Research

Download

In September 2018, the County Board in Arlington voted to launch a nine-month SMD demonstration project (pilot) intended to evaluate the community impacts of dockless electric-assist (e)-bikes and electric stand-up scooters, together referred to in the pilot program and throughout this report as shared mobility devices (SMDs). In June 2019, the pilot was extended for another six months until December 31^st 2019, to allow enough time for staff to complete the necessary evaluation and recommendations.

By doing so, Arlington County joined many other cities, such as Portland, Oregon and Santa Monica, California, in piloting SMDs in their respective jurisdictions. Results from pilot programs undertaken in other cities indicate the potential for SMDs to advance sustainability, promote equity, and increase accessibility and mobility. They also document potential challenges such as community complaints pertaining to sharing the right-of-way and safety.

This report provides the results of the evaluation of the nine-month Arlington County pilot program, including trends in deployment, utilization and feedback from the community to understand SMD adoption and system performance in the context of Arlington specifically. SMD performance was primarily evaluated against Arlington County’s transportation goals as documented in the Master Transportation Plan (MTP), pertaining to increased mobility, accessibility, equity, sustainability and efficient management of transportation options.

The analysis proceeded in three main steps, looking first at pilot operations (i.e. the supply-side), then service utilization (i.e. the demand-side) and finally the community’s reaction to the pilot for both SMD riders and non-riders.

The key takeaways of this evaluation report are threefold. First, deployment and utilization of SMDs in Arlington have increased over the duration of the pilot with a firmly positive response from riders in Arlington. This report supports evidence pointing to SMDs providing a viable complement to the County’s transportation ecosystem that increases mobility options and provides potential sustainability benefits. Second, certain aspects of the pilot have shown mixed results for the community, including the focus on equity concerns, with one measure being a disparity in deployment (normalized by residential population) between North and South Arlington), and the need for clearer communication of rules and regulations to the Arlington community. Finally, the third key takeaway is that there remain some challenges with the integration of SMDs in Arlington that will need to be addressed. This includes safety concerns from the standpoint of riders, pedestrians and drivers in Arlington, pointing to the need for more adequate infrastructure (e.g. protected bike lanes), and community concerns over parking and clutter on the sidewalk resulting from the program.

Based on the results, eight main recommendations were derived as follows:

I. Accelerate infrastructure investments to address rider and community safety and comfort concerns; focus on available route detail data Evaluate the possibility of increasing the share of protected bike lanes in key SMD corridors with the Rosslyn-Ballston corridor as a high priority given high ridership and elevated vehicle and pedestrian traffic. II. Continue working on innovative ways to address parking Communicate more stringent parking restrictions for operators – if addressable through technology – such as systematic restrictions by operators from parking at or near an intersection, outside residential or commercial entrances, in the middle of a sidewalk or near handicap parking space. Provide operators with map of desired deployment areas in each neighborhood and conversely of no-parking areas. Monitor and enforce operator response time in addressing parking complaints, where applicable. Examine further potential for SMD-specific parking infrastructure such as corrals or lock-to devices. III. Create, monitor, and refine equity expectations, go beyond geography Monitor and enforce as required proportional deployment in specific target areas. Perform more detailed equity and access analysis to ensure SMDs are being deployed in lower-income areas. Aim to assess equity from three standpoints (1) accessibility (in terms of location and the need for a smartphone to unlock the mobility service), (2) existence of equity programs, and (3) payment methods (e.g. needing a credit card). IV. Focus on and invest in communicating the rules and regulations to the public, including riders and non-riders Establish clear guidelines and messaging that is consistent across county resources and operator information platforms (websites, apps, and devices). Monitor operators’ messaging to ensure rules, regulations and rider resources are clearly communicated. Suggest or mandate creative ways in which operators can better communicate rules and regulations including, more innovative methods such as quizzes. Clearly state when rules are different from neighboring jurisdictions such as Washington, DC. Continue to conduct community outreach events, soliciting feedback and communicating how the county is addressing key community concerns flagged in this evaluation. Any such initiative should be done while balancing the importance of the convenience to riders so as not to negatively impact demand and the overall rider experience. V. Continue monitoring operations and requiring complete and robust data from operators SMDs are growing significantly while still at an early stage in terms of technology, best practices, and operational guidelines, making close monitoring a critical requirement for their continued operation. Require operators to comply with the data template and to submit additional operational data that they have not submitted yet (e.g. idle time, thefts and vandalization, broken SMDs, vehicle-specific trip and incident-level data). Require unified data (in terms of variables provided and format) from operators, allowing easier processing and cleaning of the data, which would leave more room for research and analysis. Monitor the difference between Washington DC and Arlington in terms of the service level (SMDs per 1,000 people). Monitor the difference between SMDs and Capital Bikeshare bikes in terms of the service level (SMDs per 1,000 people). Monitor incident rates such as broken SMDs and crashes with a specific focus on systematic or operator-specific patterns pointing to structural challenges. VI. Share results and county initiatives with the public, make the integration of SMDs into Arlington an inclusive and interactive conversation Share key SMD-related studies with the public including how the County is thinking about sidewalks, the rationale behind opening them up to SMD ridership and how it envisions the coexistence between SMDs, pedestrians and bikers. Inform the public on how the County is dealing with speed limitations without compromising on safety, including how speed limits, if applicable, are monitored and how operators are held accountable. Address perceptions of lack of safety, a key challenge to SMD popularity or even acknowledgment. This could include undertaking a specific study on SMD safety, exploring alternatives available (e.g. helmets, bike lanes) and misconceptions, and share findings with the public. With assistance from Arlington law enforcement, provide insight into the SMD enforcement process and potential deterrents for infractions. Share the results of this pilot evaluation as well as experiences from pilots in other cities to provide comparison and benchmarking, which are critical with early-stage technologies. VII. Collect or compile more robust data within and outside the SMD program and mandate periodic evaluation of SMD trends A broader array of data sources could enable more accurate analysis of SMDs in Arlington. This could include: Daytime population for Arlington using more granular measures than county-wide can help with a better comparison of (1) who actually is demanding SMDs at any point and (2) between areas that receive higher levels of commuters/workers (e.g. North vs. South Arlington). More detailed income data than above or below median household incomes could help examine equity concerns more accurately. Request crash data from law enforcement and health services to start differentiating between scooters and other modes when dealing with incidents to improve tracking. Repeat SMD evaluations to assess SMD trends and truly characterize the service and its long-term evolution (e.g. crashes). VIII. Undertake additional research or studies including more detailed analysis of specific issues of interest flagged in the pilot evaluation This evaluation provides a valuable starting point in terms of flagging the most critical issues but has foregone detailed focus on specific issues in the interest of a holistic assessment of the SMD pilot. Several more detailed analyses could be undertaken with available data and separate longer-term studies and/or surveys incorporating learnings from this evaluation could help improve SMD system performance, rider experience and community responses, including: Short term studies with available data Examining key results (e.g. perceptions) by sample segment including perception and experience by gender, primary mode, and frequency of use. Examining trip characteristic differences by corridor. Looking at trip characteristics by time of the day and weekends versus weekdays. Examining geographic distribution of operational problems – are incidents concentrated in one or more areas in Arlington? Do they correlate with elements of the infrastructure or land use? Taking a closer look at “late night travel”, potentially complemented with an intercept survey to characterize such trips and their link to accessibility. Examining the community’s reaction to the pilot before and after the installation of corrals – did complaints, operational challenges and trip change after the installation of corrals? Performing more sophisticated modeling of SMD behavior using attitudinal and demographic variables in order to understand the determinants of satisfaction, frequency, trip purpose etc. Conducting more sophisticated content analysis of the Mobility inbox data. Conducting more sophisticated correlation analysis based on bivariate maps obtained and discussed in this report. Long-term studies with additional data Studying the impact of e-scooters on accessibility and comfort for people with disability. Collaborating with other pilot programs and leveraging findings from Arlington utilization rates to estimate an “adequate” level of service that planners should aim for in designing SMD programs. This would be a similar effort to the ITDP bike share planning. Evaluating acceptable levels of broken SMDs for new technologies or a new mobility service. Evaluating communication techniques for best retention rates within apps (tests, games etc.). Examining travel behavior from the perspective of mode substitution between cars/TNCs and SMDs. Developing performance measures for shared mobility devices. Developing a scoring system/service standard for performance measures to rank and evaluate operators, mandating a minimum service level for continued operation in Arlington County.

The results and the recommendations of this report should be read within the context of Arlington County and the data collected during the pilot. The limited time SMDs have been in operation and the corresponding limited data and research means that the characterization of SMDs and how cities manage them will continue to evolve. This makes it important for local policymakers to continue monitoring and collecting data in order to derive structural and systemic trends, accurately characterize these services and ensure their integration into the Arlington County transportation landscape that yields desired benefits while mitigating negative externalities.

Arlington County’s SMD pilot program: key highlights

Service operations (i.e. “Supply”)

• The pilot was launched with an average of 706 daily SMDs deployed with two operators and ended with 806 SMDs deployed in June with six operators. It fluctuated in between, with the lowest deployment occurring during winter months.
• Most of the SMD deployment was concentrated in the Rosslyn-Ballston (45%) and Route 1 corridors (10%) with Columbia Pike less well-served by SMDs when accounting for respective residential population.
• Arlington County received more service per population (4.0 SMDs per 1,000 people) than Washington DC (2.4 SMDs per 1,000 people) and Capital Bikeshare bikes in Arlington (3.1 SMDs per 1,000 people).
• North Arlington received 1.3 to 2.5 times more service than South Arlington.
• Ten main operational challenges were identified in the pilot consisting of: (1) inconsistent deployment of SMDs, (2) problematic deployment sites such as bus stops and pedestrian right-of way on sidewalks, (3) high operating speed, (4) sidewalk riding, (5) broken SMDs, (6) stolen and vandalized SMDs, (7) idle SMDs , (8) incorrectly parked SMDs, (9) crashes and injuries and (10) data. Five of the operational challenges are also breaches of the MOA.
• There were 69 crashes in total between October 2018 and June 2019. Those resulted in approximately 29 injuries.
• In terms of the adequacy of the pilot-related information supplied during to the pilot:
  • 20% of non-riders received their information on the SMD pilot from Arlington County’s website.
  • 20-22% of SMD riders and 43% of non-riders did not know what the “laws” are.
  • Less than half of respondents (45%) indicated that they had received information from operators on local regulations, and less than a third (30%) indicated that they received information from operators on how to file a complaint.

Service utilization (i.e. “Demand”)

• There was a total of 453,690 SMD trips in Arlington County between October 2018 and June 2019.
• The lowest-trip months were January and February (around 23,000 trips each month) and the highest trip month was May (around 80,000 trips).
• The average trip distance was 0.94 miles and 90% of trips were for less than two miles. SMD riders traveled a total of 425,124 miles in Arlington between October 2018 and June 2019.
• The average trip duration was 14 minutes.
• 25% of trips occurred during peak travel times with 12% of trips taking place during the morning peak and another 13% of trips (approximately 58,500 trips) during the afternoon.
• 70% of trips took place during weekdays while 30% of trips took place on weekends, although Saturday ridership was the highest day of ridership over the nine-month period
• Most riders remain within the bounds of the County, with 89% of trips starting and ending in Arlington.
• Most trips and routes clustered around the two main transit/commercial corridors – the Rosslyn-Ballston corridor and Route 1 corridor.
  • The Rosslyn-Ballston corridor included 60% of trip origins and 55% of trip destinations. Within the Rosslyn-Ballston corridor, key e-scooter arterials included the Key Bridge (in and out of Washington DC), N Lynn Street, Wilson Boulevard, Clarendon Boulevard, and 9^th Street in between Clarendon and Wilson boulevards is also used.
  • The Route 1 corridor included 17% of trip origins and 35% of trip destinations. For this corridor, 12^th St S, S Eads St, and S Crystal Dr were areas of highest use.
  • The Columbia Pike corridor recorded fewer trips, with 4% of trip origins and 5% of trip destinations.
• In terms of infrastructure use, bike lanes were most used with 62% of e-scooter riders always-to-often using bike lanes, followed by shared lanes with cars (24%). The least-used facility was trails.
• SMD- riders preferred to ride on protected bike lanes (67% of respondents chose it as a top or second choice) followed by regular bike lanes (47% of respondents chose it as either a first or second choice). The least- preferred facility was sharing travel lanes with cars, and sidewalks were second-least preferred.
• Trips occurred in areas of high transit supply with scooter trips originating 0.38 miles away from a Metrorail station and ending 0.48 miles away from a Metrorail station. Ballston Metrorail has the most trips in its vicinity (78,000 parking events within 500 meters).
• The feedback form pointed to social and/or entertainment (21% of e-scooter riders) as the category most cited as a primary trip purpose for using e-scooters in Arlington, followed by shopping or errands (18% of e-scooter riders) and connecting to Metrorail (18% of e-scooter riders).
• When asked about the mode they would have used to make the trip, 37% of e-scooter riders and 22% of dockless e-bike riders indicated replacing walking, while one in five SMD riders indicated replacing a ride-hailing trip, and 13% indicated replacing a personal car or other motor vehicle (for a total of 32% of e-scooter riders having replaced an automobile trip).
• The online feedback form provided some preliminary insights into SMD rider In particular, a larger proportion of rider respondents were male (63%) than were female (37%), and rider-respondents reported a relatively lower average age than non-rider (more than 63% of e-scooter riders born after 1980 (compared to 22% for non-SMD riders). In terms of occupation and education, the largest proportion of riders was made up of full-time employees (66% for e-scooters and 63% for dockless e-bikes) and with a lower rate of advanced degrees (34% for e-scooter riders) than non-rider respondents (51%), yet still educated.

The community’s reaction to the pilot

• When asked about why they use SMDs in Arlington County, the majority of e-scooter rider respondents (55%) selected “to get around faster” as one of their top three choices. This was followed by “convenient” (44%) and “fun to ride” (36%).
• When asked about why they haven’t used e-scooters in Arlington in a close-ended form, the first popular choice was “I don’t think e-scooters are safe” selected by 58% of non-SMD riders and 32% of dockless e-bike riders and the third most popular choice was “I feel unsafe riding in the street” selected by 36% of non-SMD riders and 21% of dockless e-bike riders. This suggests that the main barrier to using e-scooters in Arlington pertain to the adequacy of the infrastructure or a safe place to ride.
• When asked about specific measures that could lead them to start using SMDs, most non-SMD riders (68%) said that “none of these changes would encourage them to start using SMDs”. For e-scooter riders, the most popular responses for what would make them use e-scooters more often were “safer places to ride” (51% for e-scooter riders and 44% for dockless e-bike riders),  and “more e-scooters available in Arlington” (42% for e-scooter riders and 27% for dockless e-bike riders).
• When asked what infrastructure would make them feel safer, most SMD riders (e-scooter riders and dockless e-bike riders) wanted bike lanes separated from motor vehicles traffic with a physical barrier while most non-SMD riders wanted designated e-scooter parking.
• When asked which type of problems they encountered, 36% of e-scooter rider respondents chose “none of the above”. Of the remaining 64%, the majority (60%) encountered either mechanical issues with their e-scooters or issues unlocking/locking e-scooters via the mobile app.
• When asked about safety and comfort around e-scooters as pedestrians and drivers, the analysis revealed a difference in perception between SMD riders and non-SMD riders.
  • 73% of non-SMD riders who responded to the survey did not feel safe as pedestrians around riders on e-scooters as opposed to 41% of dockless e-bike riders, and just 15% of e-scooter riders.
  • 65% of non-SMD riders reported often to always encountering blocked sidewalks due to e-scooters being improperly parked compared to 43% of dockless e-bike riders and only 16% of e-scooter riders.
  • 76% of non-SMD riders reported being very uncomfortable to uncomfortable as drivers in Arlington County around riders on e-scooters compared to 47% of dockless e-bike riders and only 21% of e-scooter riders.
• The online feedback form also included open-ended questions about the impact of improperly parked SMDs. Out of the people who responded to this question (2,876, 71%), a plurality (884 responses, or 31% of total open-ended responses received) qualitatively suggested that SMDs block the path of pedestrians in sidewalks, driveways, and other common-use areas in Arlington County. After that, the responses were mixed with the same share of respondents (14%) qualitatively stating a safety concern on one hand and no to minimal negative impact on the other.
• The analysis also examined voluntary, self-initiated emails received to the Mobility Inbox (mobility@arlingtonva.us). A total of 727 emails were received to the Mobility inbox. The number received each month decreased significantly over the course of the pilot, from October (226 e-mails) to June (38 e-mails). This is consistent with staff’s experience running a similar email account for the free-floating car-sharing services during that pilot program.
• Key topics in the emails received to the Mobility inbox included: “parking” followed by sidewalk riding, safety, rider behavior and underage riding.
• Additionally, outreach community events in Arlington were an important source of community feedback. Most intercepted respondents confirmed that the pilot has challenges, but the majority expressed an interest in resolving issues rather than ending the program. This summarizes an important (qualitative) perspective of people who potentially did not have access to the online feedback form. The documented difference or more positive reaction to the pilot should be taken into consideration when evaluating the Arlington Community’s reaction to the pilot, potentially offsetting in part some biases in self-selected complaints received through the Mobility Inbox or the feedback form.

SMDs vs. Capital Bikeshare: key measures

• Total Trips:
  453,690 SMD (pilot period)
  261,129 Capital Bikeshare (2018)
• Total Distance (miles):
  409,548 SMD (pilot period)
  511,887 Capital Bikeshare (2018)
• Average Trip Distance (miles):
  0.94 SMD (pilot period)
  1.96 Capital Bikeshare (2018)
• Average Trip Duration (minutes):
  14 SMD (pilot period)
  16 Capital Bikeshare (2018)
• Service Level:
  4.0 SMDs/1000 people SMD (pilot period)
  3.1 SMDs/1000 people Capital Bikeshare (2018)

Footnotes:  

Any such initiative should be done while balancing the importance of the convenience to riders so as not to negatively impact demand and the overall rider experience.

See body of the report for a discussion of crash data and safety analysis limitations.

500 meters is equivalent to 0.31 miles.

MORE RESOURCES FOR THIS REPORT:

• Read the blog post.
• View Arlington County’s SMD website 

Note: On 11/14/2019 the report has been updated on page 20 to edit our review of the Santa Monica research into “Several studies published in 2019 looked at the safety of scooters. Santa Monica researchers found that people going into the hospital for scooter injuries was around 50 more times than for bike injuries in the same year (249 vs. 195)”.