Results From a Campus Population Survey of Near Misses,Crashes,and Falls While E-Scooting,Walking,and Bicycling

Abstract

Dockless e-scooters were used for 86 million trips in the United States in 2019, indicating great potential as a new transportation mode in U.S. cities and on university campuses. Yet, little is known about how e-scooter users interact with people walking, bicycling, and driving. Although several studies have examined e-scooter injuries reported in hospital data, transportation-related near misses are chronically understudied in general, and even more so for this newer mode of transportation. In this paper we present the results of an online survey of 1,256 university staff (22% response rate) in Tempe, AZ. Using a single population, we compared the prevalence of self-reported incidents and injuries among those who use e-scooters, walk, and bicycle. Our results indicated a higher percentage of respondents reported incidents associated with walking (25%) than e-scooting (11%) or bicycling (9%), but e-scooter users were the most likely to report incidents resulting in a crash. E-scooter users were also more likely to report issues related to pavement, equipment, or losing control, whereas people walking and bicycling were more likely to report conflicts with other roadway users. Our findings suggest important areas for policy and infrastructure innovation, including prioritizing separate space for e-scooters to mitigate conflicts with pedestrians, and continuing to evolve rider training and speed governance to help keep e-scooter users safe. Other findings underscore the importance of measuring near misses to develop a comprehensive picture of transportation safety.

Keywords

pedestrians bicycles human factors safety accessibility walkability sustainability and resilience transportation and society equity in transportation micromobility and active transportation

Over the last few years, dockless electric scooters (e-scooters) have become a fixture of the transportation system in many cities and on college campuses. Before the pandemic, e-scooter ridership in the United States grew dramatically, from 40 million trips in 2018 to 86 million trips in 2019 ( 1 , 2 ). However, the pandemic led to dramatic changes in transportation trends in 2020: micromobility demand initially increased in 2020 ( 3 ), probably as an alternative to public transportation at the onset of the pandemic, but most e-scooter programs eventually shrank or closed in response to a drop in ridership because of mandatory stay-at-home orders and potentially fears related to shared transportation ( 4 ). Yet as cities opened back up in 2021, e-scooters began to reemerge to once again fill a gap in urban transportation services. This reemergence presents an opportunity to evaluate how e-scooters work with other transportation modes and what policies could help maximize transportation options and minimize conflict in the future.

Because of the pandemic, most e-scooter research is based on data collected before 2020. Studies suggest that at that time, e-scooters were generally popular or seen as providing a valuable service, even among nonusers ( 5 – 8 ). However, safety emerged as an early and important issue owing to spatial conflicts between e-scooter users and other road users who shared (often insufficient) sidewalk and roadway space with them. For example, about 65% of current and past riders in an online survey of university staff in Tempe, AZ, reported feeling “somewhat” or “very” safe while riding an e-scooter ( 9 ). The main barriers to e-scooting included traffic safety concerns such as worrying about interactions with other roadway users, feeling unsteady, out of control, or worried about falling, and not having enough safe places to ride. Pimentel and Lowry found that 92% of respondents to an online panel survey of Washington, Oregon, and Idaho residents in cities with micromobility systems named auto traffic a deterrent to e-scooting and biking; the deterrent was stronger for e-scooter use than for bike or e-bike use ( 10 ).

Other research examined the perceived safety of roadway and sidewalk users when walking, biking, or driving around e-scooter riders. A survey in Rosslyn, VA, found that 56% of respondents felt unsafe or very unsafe walking near e-scooter riders ( 11 ). These results were strongly negatively correlated with past e-scooter use and were stronger than for walking near bicyclists, a result mirrored for drivers. It is not clear how often respondents considered bike lanes, which have been found to significantly improve driver comfort around bicyclists and would theoretically have a similar effect for e-scooter riders ( 12 ).

In several surveys, e-scooter users requested bicycle infrastructure as a way to feel safe riding in the street, often explicitly in contrast to riding on the sidewalk ( 1 , 6 , 10 , 13 –15), although not all e-scooter riders report feeling safe in bike lanes ( 16 ). Providing bike lanes for e-scooter users would complement pedestrians’ desire that e-scooter riders keep off the sidewalks, as documented in several pilots ( 6 , 7 , 17 ). Public feedback about e-scooters, sidewalks, and near misses between pedestrians and e-scooter riders was so strong in Denver, for example, that the City of Denver changed its ordinance to allow e-scooter users to use bike lanes mid-pilot ( 6 ). At the same time, some e-scooter riders also noted that riding on the sidewalk felt unsafe ( 18 ). Additionally, several studies have noted higher injury incidence on the sidewalk ( 19 , 20 ), although this is likely to be because this is where e-scooter users feel comfortable riding.

Studies examining e-scooter injuries have found mixed results, depending on the study years and location. More recent studies, which are likely to better reflect trends with shared e-scooters, have found that head injuries and facial and extremity fractures are the most common injury type ( 21 – 23 ). Injuries were caused by a variety of actions, including falls (over 80% of incidents in Portland and Los Angeles [LA] [ 5 , 21 ]), and losing control of a scooter on poorly maintained pavement or for another reason (56% of incidents in a multicity survey [ 16 ]). A smaller percentage of incidents involved crashing into another object (such as a car, street lamp, curb), although this varied by place. For example, collisions with inanimate objects were associated with 17% of e-scooter incidents in Austin and 11% of incidents in LA ( 20 , 21 ). Fourteen percent of incidents in Portland resulted from a collision with a car ( 5 ), whereas 12% of incidents in the multicity survey resulted from being hit by a driver or someone opening a car door ( 16 ), and 9% in LA resulted from being hit by a moving vehicle or object ( 21 ). Some incidents, even if they involve an inanimate object, are reportedly the result of e-scooter users practicing defensive scooting to avoid a collision with a car ( 19 , 20 ). A few studies have compared injuries across modes, despite difficulties from often lacking exposure data. Cicchino et al. found similar results for injury severity among bicyclists and e-scooter users treated in the emergency department, but e-scooter riders were more likely to sustain concussions and other head injuries, probably because of a lower incidence of helmet usage ( 24 ).

Despite the high incidence of complaints about e-scooters by pedestrians, collisions involving the two have tended to be rare (less than 5% in most cases [ 5 , 21 ]). In the Denver pilot survey, 34% of respondents indicated that they were pedestrians who were hit or almost hit by a scooter ( 6 ), but this figure is most likely inflated compared with other cities’ numbers because it includes near misses. Research has found that near misses are significantly related to perceived safety for bicyclists ( 25 ), which could plausibly be applied to pedestrians with regard to e-scooters (or other users)—particularly for those who are less able to evade oncoming e-scooters or who may be more likely to be injured if hit. Additionally, carelessly parked e-scooters can impede pedestrian traffic and create tripping hazards or contribute to injury from needing to be moved. In LA, 2% of the injury reports resulted from a pedestrian tripping over a parked scooter, and another 2% resulted from pedestrians attempting to lift or carry a scooter out of their path ( 21 ).

In this paper we present an exploratory analysis of survey responses about crashes and near misses while e-scooting, walking, and bicycling within a single large sample of university staff; the use of a single sample allowed us to compare the prevalence of incidents and incident types, crashes, and injuries and injury types, particularly as they related to bicyclists and pedestrians. Our findings contribute to the body of e-scooter safety literature by providing insights into the safety issues associated with these three modes, particularly relative to environments where e-scooter users, pedestrians, and bicyclists may share space, like college campuses, urban areas without bicycle networks, and trails. We hope these findings will help cities, universities, and e-scooter companies develop strategies to manage conflicts and create safe, comfortable travel for all roadway users.

Data and Methods

Background

The survey was conducted among Arizona State University (ASU) staff in Tempe, AZ. Staff were targeted because of the timing of the survey (late spring, near finals) and because they represent a greater range of income and education levels and work year-round, making them less likely than faculty or students to leave Tempe for extended periods of time (which could have influenced the survey results). ASU’s campus is multimodal and allows walking, bicycling, and skateboarding throughout most of the campus, but prohibits nonuniversity motorized vehicles including e-scooters and automobiles. Despite this prohibition, e-scooting occurred on campus when e-scooters were first introduced to Tempe; they were eventually banned from campus and removed or restricted via geofencing. This survey was conducted after e-scooters were banned. As such, the responses here should be viewed as reflecting e-scooter riding in Tempe and the surrounding areas, in contrast to on campus. That said, sidewalks near campus are often crowded with users, including e-scooters and bicyclists at times. These results may therefore be more applicable for areas with a higher density of users in shared spaces, such as on a college campus or in city without a developed network of bike lanes.

Survey Construction and Recruitment

On May 2, 2019, we administered an online survey via email to the 5,720 ASU staff (not including faculty) who work on the Tempe campus. The survey, hosted on Survey Gizmo, included a $5 Amazon e-gift card incentive for the first 200 respondents, and a chance to be entered into a draw for one of ten $20 Amazon e-gift cards for the remaining respondents. The survey was estimated to take about 15 min and received 1,385 responses, of which 1,256 (response rate: 22%) were complete and analyzed in this study.

We developed the survey based on a literature review of e-scooter-related research (5 –8, 21 ) and past related work on bicycling ( 12 , 25 , 26 ). The survey included approximately 170 questions, although skip logic reduced that number significantly for most respondents owing to the large number of questions about near misses and crashes that were inapplicable if not experienced. We asked questions about the following categories:

Experiences with crashes and near misses while traveling via various modes;

E-scooter and bicycle usage (both personal and bike share);

Benefits and barriers to using e-scooters, bicycling, and walking;

Enjoyment and perceptions of safety while traveling via various modes;

Opinions about laws associated with e-scooters and bicycling; and

Sociodemographics.

This paper focuses on the findings about near misses, crashes, injuries, and perceptions of safety. We asked only about experiences over the past year to bound the recall period and mitigate potential recall bias.

Methods

Survey data were coded and analyzed using the statistical software programs STATA SE (StataCorp, College Station, TX). Statistical tests included chi-square, Z, and Kruskal–Wallis, the latter of which is a nonparametric version of analysis of variance that accommodates ordinal variables. Note that some analyses are based on small sample sizes, which may limit the generalizability of related findings.

Results

Characteristics of the Survey Population

In comparison to the city of Tempe, the sample population overrepresents women and households with young children, and underrepresents participants aged 18 to 24 and over 65, households earning less than $35,000 a year, and respondents who are Black or African American alone or Asian alone. (For more detail about the populations, see research by Sanders et al. [ 9 ].)

Table 1 shows the demographics of the survey population by e-scooter usage, walking, and bicycling. The pedestrian sample matched the entire sample given that all respondents walked at least occasionally, and the large number of bicyclists (94% of respondents having ridden a bicycle) closely matched the pedestrian data in most cases. However, there were a few significant differences between those groups and e-scooter users, the latter of which only comprised about one-third of the survey population. For example, younger users predominated for all activities, but comprised a significantly larger percentage of e-scooter riders, with 55% of riders aged 18 to 34. Women comprised the majority of the sample, but the gender split was significantly more even for e-scooter riders. There were no significant differences between activity by race. E-scooter users were significantly less likely to use a personal vehicle as their main mode of travel. Income was significantly lower among e-scooter users (29% with incomes less than $35,000, compared with 20% for pedestrians and bicyclists). Having children below the age of 16 in the household was slightly less common among e-scooter riders, but not significantly so.

Table 1.

Survey Population Characteristics, by Mode

	E-scooter rider (n = 407) %	Pedestrian^a (n = 1,256) %	Bicyclist (n = 1,186) %
Age
18–24	10	5	5
25–34	45	31	30
35–44	26	25	25
45–54	11	19	19
55+	6	17	18
Missing	1	2	2
Total	100^b	100	100
	Z test significant p ≤ 0.001
Gender identification
Man	45	34	35
Woman	53	62	62
Other/decline to say	2	4	4
Total	100	100	100
	Z test significant p ≤ 0.01
Race/Ethnicity^c
White alone	68	67	69
Hispanic/Latino alone	11	10	9
Black/African American alone	3	3	3
Asian alone	4	5	5
Native American alone	1	1	1
Mixed race	7	6	6
Some other race alone	2	1	1
Decline to say	6	7	7
Total	100	100	100
	Z test not significant
Main mode of travel
Walk	7	7	7
Bicycle	9	7	7
E-scooter	2	1	1
Public transit	5	5	5
Personal vehicle	73	78	78
Other	3	2	2
Decline to say	0	0	0
Total	100	100	100
	Z test significant p ≤ 0.05
Annual household income
Less than $35,000	9	5	5
$35,000–49,999	20	15	15
$50,000–74,999	25	22	22
$75,000–99,999	15	14	15
$100,000–149,999	13	17	17
$150,000 or more	8	10	10
Decline to say/missing	10	16	15
Total	100	100	100
	Z test significant p ≤ 0.05
Children younger than age 16 in household
Yes	27	30	30
No	69	65	65
Decline to say/missing	4	5	4
Total	100	100	100
	Z test not significant

Data for pedestrians matched that for the entire sample, since all people walked.

Column totals may equal 99 or 101 because of rounding.

All race values presented are those for the non-Hispanic race alone. All respondents identifying as Hispanic were classified as “Hispanic/Latino alone.”

Trends Related to Reported Incidents

To understand incident patterns within the sample, survey participants were asked a series of questions about falls, near misses, and crash experiences (“incidents”) in the past year.

Prevalence of Incidents Within the Sample

Table 2 shows the percentage of respondents of each group who had experienced a crash, near miss, and/or fall while walking, e-scooting, or bicycling in the past year. In general, the data indicate that a higher percentage of people experienced an incident while walking than while e-scooting or bicycling over the past year (25% compared with 11% and 9%, respectively). How often the respondent walked or biked in the last month was positively and significantly (p ≤ 0.0001) correlated with reporting an incident while walking and bicycling, respectively (data not shown), which fits with expectations related to increased exposure. However, the association between e-scooting frequency and reporting an incident was only marginally significant (p ≤ 0.10), indicating a potentially different relationship with exposure compared with other modes.

Table 2.

Prevalence of Incidents, Crashes, and Injuries Within the Sample

Mode	Respondents experiencing a crash, near miss, and/or fall in past year				Incident statistics
Mode	Once (%)	Twice (%)	Three or more times (%)	Total* (%)	Reported number of incidents	Incidents resulting in crash (%)	Crashes resulting in injury (%)
Walk (n = 1,256)	109 (9%)	78 (6%)	125 (10%)	313 (25%)	640	64 (10%)	57 (89%)
E-scooter (n = 407)	34 (8%)	6 (2%)	1 (0.2%)	43 (11%)	49	17 (35%)	14 (82%)
Bike (n = 1,186)	62 (5%)	17 (1%)	32 (3%)	112 (9%)	192	54 (28%)	29 (54%)

Totals exceed the prior columns owing to minor nonresponse.

Of those who reported incidents, e-scooter users were significantly more likely than pedestrians to experience a crash (35% compared with 10%, respectively). There was no significant difference between pedestrians and e-scooter users in relation to experiencing an injury as a result of a crash, and both were significantly more likely to experience an injury than bicyclists.

Incident Type

Those who reported incidents were then asked which of a list of incidents had occurred. Table 3 shows two types of incidents—those that involve another person and those that are solo incidents. For e-scooter users, the two most common incidents were solo incidents: falling or almost falling without hitting something else (33%) and hitting or almost hitting something on the ground (29%). Being hit or almost hitting another person was a close third (27%). In contrast, being hit or almost hit by another person was the top incident by far for people walking and riding a bicycle (79% and 60%, respectively), with the next most prevalent incident falling or almost falling because of something on the ground for people walking, and falling or almost falling without hitting something else for people biking.

Table 3.

Incident Type by Mode

What happened in the crash, near miss, or fall? (nonexclusive)	E-scooting (n = 49) (%)	Walking (n = 640) (%)	Bicycling (n = 192) (%)
Incidents involving another person
Person hit/almost hit me	27	79	60
I hit/almost hit person	18	6	9
Solo incidents
I fell/almost fell without hitting something else	33	na¹	15
I hit/almost hit/fell/almost fell because of something on ground	29	12	8
I hit/almost hit something above ground	12	3	7
I slipped/almost slipped on wet/slippery/unstable surface	8	6	7
Other	6	7	7
E-scooter/bicycle malfunction	10	na	3

¹“na” indicates that this row was not presented for walking incidents.

We further explored the incidents involving other people by asking how the other person was traveling (see Table 4); in all cases, users were more likely to report being hit or almost hit than hitting or almost hitting someone else. E-scooter users were most likely to report being hit or almost hit by another e-scooter user or a person walking or running, followed by a person driving. Among e-scooter users who said they hit or almost hit someone, nearly 80% of those incidents involved a person walking or running.

Table 4.

Travel Mode of Other Party Involved in Incident by Mode

Travel mode of person who hit or almost hit respondent	E-scooting (n = 13) (%)	Walking (n = 508) (%)	Bicycling (n = 115) (%)
Driving	23	14	67
Bicycling	9	25	13
E-scooting	31	45	13
Walking/running	31	1	2
Riding a skateboard	8	11	1
Travel mode of person respondent hit or almost hit	E-scooting (n = 9) (%)	Walking (n = 39) (%)	Bicycling (n = 17) (%)
Driving	11	13	24
Bicycling	11	31	18
E-scooting	0	39	18
Walking/running	78	8	24
Riding a skateboard	0	10	24

For people walking or running, 45% reported being hit or almost hit by an e-scooter user, followed by bicyclists and then car drivers. Among those who said that they hit or almost hit someone, 39% of those incidents involved an e-scooter user and 31% involved a bicycle. For bicyclists, however, 67% of incidents involving another person hitting or almost hitting them were attributed to drivers, suggesting conflicts on the roadway, at junctions, or both. Of the few bicyclists who reported hitting or almost hitting someone else, 24% reported drivers, someone walking or running, and someone skateboarding.

Injury Statistics

Table 5 shows the location of injuries for each mode. Only one injury each for e-scooting and walking was reported as severe; all others were reported as minor injuries. Most of the injuries, regardless of mode, occurred to the legs (including knees, ankles, and feet). For e-scooter users and bicyclists, injuries to the arms (including wrists and hands) were also common. In contrast to several studies on e-scooter injury data ( 21 – 23 ), the data reported here reflect injuries that were rarely reported to a hospital (only 29%, 25%, and 7% of injuries were reported for e-scooter users, people walking, and bicyclists, respectively).

Table 5.

Injured Body Part by Mode

Injured body part (nonexclusive)	E-scooting (n = 14) (%)	Walking (n = 57) (%)	Bicycling (n = 29) (%)
Arms/wrists/hands	64	33	62
Chest/torso	0	14	3
Head/neck	0	9	10
Hips/back	0	9	10
Legs/knees/ankles/feet	79	72	62

Factors Contributing to Incidents

We also asked respondents about any contributing factors to their incidents, to understand how various dynamics may be associated with crashes, near misses, and falls (Table 6). Relatively few people reported being distracted, particularly among people who had incidents while walking or bicycling. Even fewer people reported intoxication as a contributing factor, fitting with low alcohol usage reported elsewhere ( 27 ). It is unclear whether and to what degree the low reported numbers for distraction and intoxication reflect biased data. However, 35% of people reporting an incident while e-scooting indicated that going too fast and losing control contributed to the incident (compared with only 5% of bicyclists reporting an incident), suggesting that speed and handling are issues that should be addressed among e-scooter riders.

Table 6.

Contributing Factor by Mode

Contributing factor (nonexclusive)	E-scooting (n = 49) (%)	Walking (n = 640) (%)	Bicycling (n = 192) (%)
I was distracted	10	3	3
I was intoxicated	2	1	1
I was going too fast and lost control	35	n/a	5
None of the above	41	79	73
Other	22	15	14

Overall, 79% and 73% of people who reported incidents while walking and bicycling, respectively, reported no contributing factor (either listed or within the “other” option), compared with 41% of respondents reporting e-scooter incidents. Among those who did report “other,” the vast majority involved another party behaving in an unsafe manner.

Perceived Safety Compared with Objective Safety

We also explored the association between perceived safety and reported incidents among users in our sample, as shown in Table 7. We found a highly significant (p ≤ 0.001) association between having experienced an incident and both feeling safe walking and the perceived likelihood of experiencing an injury while walking. Having experienced an incident while e-scooting was marginally significantly related to perceptions of e-scooter safety, but significantly (p ≤ 0.05) related to beliefs about the likelihood of experiencing an e-scooter injury. Having experienced an incident while bicycling was not related to perceptions of bicycle safety, but was significantly (p ≤ 0.05) related to beliefs about the likelihood of experiencing an injury while bicycling. The number of incidents was not significantly related to perceptions of safety or beliefs about the likelihood of injury for any mode.

Table 7.

Perceptions of Safety Relative to Reporting an Incident by Mode

How safe do you generally feel…	Walking*** (n = 1,235)		E-scooting (n = 394)		Bicycling (n = 1,102)
	No incident (%)	Incident (%)	No incident (%)	Incident (%)	No incident (%)	Incident (%)
Very safe	43	31	11	14	17	17
Safe	49	54	57	40	62	61
Unsafe	7	12	34	28	18	17
Very unsafe	1	3	9	19	3	5
What is the likelihood you’ll ever be injured…	Walking*** (1,208)		E-scooting* (n = 388)		Bicycling* (n = 1,126)
	No incident (%)	Incident (%)	No incident (%)	Incident (%)	No incident (%)	Incident (%)
Not at all likely	55	39	11	2	12	9
Not very likely	29	37	29	20	37	28
Somewhat likely	13	20	44	46	43	47
Very likely	3	4	16	32	8	17

Note : Significant relationships between frequency of activity and variable of interest indicated by the following: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Discussion

To our knowledge, this study is the first to compare crashes, near misses, and falls (“incidents”) while walking, e-scooting, and bicycling among a group of people in North America. The findings suggest some important takeaways for future efforts to include e-scooters or other micromobility modes as part of the urban transportation system.

Prevalence of Incidents

Within our sample, respondents were much more likely to report having experienced an incident while walking than while bicycling or using e-scooters. The majority of reported incidents did not result in a crash, although the percentage that did was higher for e-scooter users and bicyclists than for people walking (35% and 28% compared with 10%, respectively). Most of the reported crashes resulted in an injury (significantly more so for pedestrians and e-scooter users than for bicyclists), but the vast majority of injuries were minor and most were unreported and therefore unrepresented in hospital or clinic records.

Incident Type

Findings about incident type suggest three important areas for safety-related action. First, in all three groups, but particularly among people walking and bicycling, a high percentage of respondents reported being hit or almost hit by another person at some point. As the crash statistics indicate, the large majority of these incidents were near misses. For people walking, 45% of these incidents involved an e-scooter user, fitting with studies documenting pedestrian complaints about sharing the sidewalk with e-scooter users during various city pilots ( 5 , 6 , 13 , 14 ). Only 14% of pedestrians reported an incident involving a driver, a somewhat surprising finding given that drivers are responsible for the majority of pedestrian injuries in crash records (see, e.g., tims.berkeley.edu), but perhaps indicative of reduced exposure to vehicles on and around the university campus for this population. Although e-scooters are banned from campus, their presence along the campus edge and in the walkable downtown area may lead to numerous sidewalk conflicts. Conflicts have been shown to be especially dangerous for pedestrians typically underserved by transportation infrastructure including people with vision and/or hearing impairment, young children, and older adults ( 28 ).

In contrast, bicyclists were much more likely to list drivers as the party who hit or almost hit them, perhaps reflecting roadway conditions that are not sufficiently safe. It may also be that the speed differential between e-scooters and bicycles—which plausibly impacts perceived comfort and safety—is relatively small, in contrast to that between e-scooters and people walking, and therefore interactions between bicyclists and e-scooters are less noticeable. The speed differential between drivers and bicyclists in the street, in contrast, is likely to be quite impactful and as previous bicycling research shows, people are more likely to report near misses when those incidents involve vehicles ( 29 ).

Relatively few respondents (only about 5% of the sample) reported hitting or almost hitting someone else, which may accurately reflect each person’s experience, but may also belie a bias toward remembering incidents in which the person felt fear or discomfort as opposed to potentially causing those feelings for someone else. However, among those who said they hit or almost hit someone, the data suggest particular conflicts between e-scooter users and pedestrians, which may reflect e-scooter usage on sidewalks, particularly when roads feel unsafe, and potentially use on campus before the prohibition. Our findings support policies and efforts to provide separate space for e-scooter users and pedestrians, and are in line with previous research that has highlighted bicycle and pedestrian mode separation as important for safety even when vehicles are not present ( 30 ).

The other major incident group involved solo incidents, in which the respondent hit or almost hit or fell or almost fell without interacting with another person. These included incidents involving something on the ground, like a pothole; a slippery, wet, or unstable surface; something above the ground like a pole; an e-scooter or bicycle malfunction; or simply a fall without hitting something else. In bicycling, collisions and falls that do not involve a motor vehicle have been found to account for one-third of incidents ( 31 ) and have been shown to be associated with injury ( 32 ). Similarly in our findings, members of each respondent group reported solo crashes and falls, though reports were proportionately higher for e-scooter users. More stable equipment and enhanced rider training could potentially mitigate the likelihood of these incidents for e-scooter users, and any efforts to improve pavement conditions and unstable surfaces would also benefit other roadway users.

Additionally, these findings suggest a need for a near miss monitoring or mapping system, such as BikeMaps.org, where users can log complaints to prompt action before a crash occurs. In crowdsource reports of incidents, near misses are reported at approximately 3:1, compared with crashes and near miss data ( 33 ). An even better system could be one that systematically captures near misses, as some sensor and video networks are working toward.

Injury Statistics

Despite the attention that e-scooters have received for injuries ( 21 – 23 ), only 3% of e-scooter users within the sample reported experiencing an injury, similar to the prevalence of injury among people walking and bicycling. That said, both e-scooter users and people walking were significantly more likely than people bicycling to report experiencing an injury in the event of a crash. Among those injuries, only one was severe for e-scooter users and people walking and few were reported to a hospital, supporting prior research finding underreporting—and a lack of comprehensive understanding—of injuries for nonmotorized users ( 34 ). Within those injuries, e-scooter users indicated injuries to the arms (including wrists and hands) and legs (including knees, ankles, and feet), but no injuries to the head. The location and severity of injury for e-scooter users contrasts with multiple studies finding head and facial fractures, and fractures in general, to be common among injured e-scooter users ( 21 , 22 )—but those studies also reflect injuries serious enough to present at the hospital. A similar pattern can be observed for bicycling injuries, both within our sample (mostly arm and leg injuries) and within larger bicycling injury studies; research on comprehensive pedestrian injury patterns is lacking. More research is needed to understand the potential differences with regard to the occurrence of various injuries within a population and injury severity.

Additionally, our data support the Hydén Safety Pyramid ( 35 ), adapted slightly in Figure 1 ( 36 ), which proposes that most incidents are not serious, and that frequency decreases as injury severity increases. However, the less serious incidents can be important for teaching us about more serious incidents.

Figure 1.

Adapted Hydén Safety Pyramid ( 35 , 36 ).

Contributing Factors

Our data indicate that neither distraction nor alcohol usage were prevalent in the study, although distraction was more prevalent within e-scooter incidents than those for people walking and bicycling. More salient, however, is the finding that 35% of people reporting an incident while e-scooting indicated that going too fast and losing control contributed to the incident (compared with only 5% of bicyclists). This dynamic may reflect the relative newness of e-scooters and potential operational problems that can occur with a lack of familiarity; indeed, research by the CDC and the City of Austin found an overrepresentation of crashes among users who had ridden e-scooters just a few times compared with those who had ridden more frequently ( 20 ). This result suggests a need for speed governance and potentially enhanced rider training.

Limitations

As is common in surveys, there may be bias from respondents who are more interested in a subject and therefore more likely to participate in related research. Additionally, all of the near misses and crashes reported in this paper were subject to recall bias, which may have been stronger in certain circumstances (e.g., more recent events or scarier events) than others. It is also possible that serious injuries (and certainly deaths) are less likely to be represented in these data. Additionally, the small sample sizes reported for certain subcategories in this paper warrant cautious interpretation; these small sample sizes also precluded further analysis, such as by gender or age, that could provide additional insights into some of the dynamics explored in this paper. Furthermore, a study combining near miss reporting, near miss observations, and crash data could further refine our understanding of the prevalence, type, and safety effect of near misses; including site characteristics in such a study could allow researchers to pinpoint potential effects related to infrastructure. Future research that aims for a larger sample and a mixed-methods approach would further substantiate and nuance the findings presented here. Finally, although the e-scooter statistics described in this paper do not necessarily represent travel in a university setting, it is possible that they reflect that setting more than a survey of city residents.

Conclusions

This paper provides a first look at the prevalence of incidents, crashes, and injuries related to e-scooting, walking, and bicycling within a single sample in the United States. We found different incident and injury dynamics between e-scooter users, pedestrians, and bicyclists, results that can help policy makers and public health and transportation professionals develop more tailored safety strategies going forward. Although future research will strengthen the evidence base, our findings suggest a few key areas for improving roadway users’ experiences. These areas include strategies to provide e-scooters space separate from the sidewalk and to more proactively work to address conflicts between drivers and bicyclists in the roadway. E-scooter speed governance and enhanced training is also likely to lead to improved safety outcomes. Finally, near miss data provide rich insights that complement crash data, and there is a need to develop a more consistent monitoring system to allow for the collection of these data.

Footnotes

Acknowledgements

Thanks to the Bikemaps team for input on the survey design and manuscript, and to A.A., K.C., T.G., and the anonymous reviewers for their comments.

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: R. Sanders, T. Nelson; data collection: R. Sanders; analysis and interpretation of results: R. Sanders; draft manuscript preparation: R. Sanders. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by Dr Trisalyn Nelson’s Foundation Professorship at ASU.

ORCID iD

Rebecca L. Sanders

References

National Association of City Transportation Officials (NACTO). Shared Micromobility in the U.S.: 2018. 2019. https://nacto.org/wp-content/uploads/2019/04/NACTO_Shared-Micromobility-in-2018_Web.pdf. Accessed July 15, 2019.

National Association of City Transportation Officials (NACTO). Shared Micromobility in the US: 2019. National Association of City Transportation Officials, New York City, NY, 2020.

Bureau of Transportation Statistics. Effects of COVID-19 on Bikeshare (Docked and Dockless) and E-Scooter Operations. U.S. Department of Transportation Bureau of Transportation Statistics. 2020. https://data.bts.gov/stories/s/Docked-and-Dockless-and-E-Scooter-System-Changes-2/kar5-6dpn#effects-of-covid-19-on-bikeshare-and-e-scooter-operations-march-through-august-2020.

Shared-Use Mobility Center. Case Study: Status Update, April 8: COVID-19 Crisis Impact on Transit & Shared Mobility. Shared-Use Mobility Center. April 8, 2020. https://learn.sharedusemobilitycenter.org/casestudy/status-update-april-8-covid-19-crisis-impact-on-transit-shared-mobility/.

Portland Bureau of Transportation (PBOT). 2018 E-Scooter Findings Report. 2019. https://www.portlandoregon.gov/transportation/article/709719. Accessed March 3, 2019.

Denver Public Works. Denver Dockless Mobility Program: Pilot Interim Report. 2018. https://denvergov.org/Government/Agencies-Departments-Offices/Agencies-Departments-Offices-Directory/Department-of-Transportation-and-Infrastructure/Programs-Services/Transit/Micromobility-Program. Accessed March 3, 2019.

City of Austin. Dockless Mobility Community Survey Report. 2019. https://www.austintexas.gov/sharedmobility. Accessed June 15, 2019.

Baltimore City Department of Transportation. Dockless Vehicle Pilot Program Evaluation Report. 2019. https://transportation.baltimorecity.gov/sites/default/files/Pilot%20evaluation%20report%20FINAL.pdf. Accessed July 15, 2019.

Sanders

R. L.

Branion-Calles

Nelson

T. A.

To Scoot or not to Scoot: Findings From a Recent Survey About the Benefits and Barriers of Using E-Scooters for Riders and Non-Riders. Transportation Research Part A: Policy and Practice, Vol. 139, 2020, pp. 217–227.

10.

Pimentel

R. W.

Lowry

M. B.

If you Provide, Will they Ride? Motivators and Deterrents to Shared Micro-Mobility. PacTrans, Seattle, WA, 2020.

11.

James

Swiderski

Hicks

Teoman

Buehler

Pedestrians and E-Scooters: An Initial Look at E-Scooter Parking and Perceptions by Riders and Non-Riders. Sustainability, Vol. 11, No. 20, 2019, p. 5591. https://doi.org/10.3390/su11205591.

12.

Sanders

R. L.

We can all get Along: The Alignment of Driver and Bicyclist Roadway Design Preferences in the San Francisco Bay Area. Transportation Research Part A: Policy and Practice, Vol. 91, 2016, pp. 120–133.

13.

Glenn

Bluth

Christianson

Pressley

Taylor

Macfarlane

G. S.

Chaney

R. A.

Considering the Potential Health Impacts of Electric Scooters: An Analysis of User Reported Behaviors in Provo, Utah. International Journal of Environmental Research and Public Health, Vol. 17, No. 17, 2020, p. 6344. https://doi.org/10.3390/ijerph17176344.

14.

Portland Bureau of Transportation and Alta Planning & Design. City of Portland Bureau of Transportation E-Scooter Findings Report. Portland Bureau of Transportation, Portland, OR, 2020.

15.

Chang

A. Y. J.

Miranda-Moreno

Clelow

Sun

Trend or Fad? Deciphering the Enablers of Micromobility in the U.S. SAE International, Warrendale, PA, 2019.

16.

Populus Technologies Inc. Populus Groundtruth Data. Populus Technologies Inc., San Francisco, CA, 2019.

17.

San Francisco Municipal Transportation Agency. Powered Scooter Share Mid-Pilot Evaluation. SFMTA, San Francisco, CA, 2019.

18.

City of Chicago. Chicago E-Scooter Pilot Evaluation Summary. City of Chicago, Chicago, IL, 2020.

19.

Cicchino

J. B.

Kulie

P. E.

McCarthy

M. L.

Severity of E-Scooter Rider Injuries Associated With Trip Characteristics. Journal of Safety Research, Vol. 76, 2021, pp. 256–261.

20.

Austin Public Health. Dockless Electric Scooter-Related Injuries Study. 2019. http://www.austintexas.gov/sites/default/files/files/Health/Web_Dockless_Electric_Scooter-Related_Injury_Study_final_version_EDSU_5.14.19.pdf. Accessed June 3, 2019.

21.

Trivedi

T. K.

Liu

Antonio

A. L. M.

Wheaton

Kreger

Yap

Schriger

Elmore

J. G.

Injuries Associated With Standing Electric Scooter Use. JAMA Network Open, Vol. 2, No. 1, 2019, p. e187381. https://dx-doi-org.web.bisu.edu.cn/10.1001/jamanetworkopen.2018.7381.

22.

Kobayashi

L. M.

Williams

Brown

C. V.

Emigh

B. J.

Bansal

Badiee

Checchi

K. D.

Castillo

E. M.

Doucet

The E-Merging Epidemic of E-Scooters. Trauma Surgery & Acute Care Open, Vol. 4, No. 1, 2019, p. e000337. https://doi.org/10.1136/tsaco-2019-000337.

23.

Farley

K. X.

Aizpuru

Wilson

J. M.

Daly

C. A.

Xerogeanes

Gottschalk

M. B.

Wagner

E. R.

Estimated Incidence of Electric Scooter Injuries in the US From 2014 to 2019. JAMA Network Open, Vol. 3, No. 8, 2020, p. e2014500. https://doi.org/10.1001/jamanetworkopen.2020.14500.

24.

Cicchino

J. B.

Kulie

P. E.

McCarthy

M. L.

Injuries Related to Electric Scooter and Bicycle Use in a Washington, DC, Emergency Department. Insurance Institute for Highway Safety, Arlington, VA, 2020.

25.

Sanders

R. L.

Perceived Traffic Risk for Cyclists: The Impact of Near Miss and Collision Experiences. Accident Analysis and Prevention, Vol. 75, 2015, pp. 26–34. https://doi.org/10.1016/j.aap.2014.11.004.

26.

Dill

Voros

Factors Affecting Bicycling Demand: Initial Survey Findings From the Portland, Oregon, Region. Transportation Research Record: Journal of the Transportation Research Board, 2007. 2031: 9–17.

27.

Shah

N. R.

Aryal

Wen

Cherry

C. R.

Comparison of Motor Vehicle-Involved E-Scooter and Bicycle Crashes Using Standardized Crash Typology. Journal of Safety Research, Vol. 77, 2021, pp. 217–228. https://doi.org/10.1016/j.jsr.2021.03.005.

28.

Sikka

Vila

Stratton

Ghassemi

Pourmand

Sharing the Sidewalk: A Case of E-Scooter Related Pedestrian Injury. The American Journal of Emergency Medicine, Vol. 37, No. 9, 2019, pp. 1807.e5–1807.e7.

29.

Branion-Calles

Nelson

T. A.

Winters

Comparing Crowdsourced Near-Miss and Collision Cycling Data and Official Bike Safety Reporting. Transportation Research Record: Journal of the Transportation Research Board, 2017. 2662: 1–11.

30.

Teschke

Harris

M. A.

Reynolds

C. C.

Winters

Babul

Chipman

Cusimano

M. D.

, et al. Route Infrastructure and the Risk of Injuries to Bicyclists: A Case-Crossover Study. American Journal of Public Health, Vol. 102, No. 12, 2012, pp. 2336–2343.

31.

Winters

Branion-Calles

Cycling Safety: Quantifying the Under Reporting of Cycling Incidents in Vancouver, British Columbia. Journal of Transport & Health, Vol. 7, 2017, pp. 48–53.

32.

Teschke

Frendo

Shen

Harris

M. A.

Reynolds

C. C.

Cripton

P. A.

Brubacher

, et al. Bicycling Crash Circumstances Vary by Route Type: A Cross-Sectional Analysis. BMC Public Health, Vol. 14, No. 1, 2014, pp. 1–10.

33.

Laberee

Nelson

Branion-Calles

Ferster

Winters

Crowdsourced Bicycling Crashes and Near Misses: Trends in Canadian Cities. Urban, Planning and Transport Research, Vol. 9, No. 1, 2021, pp. 450–464.

34.

Stutts

J. C.

Hunter

Police-Reporting of Pedestrians and Bicyclists in Hospital Emergency Rooms. University of North Carolina Highway Safety Research Center, Chapel Hill, NC, 1996.

35.

Hydén

The Development of a Method for Traffic Safety Evaluation: The Swedish Traffic Conflicts Technique. Lund Institute of Technology, Sweden, 1987.

36.

Laureshyn

Application of Automated Video Analysis to Road User Behaviour. Doctoral dissertation. Lund Insitute of Technology, Department of Technology and Society Traffic Engineering, Sweden, 2010.