by parents and guardians who are often gatekeepers of youth travel mode choices (Giles-Corti et al., 2009;Mandic et al., 2020;Ross & Wilson, 2021).A review of over 20 studies shows that negative perceptions of built environment features (e.g., street connectivity) and traffic safety are the primary parental barriers to youth cycling (Aranda-Balboa et al., 2020).
The literature on the cycling gender gap is less conclusive, although there is good evidence that safety considerations are especially salient for women.One analysis of 10 million journeys on a bike-share scheme in London showed that women preferentially select routes with slower traffic and those that are offset from major roads (Beecham et al., 2014;Mitra & Nash, 2017).The primary issue with the available evidence is that, while gender gaps have been identified, there has been little consideration of gendered processes underlying the differences, such as social norms (Gorrini et al., 2021) or the greater complexity of travel patterns required by women due to their typically larger share of household responsibilities (Ravensbergen et al., 2019).One analysis of travel patterns in New Zealand showed that women indeed tend to use more diverse routes (Shaw et al., 2020), potentially reflecting women's greater propensity to engage in multi-purpose trips (Craig & van Tienoven, 2019).Census analysis of cycle frequency in Dublin suggests that women are more sensitive to proximity effects of cycle lanes compared to men (Carroll et al., 2020).Hence, improving safety, ensuring different destinations can be reached and providing appropriate facilities (e.g., cargo e-bikes) may be especially important for women, although more research is needed on potential cultural influences (e.g., Egan & Hackett, 2022).
Despite the potential for larger marginal benefits for women and older people of behaviorally-informed designs, the evidence for who benefits from changes to ATI is mixed.One reason is that many infrastructural changes aim to facilitate active commuting, which may favor men and younger people.UK analyses suggest that cities with growing rates of cycling do not necessarily see greater diversity among cyclists following improved infrastructure (Aldred et al., 2016).Research in New Zealand showed those who already cycled simply did so more, rather than changing the behavior of those who did not already engage in active travel (Keall et al., 2018).One systematic review claimed that infrastructure changes benefit socio-economically advantaged groups, although the authors acknowledge most of the individual studies are of low quality (Smith et al., 2017).By contrast, better quality PPQE evidence from New Zealand suggests the opposite, that groups least likely to travel by active modes benefit more (Keall et al., 2022; see also Hansmann et al., 2022).This variability in the literature with respect to who benefits from ATI may be attributable to differences in design focus; it is unclear from the available evidence what design elements and safety aspects were considered within studies that have sought to analyze uptake among disadvantaged groups.

| Summary
On balance, the evidence suggests that investment in ATIs lead to behavior change, but the degree of change depends on multiple additional variables.The empirical literature reviewed in this section highlights the importance of infrastructure design on uptake, particularly among groups less inclined to cycle.Based on the above evidence, the best way to increase uptake is to design routes that are easily accessible from places of residence.Built environment components (e.g., such as perceived safety at intersections) can also influence how safe active travelers feel on their journey.Improving perceived safety is likely to increase uptake by all sociodemographic groups, but especially women, older adults and children.We examine the literature on design features in more detail in the next section.

| LESSONS FOR DEVELOPMENT AND DESIGN
The predictors of uptake described in the previous section may not always be amenable to intervention (e.g., the sociodemographic profile of the neighborhood, natural barriers to proximity).However, there is strong evidence that flexible design decisions influence the degree of behavior change that can be expected from ATIs.This section analyses the available evidence on ways to design ATI to improve perceived and objective safety, as well as network design considerations to promote uptake.

| Safety
In general, ATI reduces the incidence of acute injury among cyclists (Schepers et al., 2015).For example, an expansion of the bike network in the United States-which is more dangerous for cyclists compared to most European countries-was associated with a fall of between 25 and 75 per 100,000 trips in crashes, severe injuries, and fatalities (Pucher & Buehler, 2016).Similarly, retrospective studies have shown an association between increased bike lane milage and decreased cycling-related trauma admissions (Goerke et al., 2020).Separated bike lanes are safer than mixed traffic, particularly as they remove the possibility of car-door crashes (Nanayakkara et al., 2022), as well as reducing vehicle-cycle collisions (Ling et al., 2020).
A systematic review of the effects of cycle tracks on objective safety measures concluded that cycle tracks boost safety, but one-way cycle lanes tend to be safer than two-way ones (Thomas & DeRobertis, 2013).Where it is not feasible to implement one-way tracks, intersections can be designed to improve cyclist safety.Some of the most effective ways to improve intersection safety include using advance stop lines, providing dedicated priority lights for cyclists, raising cyclist crossings to the pavement level and traffic calming measures such as speed limits of 30 km/h (Pucher & Buehler, 2016;Thomas & DeRobertis, 2013).In general, policies that incentivize cycling and disincentivize driving, for example by providing deliberately circuitous routing for cars and direct routing for cyclists, improve cycling safety (Thomas & DeRobertis, 2013;Xiao et al., 2022).
Assessing ways to design infrastructure to reduce crashes is relatively straightforward given the availability of statistics for most countries.However, these design features may not be sufficient to increase uptake, as Section 2 outlined how perceptions of safety determine whether most people choose to engage with ATI, particularly when cycling.The following sections outline the available evidence on safety perceptions.

| Segregated lanes
For confident cyclists, the primary decision factor when choosing to cycle and which route to take is estimated travel time (Caulfield et al., 2012;Gutiérrez et al., 2020).For others, however, the nature of the infrastructure on their journey makes the largest difference.Cyclists in multiple countries show a strong preference for cycle lanes that are segregated from other forms of traffic.A common design choice is to segregate bike lanes from car traffic by combining cycle and bus lanes, which may actually decrease rates of cycling (Echiburú et al., 2021).The relevant body of evidence is substantial, including qualitative and quantitative methods, reviews and meta-reviews (Berghoefer & Vollrath, 2022;Buehler & Dill, 2016;Götschi et al., 2017;Rossetti et al., 2019).Safety perception-boosting segregation can be achieved by building ATI away from roads or by installing buffers or planters between cycle lanes and traffic (Knight & Charlton, 2022;M arquez et al., 2021).Planters may be especially beneficial for cyclists due to incidental green exposure, although visibility issues need to be considered (Beery et al., 2017).
Safety perceptions are typically elicited using experimental surveys, where participants are asked to make judgments about images or videos of cycle lanes that have been digitally altered to test different designs.Findings from these types of studies are supported by other studies that have focused instead on biological indicators of perceived safety, by recording biological stress responses while people cycle on roads or in simulators.These studies also show that stress responses are significantly lower on segregated cycle lanes and higher on busy roads, at intersections, on cobbled surfaces and in noisier places (Cobb et al., 2021;Teixeira et al., 2020).

| Lane color
A relatively easy infrastructural change with implications for perceived safety is to paint cycle lanes (M arquez et al., 2021).There is strong evidence in favor of painted lanes from a PPQE, multimethod analysis of the effect of painting an existing cycle lane red in Norway (Fyhri et al., 2021).The study included video analysis of cyclist and motorist behavior as well as GPS analysis of almost 2500 journeys and a survey of over 1500 residents.After painting, cyclists were more likely to cycle in the lane than on the pavement which, coupled with survey results, implied greater perceived safety.GPS data showed that the first streets with painted cycle lanes saw significant increases in cycling.The painted cycle lanes also affected motorists' behavior.Following painting, they were less likely to park in the cycle lane and routinely kept a greater distance from the lane when driving (Fyhri et al., 2021).
Evidence on whether the color itself matters is inconclusive (Vera-Villarroel et al., 2016).A web survey of 560 cyclists and motorists in Norway identified preferences for colored over uncolored bicycle lanes, but whether red or green lanes were favored depended on familiarity: participants living in areas with existing red lanes preferred those lanes, whereas participants living in areas without colored lanes preferred green ones (Karlsen & Fyhri, 2020).Some studies have also sought to evaluate the effect of transverse lines on cycle lanes as a way to reduce cyclist speed at specified points.One PPQE study in Sweden showed limited results, with a small effect observed only on recreational cyclists (Kovaceva et al., 2022).

| Combining safety features
Most studies on safety perceptions tend to focus on effects of removing or adding one specific feature, but the interaction effect from combining multiple safety features is also important: what is the marginal benefit of painting a cycle lane if it is already segregated?To answer questions like this, one study in Germany asked over 21,000 participants to judge multiple digitally enhanced road and street configurations, which varied along multiple factors: path width, segregation, boundaries, the presence of parked cars, lane color, busyness, type of road (Gössling & McRae, 2022).The study also varied the participant's perspective, by showing images from the perspective of motorists, cyclists or pedestrians (assigned based on the participant's typical mode of travel).Supporting the findings above, separation from other forms of traffic was judged to be the most important safety feature, particularly when achieved using clearly demarcated boundaries.Path width emerged as the second most important feature.Lane coloring had larger effects on safety perceptions where other features were missing rather than present.In other words, where physical separation is not possible, coloring the lane becomes more important for maintaining perceived safety.Wide, painted lanes with physical boundaries away from parked cars and commercial activity received almost universal safety ratings (98%), regardless of whether they were on a major road, alongside pavement or on a side street.The setup of the experimental survey allowed for the marginal effects of individual factors to be estimated.For example, even with physical boundaries, safety ratings dropped to 78% if the lane narrowed or if the lane ran alongside parked cars.Removing the boundary led to a further drop to 50%.Importantly, cyclists, motorists and pedestrians were in agreement about what constituted safe travel infrastructure, including the presence of parked cars.

| Connectivity
Better connectivity-typically measured by the density of intersections and scarcity of "discontinuities" (i.e., breaks in cycle lanes where stretches of road have no cycle lane)-is one of the strongest predictors of the success of ATI (Buehler & Dill, 2016;Giles-Corti et al., 2016;Stinson & Bhat, 2003).The evidence for this relationship derives from systematic reviews as well as meta-level reviews of systematic reviews (Götschi et al., 2017;Kärmeniemi et al., 2018;Panter et al., 2019).The general idea is that being able to cycle full journeys through multiple routes correlates positively with active travel frequency, controlling for other sociodemographic and environmental factors (Verduzco Torres et al., 2022;Winters et al., 2017).Modeling studies suggest that the benefits of network infrastructure are observed across gender, age and income groups (Standen et al., 2021).Another possibility is that increased density of intersections allows cyclists to find shorter routes for most trips, increasing the appeal of traveling by bike.Example studies include a longitudinal survey from Australia (n = 909) showing that objective measures of street connectivity strongly correlated with selfreported use of cycling infrastructure for both recreation and utility purposes (e.g., commuting) (Badland et al., 2013).A cross-sectional survey of over 2000 residents in small towns in Germany found street connectivity to the be the strongest environmental predictor of self-reported cycling frequency (Brüchert et al., 2022).Similarly, a survey of over 8000 people across 8 countries in Latin America showed that better connectivity predicted higher frequencies of recreational cycling (Ferrari et al., 2020).Hence, active travel routes that minimize discontinuities in pathways and cycle lanes should be prioritized when designing new infrastructure.The effect of connectivity may be strengthened by facilitating shortcuts between popular locations for active travelers that are not accessible to motorists (Piatkowski et al., 2019).
However, the relationship between connectivity of the infrastructure and perceived safety is important to note.While a higher density of intersections may lead to fewer high-traffic streets, where streets do have high volumes of traffic or where specific intersections are perceived as unsafe, cyclists often engage in detours to avoid those intersections or crossings (Gössling et al., 2019;Van Cauwenberg et al., 2018).Hence safety perceptions have the potential to undermine positive effects of connectivity, unless intersections are designed in ways that cyclists perceive as safe.

| End-of-journey facilities
A systematic review of 39 interventions to increase cycling suggested that adding facilities, such as secure parking facilities, can have relatively large effects on use of cycle lanes (Do gru et al., 2021; see also Egan et al., 2022;Hamre & Buehler, 2014;Hunt & Abraham, 2007).Other reviews have suggested that secure parking facilities can help to address the gender and age gap in cycling (Goel et al., 2022;Tilahun et al., 2007).For commuters, shower availability at work can encourage active travel (Hamre & Buehler, 2014;Tilahun et al., 2007).

| Summary
Marginal gains in active travel are likely to be enhanced through behaviorally-informed design.The evidence above provides reasonably strong support for the importance of segregating wide, painted cycle lanes from traffic using physical boundaries.Importantly, the literature on reducing crashes focuses on intersection features, whereas the literature on safety perceptions tends to focus on stretches of road.Hence these safety features should be combined to maximize real and perceived safety.
One limitation of the research reviewed in this section is that it generally considers the impact of infrastructural change independently of how visible the changes are to the target users.It is essentially unclear how cognizant potential cyclists are of features installed to improve safety.This is important, because the wider behavioral science literature suggests that drawing attention to safety features and improvements is likely to further encourage uptake.However, this idea would benefit from robust evaluation, ideally through a PPQE design.
Aside from designs to improve safety, ATI clearly benefits from careful consideration of how the infrastructure functions as a network.Designs should allow for full journeys to be made safely by foot or bike between popular areas, with limited breaks in infrastructure.End-of-journey facilities such as bike parking near public transport or shower facilities at work places are both useful to promote use.

| LESSONS FOR IMPLEMENTATION
Behavioral science not only offers lessons for designing ATI in ways to maximize use, but to also inform how to implement schemes in ways that avoid biased reactions.In this section, we outline relevant behavioral science research on public responses to potential infrastructural change, informed by the psychology of opinion formation and research on climate policy support.

| Public response
Although the general public in many countries report high levels of support for climate change mitigation and ATI more specifically, initiatives often face unexpectedly high levels of opposition, or "bikelash," during planning and consultation stages (Aldred et al., 2019;Timmons & Lunn, 2022;Wild et al., 2018).Negative reactance to infrastructure changes that seek to accommodate cyclists may stem from negative public opinion of cyclists more generally.Some scholars have argued that being a "cyclist" is a stigmatized social identity, with characteristics that manifest themselves in ways similar to other minority groups (Aldred, 2013;Prati et al., 2017).
While motorist opinion of cyclists has been shown to correlate with aggressive behavior (e.g., Fruhen et al., 2019), it has not yet emerged as a predictor of wider resistance toward changes to infrastructure.Indeed, there is a relatively small body of literature that has explicitly investigated this phenomenon.The majority are case studies conducted in areas where active travel schemes either have been implemented or are at the proposal stage.Some studies have sought to understand public opposition through interviews and surveys with small samples of opponents, planners, and stakeholders (Field et al., 2018;Lambe et al., 2017;Parajuli & Pojani, 2018;Vreugdenhil & Williams, 2013).Most studies are conducted after the implementation of a scheme and use interviews (Field et al., 2018;Melia & Shergold, 2018) or field surveys as the research method (Aldred & Croft, 2019;Castillo-Manzano et al., 2014;Larson et al., 2016;Noland et al., 2022).Some studies have recorded public expectations before implementation and attitudes after implementation, again through interviews (Crane et al., 2016;Lambe et al., 2017) or field surveys (Melia & Shergold, 2018).
Hence, the above research predominately takes the form of interviews with a relatively small number of participants and the strength of the relationship between expectations and support has been inferred qualitatively.While such approaches can provide rich, scheme-specific data, they do not permit general inferences that are likely to be widely applicable.However, we have located a few quantitative studies that have investigated public opinion of active travel schemes generally and individual predictors of support (Cradock et al., 2018;Gase et al., 2015;Gustat et al., 2014;Rissel et al., 2018;Semple & Fountas, 2022).As structure for the following sections, we use the framework of considering climate change as a complex collective action problem and consider factors that drive cooperation in such problems: individuals' perceptions of cooperation effectiveness, fairness and self-interest (Hormio, 2023).We end with discussing how biases can play a role in opinion formation and how to counter them.

| Perceptions of effectiveness
Both qualitative and quantitative studies on public opinion of active travel initiatives highlight expected outcomes as drivers of support.While we have found no evidence that people doubt whether active travel is effective in reducing emissions, other research suggests that the public in multiple countries find it difficult to estimate the relative effects on emissions of different pro-environmental behaviors, often underestimating the impact of more effective actions (Cologna et al., 2022;Timmons & Lunn, 2022;Wynes & Nicholas, 2017).Hence, while the public may not doubt that active travel has environmental and health benefits, they may underestimate the scale of these benefits.
Underestimation of the environmental and health benefits of active travel initiatives may be compounded by the frequently-cited expectation of negative economic effects.Qualitative interviews from multiple countries before and after the implementation of active travel initiatives show that retail traders strongly associate the presence of car parking spaces with turnover (Crane et al., 2016;Ferster et al., 2021;Field et al., 2018;Lambe et al., 2017;Wild et al., 2018).Such concerns regarding retail turnover do not align with the empirical evidence.For example, while car drivers may spend more per trip, cyclists tend to spend more frequently, which can lead to higher accumulated spend over time (Lambe et al., 2017;Volker & Handy, 2021).
Reported opposition from residents cites concerns about the local economy too, but also often highlights concerns around traffic safety (Vreugdenhil & Williams, 2013;Wild et al., 2018) and expected increases in traffic congestion (Crane et al., 2016;Field et al., 2018).As discussed in Section 3.1, concerns regarding safety does not in general align with available evidence, as cycling infrastructure is in large safer for users.
That people may hold inaccurate expectations is unsurprising, as there is strong evidence that people are often poor at predicting the consequences of both large and small events-including how they will feel (Wilson & Gilbert, 2005).Inaccurate projections of future feelings can explain why people do not make behavior changes that would improve their well-being.For example, people underestimate the boost in happiness from spending more time in nature (Nisbet & Zelenski, 2011).Attitudes for climate policies have also been shown to become more positive after schemes are introduced and people witness the actual change (Hansla et al., 2017).Survey evidence suggests that public underestimation of the effectiveness of climate policies helps to explain some opposition to those policies (Nilsson et al., 2016;Schuitema et al., 2010).
People can also be poor at predicting the future trajectory of numerical data.There is a "conservative bias" in many predictions, where people become anchored to the starting point and do not adequately alter this value to reflect change (Stango & Zinman, 2009).This phenomenon can help explain why people underestimate the effect of traffic disincentives at reducing car numbers (Schuitema et al., 2010).
Another reason an individual may discount the evidence in favor of active travel schemes is lack of trust in the implementing body to carry out the work with sufficient competence.Trust in the implementing institution is positively related to policy acceptance across a range of climate policies (Bergquist et al., 2022) and there is some evidence that low trust in government can translate to low perceived effectiveness of transport policy (Huber & Wicki, 2021).However, most of the research on trust and climate policy is in the context of carbon taxes.It generally finds that much of the opposition to carbon taxes is due to distrust of politicians (Fairbrother, 2019).People suspect government of using carbon taxation as a way of raising public revenue generally, rather than as a way to lower emissions.Research has implicated low levels of trust as the reason why the public prefer ring-fencing carbon tax revenues for green investment projects, even though from an optimal taxation viewpoint this approach is not ideal (Carattini et al., 2018).

| Perceptions of fairness
Opponents of bike lanes often report feeling that they had no voice or control over the changes being made (Crane et al., 2016) and complain of a lack of meaningful consultation (Field et al., 2018).These views are reiterated from retailer evaluations of active travel schemes in Ireland: feeling that one's concerns are being ignored is cited as a source of opposition and distrust (Lambe et al., 2017).
An important point to note is that, when judgments of fairness are made post-hoc, they may be influenced by the outcome rather than an objective evaluation of the process, known as the "outcome bias" or "hindsight bias."The results of courses of action color peoples' judgments, including ethical or moral evaluations (Byrne & Timmons, 2018;Fleischhut et al., 2017;Roese & Vohs, 2012).Outcome biases are difficult to overcome, but one avenue potentially worth exploring is to crowd-source ideal consultation processes using similar methods to those used for public consultations on active travel implementation itself.Communicating that the consultation process was agreed prior to any decisions on infrastructure may help (Perlaviciute, 2022).
Other important aspect of fairness is whether the distribution of costs of the policy is perceived as fair (Muhammad et al., 2021).In a meta-analysis on public opinion about climate change laws and policies, perceived distributional fairness was one of the strongest predictors of support (Bergquist et al., 2022).Most of the controversy around active travel initiatives concern incentives that relocate and transform spaces reserved for vehicles to bikes and pedestrians (Wilson & Mitra, 2020).It is not evident whether these initiatives are viewed as resulting in an unfair cost for drivers in favor of pedestrians and cyclists.Moreover, many salient objections to changes to car access relate to accessibility for those less able-bodied, which warrant consideration at design and implementation stages.It is thus important to assess perceived and expected costs and benefits for different population subgroups, especially those with low mobility.Moreover, the potential for e-bikes to overcome mobility barriers may warrant subsidies (Jenkins et al., 2022).

| Self-interest
Implementing ATI typically requires reappropriating existing car infrastructure, either through the removal of lanes or car parking spaces.Hence existing motorists inevitably experience some losses, while current and prospective cyclists stand to experience gains.Interventions that combine disincentives to driving may be more effective than those that simply encourage cycling (Xiao et al., 2022) but naturally such disincentives may be difficult to implement with public support.Some studies have indeed shown a relationship between regular active travel and stronger support for active travel and pedestrianization investments (Gase et al., 2015;Semple & Fountas, 2022).However, this relationship is not guaranteed.Other studies have found support for the introduction of bike lanes to be similarly as high or higher among motorists as cyclists (Cradock et al., 2018;Rissel et al., 2018).This finding may be linked to preference among some motorists for greater safety precautions for cyclists (Gössling & McRae, 2022), or it could indicate a subset of motorists who would prefer to cycle if provided with appropriate infrastructure.However, regular travel by car is typically associated with reduced support for removing car parking spaces or restrictions on car travel to facilitate active travel (Rissel et al., 2018).

| Biases in reasoning
Retailers and residents often hold legitimate concerns that should be considered and addressed during consultation phases.However, evidence from behavioral science highlights how misperceptions can form and lead to biases in reasoning.
The discrepancy between public opposition toward active travel plans and the acceptance of such infrastructure after implementation (Ferster et al., 2021) points to perhaps the largest psychological barrier to public support for ATI: status quo bias.Status quo bias is the tendency for individuals to prefer things to stay as they are, even if change may be beneficial (Samuelson & Zeckhauser, 1988).Status quo effects can be large.For instance, Lang et al. (2021) investigated the role of status quo bias in preferences to join a state-level climate mitigation scheme.When the status quo was to be outside the scheme, the average "willingness to pay" to join was $170, but when the status quo was to already be part of the scheme, maintaining membership was valued at $420.
"Not-in-my-backyard" syndrome (NIMBY; Dear, 1992) is often cited as a barrier to implementing climate policies at a local level and can be attributed to the status quo bias.Others argue that the concept of NIMBYism is too simplistic and opposition should instead consider the strong emotional attachment between residents and their locality (Anton & Lawrence, 2016;Clarke et al., 2018;Devine-Wright, 2009;Sebastien et al., 2019).However, status quo bias may also drive this association between place attachment and dislike of change.
Dissipation of status quo bias has been implicated as a reason for the increased support for congestion charges postimplementation (Börjesson et al., 2016).In other words, change is initially opposed because it is a change, and then supported because it has been made, regardless of the respective costs and benefits.Similar effects are observed following the implementation of bike lanes (Ferster et al., 2021).After active travel schemes in Scotland were in place, retailers mentioned concerns about revenue decreasing due to limited parking prior to implementation of bike lanes, but not after (Crane et al., 2016).
Status quo bias can be difficult to overcome (Nightingale et al., 2022).One reassuring piece of evidence comes from the evaluation of active travel initiatives after implementation, which tend to be generally positive (Castillo-Manzano et al., 2014;Melia & Shergold, 2018;Noland et al., 2022).However, we could locate no studies that have identified the psychological mechanisms underpinning status quo bias with respect to ATI change and, relatedly, no studies that have identified ways to prevent it or tested interventions to mitigate it.This gap presents an opportunity for behavioral science to contribute to overcoming one of the primary barriers to policy implementation in developed countries.

| Preventing biased reasoning
Once initial opinions have formed, they are difficult to shift.People tend to interpret new information in ways that match their existing beliefs, to downplay information that contradicts these beliefs, and to seek out information that supports them.These effects are referred to as motivated reasoning and confirmation bias (Kunda, 1990;Nickerson, 1998).How opinions initially form is therefore important.The psychology of opinion formation suggests two important factors to overcome to reduce bias: primacy effects and messenger effects.
Initial information tends to be overweighted in attitude formation, particularly about personally relevant topics (Petty et al., 2001).This primacy effect is difficult to counteract.When trying to correct initial misperceptions, simply providing factual information tends to be ineffective and can even backfire (Lewandowsky et al., 2012;Walter & Tukachinsky, 2020).Hence, public consultations about active travel schemes should prioritize communicating accurate information on anticipated effects as early as possible, in order to "pre-bunk" against potential misperceptions (Maertens et al., 2020).
The messenger behind the information also makes a difference.Where people have difficulty evaluating the veracity of information, they often rely on mental shortcuts or heuristics rather than expending additional effort.One such heuristic is the messenger effect, whereby the source of information colors how people evaluate it.Familiar and trusted sources are, unsurprisingly, favored.This issue may be particularly relevant for active travel initiatives, as they are often viewed as the pet projects of faceless and unaccountable technocrats (Selmoune et al., 2020).One systematic review of factors that affect the successful implementation of infrastructural change highlighted the importance of influential individuals and their communication strategies with the public (Lawlor et al., 2022).Planners of such projects could thus benefit from collaboration and involvement of individuals or community groups to build public trust.
Another psychological tendency worth noting is that people often underestimate the level of support for progressive change in their communities, which can lead to self-silencing among advocates for change (Bursztyn et al., 2020).As an illustrative example, one study in the United States showed that people who underestimated other people's concern about climate change were less willing to discuss climate change with others.Correcting this misperception boosted willingness (Geiger & Swim, 2016).Similar instances of "pluralistic ignorance" have been observed in relation to other environmental beliefs and behavior (Drews et al., 2022;Ejelöv et al., 2022).Hence collecting data on privately-held support for ATI early in a consultation process may be important.

| DISCUSSION
Policymakers rarely have the luxury of waiting for perfect evidence, particularly when faced with problems as urgent as the climate crisis.When decisions need to be made, for example about how to implement ATI in a way that people will use, policymakers rely on the balance of evidence and intuition.It is role of researchers seeking to inform policy to present the former in an accessible way.Here we condense the above literature into implications for effectively designing ATI and to avoid losing public support to misperceptions, based on the balance of existing evidence.

| Effective design
The goal of infrastructure change is to change behavior.Multiple individual studies and systematic reviews show strong evidence that implementing ATI is likely to increase rates of active travel.Planning and design decisions determine the scale of effectiveness.There is strong evidence that connectivity, proximity and safety should be prioritized over other design elements.These considerations benefit all cyclists regardless of confidence, but are likely to have a greater marginal benefit among older adults, children and women.Schemes that make e-bikes and cargo bikes more readily available also hold potential to boost cycling rates within these groups.
Segregating cycle lanes from traffic is beneficial for both real and perceived safety and is favored by all road users.Real safety can be maximized by using one-way lanes, dedicated priority lights at intersections, advance stop lines, traffic calming measures and direct routes accessible only by cyclists.Perceived safety-which drives uptake-can be maximized using elements of the above, alongside physical barriers from traffic on wide, painted lanes.Taking this behaviorally-informed design approach can help increase the environmental and health benefits from ATI, as these combinations can not only increase active travel frequency but also decrease car dependency, particularly when implemented with policies that disincentivize driving.

| Public support
Design decisions can have large effects on the success of ATI, but public support for implementation is a necessary precondition; "bikelash" represents a significant barrier to the shift toward more sustainable communities that requires far more consideration from the behavioral sciences to understand underlying psychological mechanisms.While negative attitudes toward cyclists may play a role, the perceived effectiveness of active travel schemes remains a primary factor.People are more willing to support something that they think will work.This finding arises in a context where the overwhelming conclusion from international research into active travel initiatives is that they are, in fact, effective.The implications of this research suggest a need for continued efforts to communicate the supportive evidence and specific measured benefits of active travel schemes.Where initiatives are successful these can be used as demonstration projects in order to reduce uncertainty about effects of future plans.
Trust in public institutions and procedural and distributional fairness also matter for support.The conduct of public authorities that propose and undertake schemes is important for public trust.Presenting balanced views of evidence and taking multiple perspectives into account may be a more fruitful approach than expressing certainty in the righteousness of a proposed initiative.Similarly, investment in early, open and extensive consultation with community groups is likely to generate returns in the form of stronger support (Mar et al., 2023).
The relationship between self-interest and support for active travel initiations present a challenge for drawing broad implications.Instead, research undertaken within a locality is likely required, to assess perceptions in determining how people adopting a self-interested perspective are likely to view any proposals.However, even where community perceptions are measured and understood, expectations about likely impact are often inaccurate and attitudes are prone to change once ATI is actually in place.The evidence suggests that these changes can be substantial, as has proved to be the case in relation to congestion charging.This is tricky territory for policymakers, as asserting to the public that they will grow to like a scheme that they presently dislike may come across as paternalistic, or even deluded.A commitment to a trial period, where this is feasible, may be a useful solution.Stakeholders and planners would also benefit from gathering data on public perceptions and expectations of active travel plans in order to pre-bunk common misconceptions before they have cemented.This all takes place in a context where people are most strongly influenced by the information and arguments that they hear first and who they hear the arguments from.Once opinions have formed, they can be resistant to new information that challenges them.Early, clear communication from trusted sources is likely to be the best way to inoculate against misperceptions.
More generally, however, there is a need to understand and account for biases in the way information about active transport initiatives is processed.Status quo bias can be particularly influential for opinion formation and more research is needed on the role it plays in public response to transport infrastructure change.Another under-researched area is community perceptions of public support for such initiatives.People may assume that a majority opposes change, especially where that majority is quieter on the issue than a more vocal minority.Finally, people may view active travel schemes as a mere redistribution of resources between transport users and not "join the dots" to efforts to tackle climate change, which in general enjoy broad support.All of these forces potentially undermine support for change and has not received sufficient attention in the literature.

| CONCLUSION
ATI has large potential to decrease car dependency and promote healthier, more sustainable communities.The urgently required modal shift from motorized transport to cycling and walking in urban areas is unlikely to occur without it (Brand, Götschi, et al., 2021;Cuenot et al., 2012).Rapid infrastructure changes are hence essential for reversing the growth of transport-based emissions and mitigating climate change (Bleviss, 2021;Creutzig et al., 2018).However, planning and design decisions determine the scale of effectiveness.Achieving the goal of behavior change is made more likely by incorporating evidence from psychology and other behavioral sciences (Sharma & Jain, 2023) There is strong evidence that connectivity, proximity and safety should be prioritized over other design elements.Segregated, painted, one-way cycle lanes, dedicated priority lights at intersections, advance stop lines, traffic calming measures and direct routes accessible only by walkers and cyclists are ways to maximize perceived and objective safety-and in turn usage.
Planning of new active travel schemes needs to begin with early communication undertaken within an open and fair consultation (Lawlor et al., 2022).Ideally, messages should aim to challenge status quo bias and would aim to help local citizens to make up their minds about the benefits and disadvantages of change based on accurate perceptions and expectations.In particular, impacts on traffic, local businesses, and safety should be addressed.The international research that informs this conclusion is helpful in providing some principles to support this aim.However, more targeted research is needed to understand public opinion of ATI and how communities respond to change.Understanding the psychological mechanisms that underly opinion formation with respect to ATI specifically is, at present, a gap in the behavioral literature that hinders effective climate policy implementation.