Document [original]

Felix Wilhelm Siebert, Madlen Ringhand, Felix Englert, Michael

Hoffknecht, Timothy Edwards, Matthias Rötting

Braking bad – Ergonomic design and implications

for the safe use of shared E-scooters

Open Access via institutional repository of Technische Universität Berlin

Document type

Journal article | Accepted version

(i. e. final author-created version that incorporates referee comments and is the version accepted for

publication; also known as: Author’s Accepted Manuscript (AAM), Final Draft, Postprint)

This version is available at

https://doi.org/10.14279/depositonce-12674

Citation details

Siebert, Felix Wilhelm; Ringhand, Madlen; Englert, Felix; Hoffknecht, Michael; Edwards, Timothy; Rötting,

Matthias (2021). Braking bad – Ergonomic design and implications for the safe use of shared E-scooters.

Safety Science, 140, 105294. https://doi.org/10.1016/j.ssci.2021.105294

cbed This work is licensed under a Creative Commons Attribution-NonCommercial- NoDerivatives 4.0

International license: https://creativecommons.org/licenses/by-nc-nd/4.0/

This is the Accepted Manuscript of the following article published by Elsevier in Safety Science [1.

May 2021]:

Siebert, F. W., Ringhand, M., Englert, F., Hoffknecht, M., Edwards, T., & Rötting, M. (2021). Braking

bad–Ergonomic design and implications for the safe use of shared E-scooters. Safety science, 140,

105294. https://doi.org/10.1016/j.ssci.2021.105294

This manuscript is not the copy of record and may not exactly replicate the final, authoritative

version of the article.

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0

International License, http://creativecommons.org/licenses/by-nc-nd/4.0/.

Braking Bad - Ergonomic Design and Implications for the Safe Use of Shared E-Scooters

Keywords: micromobility; e-scooters; naturalistic observation; brake ergonomics

Siebert, F. W., Ringhand, M., Englert, F., Hoffknecht, M., Edwards, T., & Rötting, M.

Abstract

Shared e-scooters are introduced as a new form of mobility around the world. Alongside this rise in

micromobility, e-scooter crashes are reported, and e-scooter riders are injured and killed in traffic.

Little research has been conducted on the relation between ergonomics and the safe use of e-scooters,

and it is unclear whether e-scooter riders know about prevailing e-scooter related regulation and if

they adhere to existing regulation in traffic. We conducted a field observation (n=2972) in combination

with a questionnaire survey (n=156), to investigate the influence of ergonomics on the safe use of

shared e-scooters, and to explore riders’ knowledge and self-reported behavior. Riders’ brake

readiness, dual use (two riders per vehicle), and helmet use was registered, and specific knowledge

about the braking system of e-scooters was assessed, alongside knowledge about road rules and

reported past safety related behavior. Results reveal a clear effect of braking system design, with

significantly more riders readying the left hand brake, in comparison with the right hand or foot brake

(depending on the e-scooter model). This was found regardless of the brake-lever-to-wheel coupling,

indicating that the preference for the left hand brake can be detrimental to targeted braking of the

front or rear wheel. Only one third of respondents could correctly identify the basic braking system of

the shared e-scooter they had last used. In addition, high shares of illegal behavior were reported by

riders. Implications of these findings for the safe operation of e-scooters, their ergonomic design, and

road safety regulation are discussed.

1. Introduction

In a very short timeframe, the introduction of shared e-scooters has changed the mobility landscape

in countries around the globe (Gössling, 2020). At the same time, researchers find increased rates of

hospitalization of e-scooter users (Namiri et al., 2020; Trivedi, B. et al., 2019) with a high frequency of

head injuries (Aizpuru et al., 2019; Trivedi, T. K. et al., 2019). A plethora of potential compounding

factors in e-scooter crashes and resulting injury severity have been identified. Researchers have found

that between 16% and 36% of e-scooter riders arriving at hospitals for treatment of injuries are under

the influence of alcohol (Badeau et al., 2019; Bekhit, Le Fevre, & Bergin, 2020; Blomberg, Rosenkrantz,

Lippert, & Collatz Christensen, 2019; Puzio et al., 2020). Riders have been observed to travel against

the direction of traffic (7% on roadways in Los Angeles & Santa Monica, USA: Todd, Krauss,

Zimmermann, & Dunning, 2019). In countries where helmets are not mandatory for e-scooter usage,

only a small share of riders uses a helmet (2% in San Jose, USA: Arellano & Fang, 2019; 6% before

electric scooter helmet law in Los Angeles & Santa Monica, USA: Todd et al., 2019; 3% in Vienna, Austria:

Mayer, Breuss, Robatsch, Salamon, & Soteropoulos, 2020). To a smaller extent, the practice of dual

use of e-scooters (two riders standing on one vehicle) has been observed, interfering with their safe

use (2% in Brisbane, Australia: Haworth & Schramm, 2019; 1% in Los Angeles, USA: Todd et al., 2019;

3% in Vienna, Austria: Mayer et al., 2020).

Germany was one of the last high-income countries to allow shared e-scooters on its streets in the

Summer of 2019. To regulate this new form of mobility, Germany has enacted the

Elektrokleinstfahrzeuge-Verordnung (eKFV, engl. decree for small electric vehicles), in which technical

requirements for e-scooters as well as other regulatory boundaries are specified. Despite the

implementation of the eKFV, increasing numbers of e-scooter rider hospitalization have been found in

Germany (Störmann et al., 2020; Uluk et al., 2020).

Despite sustained international research on the safety of e-scooters, to date there is relatively little

research on ergonomic aspects of e-scooters, although ergonomic aspects play a substantial role in the

safe operation of other modes of transport (Bhise, 2012; Hawkins, 2006; Oppenheim & Shinar, 2011).

Hence, the goal of this study is to investigate the ergonomics of the brake systems of shared e-scooters

in Germany and their potential influence on riders’ safety. In addition, the knowledge of e-scooter

users about current regulations in the eKFV and related rider behavior is analyzed. To this end, a

combination of a video-based observation of and a questionnaire survey of e-scooter users in

Germany’s capital and biggest city Berlin was conducted.

2. Background

2.1. Regulation of e-scooters in Germany

As there is no global regulatory framework for the introduction of e-scooters, countries and cities have

enacted different sets of rules and regulations to increase the safety and safe use of e-scooters. The

German eKFV mandates 14 years as the minimum age for using an e-scooter in Germany, and no

driver’s license of any kind is needed (eKFV §3). E-scooters maximum speed is limited to 20 km/h (eKFV

§1 (1)), with faster e-scooters falling out of the eKFV’s scope. A bell/ acoustic signaling is required (eKFV

§6), as well as appropriate lighting and reflectors (eKFV §5). Levers for the regulation of motor power

(i.e. acceleration), are required to be self-resetting to zero-acceleration after a maximum of one second

(eKFV §7 (7)). Dual use is not permitted (eKFV §8). For road infrastructure, e-scooters are obligated to

follow the rule of the road (right hand traffic, eKFV §11 (2)), and use dedicated bicycle infrastructure

or mixed pedestrian-bicycle infrastructure within cities when it is available (eKFV §10 (1)). When no

dedicated bicycle or bicycle-pedestrian infrastructure is available, e-scooters are permitted to use the

road (eKFV §10 (1)). If there is no mechanism for indicating turns on the e-scooter, riders are required

to use their hands for turn signaling (eKFV §11 (3)). For driving under the influence of alcohol, the same

limits apply as in car use, it is illegal to drive with a blood alcohol concentration of 0.5 ‰ or higher

100

(Straßenverkehrsgesetz, StVG §24a). Since Germany employs a graduated drivers license system and

101

age-adjusted regulation, this general alcohol limit is lower (0.0 ‰ blood alcohol concentration) for e-

102

scooter riders under the age of 21 and novice drivers (license for less than three years, StVG §24a). All

103

e-scooters in Germany need to be equipped with two separately actuated brakes, which individually

104

achieve a deceleration of at least 3.5m/s2 (eKFV §4). This requirement does not necessitate that both

105

the front and back wheel are equipped with a brake, it is sufficient when two independent levers

106

actuate two independent brakes on one wheel. In addition, eKFV §4 (1) references §65 (1) of the

107

general German road safety regulation (Straßenverkehrs-Ordnung, StVO) in which an “adequate brake

108

that can be easily operated while driving” is mandated.

109

2.2. Braking system of e-scooters

110

A research need for the braking systems of the many available shared e-scooter models has been

111

identified (Garman et al., 2020), but braking systems of e-scooter models have not been researched in

112

detail. During the time of this study, six shared e-scooter providers were active in Berlin: Bird, Circ,

113

Jump, Lime, Tier, and VOI (Kraftfahrtbundesamt, 2019). All provided e-scooter models fulfill the

114

requirement of two independent braking systems, although their braking systems differ in brake lever

115

placement as well as lever-to-wheel coupling. While some models provide two hand lever brakes on

116

the handlebars of the scooter (Bird, Circ, Jump, Tier), other models are equipped with a foot-brake in

117

addition to a single left hand brake (Lime, Voi) (Figure 1). While all models are equipped with a hand

118

brake lever on the left side of the handle bar, for two models this lever actuates the front wheel (Circ,

119

Lime), for the other four (Bird, Jump, Tier, Voi) it actuates the back wheel.

120

121

Figure 1. Handlebar of a Tier e-scooter equipped with two hand-lever brakes and a highlighted

122

acceleration thumb-lever (left) and Lime scooter with single left-hand lever brake and foot brake for

123

the back wheel (right).

124

For four of the e-scooter models (Bird, Jump, Lime, Tier), one brake lever is coupled to the front wheel

125

and one to the back wheel, allowing the application of brake-power to both wheels. For two e-scooter

126

models (Circ, Voi), both brake levers are coupled to the same wheel, limiting brake-power application

127

to a single e-scooter wheel (Circ: front wheel; Voi: back wheel). Details on the brake systems are

128

presented in Table 1. For acceleration, all e-scooter models use a variant of a thumb-lever on the right

129

side of the handlebar (Figure 1). This acceleration lever does not lock in position and needs to be

130

constantly actuated to keep the e-scooter moving, with non-actuation leading to the deceleration and

131

stop of the e-scooter after a short time (as required by the eKFV).

132

Table 1. Brake system architecture of the six e-scooter models active in Berlin during the time of this

133

study.

134

Bird

Circ

Jump

Lime

Tier

VOI

E-scooter

model

Bird one

Germany

B1D

ES 200D

Lime-S 3.0

ES 200G

Voiager 1

Front wheel

brake

Right brake

lever

Left and right

brake lever

Right brake

lever

Left brake

lever

Right brake

lever

None

Back wheel

brake

Left brake

lever

None

Left brake

lever

Foot-brake

Left brake

lever

Left brake

lever and

foot-brake

135

Since e-scooters are relatively new, little research has been conducted on riders braking behavior and

136

preferences, as well as general braking efficiency. Investigating brake force application, Bierbach et al.

137

(2018) investigated the braking properties of various micromobility vehicles. With a maximum

138

deceleration of approx. 3.1 m/s2 the e-scooter used in the study (Egret One V3 – two hand brake levers)

139

performed relatively poorly in comparison with a bicycle (on average 6,5 m/s²) and a Segway (on

140

average 4,5 m/s²). For two wheelers in general, there is a difference in efficiency between front and

141

back wheel braking. The act of braking on a two-wheeler shifts the dynamic wheel load towards the

142

front wheel, hence the front wheel can exert a higher braking force on the ground than the back wheel

143

before slipping occurs between the wheel and the ground (Wilson, Schmidt, & Papadopoulos, 2020;

144

Wolff, 2017). Hence stronger deceleration can be achieved by using the front wheel brake on bicycles

145

(Beck, 2004; Mordfin, 1975; Wilson et al., 2020) although the amount of deceleration further depends

146

on the applied force on the brake lever and braking both wheels is advantageous to single wheel

147

braking (Huertas-Leyva, Dozza, & Baldanzini, 2019). Countries have differing regulations on hand-lever-

148

to-wheel coupling for bicycles, with Germany not regulating which lever actuates which brake. There

149

are no studies on hand lever preferences for braking bicycles.

150

There are no studies on e-scooter related preferences for hand or foot brake lever usage or ergonomics,

151

although it can be assumed, that using the foot brake necessitates more preparation, as the riders

152

need to shift their center of gravity to use the foot brake, while the hand brake is close to the

153

handlebars and “within reach” during normal driving.

154

155

Several challenges arise from the brake lever placement and design of shared e-scooter models active

156

in Germany. As a general issue, the novelty of e-scooters together with the dissimilarity of brake

157

actuator placement, either only as hand-lever brakes or as a combination of hand- and foot-actuation,

158

prohibits the development of conformity to user expectations (DIN Deutsches Institut für Normung

159

e.V., 2009). Hence, brake placement will have to be learnt and remembered for each individual e-

160

scooter model, since a universal mental model for lever-to-brake coupling will be incorrect for some

161

e-scooter models.

162

As a similar problem, the lack of a universal mental model for braking can lead to confusion about lever

163

and front-/back-wheel-brake coupling. Since front- and back-wheel braking produces different brake

164

forces, such confusion could in theory lead to an inadequate application of brake force. An additional

165

ergonomics challenge arises for e-scooter models equipped with a hand-lever brake on the right side

166

of the handlebar. The (eKFV mandated) need for continuous operation of the thumb-actuated throttle-

167

lever could impede the successive actuation of the right hand brake lever. While e-scooter models

168

equipped with a foot brake are not subject to this issue, their brake-mechanism necessitates a

169

repositioning and lifting of the back foot to actuate the back wheel brake, involving a repositioning of

170

the whole body on the relatively narrow e-scooter floorboard. In addition, the foot brake is rendered

171

inaccessible in cases of dual use in which the non-driving riders stands in the back of the e-scooter.

172

173

174

175

2.3. Aims of this study

176

The aim of this study is to investigate traffic safety related knowledge and behavior of e-scooters as

177

well as brake readiness in Berlin, Germany. Two hypotheses are put forward:

178

1. E-scooter riders are unfamiliar with the braking systems of the e-scooters they use, and hence are

179

unable to correctly identify which brake actuator is coupled with which wheel.

180

2. For brake preparation movements, such as riders putting their hand on the brake lever or positioning

181

their foot over the foot brake for a faster brake reaction, we expect that the right hand-lever and the

182

foot brake will be observed to have significant lower brake readiness than the left hand-lever brake,

183

regardless of brake-lever-to-wheel-coupling.

184

Apart from these two braking-related hypotheses, a further aim of this study is to collect additional

185

data on the state of knowledge of riders on the prevalent road regulation for e-scooters and observe

186

e-scooter dual and helmet use. In contrast to the brake related research hypotheses, this data

187

collection and analysis is exploratory in nature.

188

189

3. Method

190

To test our hypotheses and assess additional data on e-scooter riders knowledge about prevalent road

191

related regulation, a naturalistic observation study of e-scooter users was conducted together with a

192

quantitative questionnaire survey at three survey sites in Berlin, Germany, between 21. September

193

and 13. October 2019. In the fall of 2019, Berlin had the largest number of active e-scooters in Germany,

194

with more than 11,000 deployed shared e-scooters which are used for an average of three rides a day

195

(Tack, Klein, & Bock, 2020). The resulting three survey sites are presented in Figure 2. The observation

196

parameters will be described first, followed by methodological details of the questionnaire survey.

197

198

Figure 2. The three survey sites (including latitudinal and longitudinal coordinates) for observation and

199

questionnaire distribution in Berlin (© OpenStreetMap contributors).

200

201

Site 1:

52.504082, 13.337187

Site 3:

52.502722, 13.446836

Site 2:

52.512485, 13.377054

3.1. Observation

202

A camera-based observation was conducted at the three survey sites to register e-scooter riders’

203

behavior on the street. As the General Data Protection Regulation enacted by the European Union

204

(2016), defines a number of rules and restrictions for data collection in the public space, the

205

observation framework was developed in collaboration with the data security officer

206

(Datenschutzbeauftragte in German) of the [name of university]. Through this consultation, the video-

207

based observation was planned in a way that minimizes the amount of personal data that is collected.

208

The positioning of the observation cameras and the resulting viewing angles support these efforts, as

209

they minimize the recording of road users' faces as much as possible, while still allowing the

210

observation of e-scooter riders.

211

212

3.1.1. Observation sites

213

The sites for the observation were chosen based on two factors. During the time of the study, six

214

shared e-scooter providers were active in Berlin (see Table 1), covering different areas of service for e-

215

scooter rental. Observation sites were selected in places where all six providers were active during the

216

time of the study. As a second objective for the identification of survey sites, the frequency of e-scooter

217

traffic was considered, leading to the installation of cameras in the general vicinity of transport hubs,

218

while maintaining enough distance to presume independence of observations. The distance between

219

all sites is a minimum of 3.4 kilometers, well outside of the average travel range of e-scooter users of

220

approximately 2 kilometers (Bai & Jiao, 2020; Tack et al., 2020). Two video cameras were used to

221

collect video data of riders’ behavior. The cameras were enclosed in a grey waterproof box and

222

223

recording duration of approximately 14 hours. Videos were recorded with a resolution of 1920x1440

224

pixels and a frame rate of 30fps. Using two straps, the cameras were attached to lampposts at the

225

observation sites at a height of 4-5 meters. In accordance with the aim of limiting the recording of

226

personal data such as riders’ faces, the cameras filmed almost straight downwards. Sample frames

227

from the observation are presented in Figure 4 and Figure 7. The total recording duration was 274.5

228

hours (83.5 hours at site 1, 83.5 hours at site 2, and 131 hours at site 3), with recordings mainly

229

conducted between 12:30 pm and 02:30 am. At all sites, an information sheet was posted, informing

230

passersby of the ongoing observation.

231

232

Figure 3. Representation of camera viewing angle and position (left) and picture of camera position

233

(right).

234

235

3.1.2. Observation variables

236

Using the recorded video data, seven variables were registered using the software BORIS (Behavioral

237

Observation Research Interactive Software, Friard & Gamba, 2016) for each observed shared e-scooter

238

(private e-scooters were not registered). Variables and available codes for each variable are listed in

239

Table 2. Direction of travel refers to the fact that in Germany there is only one “correct” direction for

240

riding on a cycle path (right-hand traffic), unless an explicit exemption is made, which was not the case

241

for the observation sites in this study. Dual use driver position refers to the rider in control of the

242

handle bars, who can stand either in front of the scooter (with the passenger in the back, Figure 4) or

243

in the back of the scooter (with the passenger in the front). The registration of hand and feet position

244

for the assessment of brake-readiness will be explained in the following.

245

246

Figure 4. Observed dual use with the driver in the front position.

247

248

249

250

251

252

253

Table 2. Observational variables and available codes per variable.

254

Variable

Available codes

Scooter provider

Bird; Circ; Jump; Lime; Tier; Voi

Direction of travel

Correct; Incorrect

Helmet use

Yes; No; Not identified

Dual use

Yes; No; Not identified

Dual use driver position

Driver in front; Driver in the back; Not identified

Hand position (per lever)

Brake-ready; Not brake ready; Not identified

Feet position (for e-scooters

with footbrake)

Brake-ready; Brake-prepared; Not brake ready; Not identified

255

As shown in Table 1, the e-scooter models supplied by sharing providers in Berlin are equipped with

256

different braking systems, with some models being equipped with two hand-lever-brakes (Bird, Circ,

257

Jump, Tier) and other models being equipped with one hand-lever-brake and one foot-brake (Lime,

258

Voi). Hence, to identify brake readiness of e-scooter users, riders’ hand and feet position was analyzed.

259

For hand-brake levers, brake-readiness was defined as follows: if at least one digit of a hand was placed

260

on the brake lever, the individual brake was registered as “brake-ready”. For e-scooters that have two

261

hand-brake levers, this coding is enough to assess brake readiness for both brake levers of an individual

262

e-scooter. Examples of brake-ready and non-brake-ready hand positions are presented in Figure 5.

263

264

Figure 5. Cropped examples for brake readiness coding of hand lever brakes, not brake ready (left) and

265

brake ready (right).

266

267

For e-scooter models with a foot lever brake on the back wheel, brake readiness was assessed by

268

classifying the feet position on the floorboard of the e-scooter. If the two feet were placed in parallel

269

to each other, with a lateral overlap of more than 25%, the feet position was registered as non-brake

270

ready. If the feet of a driver were positioned so that their lateral overlap was equal to or less than 25%,

271

the code “brake prepared” was registered, as this position allows a quicker brake reaction than a

272

parallel feet position, although it is still necessary to reposition the braking foot to actuate the foot

273

lever brake. Full brake readiness for the foot-brake was registered when the two feet overlapped by

274

25% or less (as in “brake prepared"), but in addition the heel of the back foot was raised, allowing a

275

quick actuation of the foot brake. Examples for all three brake readiness positions for the foot brake

276

are presented in Figure 6. To allow a direct comparison of brake readiness for hand and foot brakes,

277

the “brake prepared” position is counted as “not brake ready” in the analysis.

278

279

Figure 6. Cropped examples of brake readiness coding for foot lever brakes, non-ready on the left,

280

preliminary readiness in the middle, and brake ready on the right.

281

282

Because of the restricted viewing angle, caused by the top down camera position (required due to the

283

European data privacy regulation), some caveats apply to the registration of the observational

284

variables listed in Table 2. Within the camera’s view, only some parts of the street’s infrastructure are

285

covered, so no general assumptions can be made on the use of a specific infrastructure for the whole

286

street. The small timeframe in which e-scooters are visible in the camera frame does not allow a

287

distinction between brake readiness and actual braking, as changes in speed cannot be reliably

288

assessed. However, we argue that hand and feet positioning for both, actual braking and brake-ready

289

hand and feet positions, can give insights into the general usage of the braking systems installed on

290

the scooters. Additional challenges for the registration of variables are present in the video data, as e-

291

scooters are sometimes not completely visible within the viewing angle of the camera or riders are

292

blurred due to poor lighting, leading to an inability to register variables such as helmet use and hand

293

position. Examples of this are presented in Figure 7. In these instances, all observable variables are still

294

registered, and “not identified” is registered for non-observable variables.

295

296

297

298

Figure 7. Examples of blurred video and e-scooter riders partly out of the video frame.

299

300

3.2. Questionnaire survey

301

The questionnaire survey was directly administered on a computer tablet at the three survey sites

302

(Figure 2) around Berlin by the authors from noon to early evening hours. In addition, small paper

303

notes with a link and a QR-code to an online version of the questionnaire were distributed at the survey

304

sites and at the [name of university]. Participation on-site versus participation through the QR-code

305

was not registered. The only prerequisite for participation in the survey was prior use of a shared e-

306

scooter and there was no compensation for participation.

307

3.2.1. Participants

308

In total, N=156 people took part in the questionnaire survey (46=female; 107=male; 1=non-binary;

309

2=no answer) between the beginning of November and the middle of December 2019. The mean age

310

of respondents was M=22.7 (SD=5.7). While 77% reported to live in Berlin, 19% reported to live in a

311

different German city and another 5% abroad. In line with the prerequisite for participation in the

312

survey, all respondents had used a shared e-scooter at least once. Of the 156 respondents, 62% (n=97)

313

had used a shared e-scooter for three rides or less, 26% (n=40) had used an e-scooter once a month,

314

8% (n=12) used it once a week, 3% (n=5) used it multiple times per week, and only 1% (n=2) reported

315

daily e-scooter use. Asked in which city they had mainly used a shared e-scooter, the majority of

316

respondents (71%) named Berlin (n=111) while an approximate third of respondents (29%) placed their

317

main use in another town (n=45). For 56% (n=87) the last e-scooter ride before the survey was more

318

than a month ago, for 24% (n=37) it was within the last 30 days, for 17% (n=26) it was within the last 7

319

days, and 4% (n=6) had used an e-scooter on the day of the survey.

320

3.2.2. Materials

321

The questionnaire consisted of 33 questions in total. To allow participation of non-German native

322

speakers, the German version of the questionnaire was translated by the authors to produce an English

323

version. Of all respondents, n=134 used the German version, and n=22 used the English version. The

324

questionnaire contains questions on the demographics of respondents, their general e-scooter use,

325

their adherence to and knowledge about safety related regulation, and questions about the braking

326

system of the e-scooter they had last used. The English items of the questionnaire can be found in the

327

result section in the corresponding tables. The order of the items in the original presentation results

328

from the numbering in Table 3 to 6.

329

330

4. Results

331

4.1. Observation

332

Within the 274.5 hours of video data, a total of 2972 e-scooters were observed. The main scooter

333

provider at the three survey sites was Lime (n= 2143), followed by Tier (n= 391), Voi (n= 316), Jump (n=

334

70), Circ (n= 34), and Bird (n= 18). The majority of scooters was observed on a bicycle lane (n= 2113;

335

71%), followed by the street (n=670; 23%), and the sidewalk (n= 174; 6%), with infrastructure not

336

identified for n=15 (0.5%) e-scooters. Of all scooters, n= 163 (6%) were driven against the direction of

337

traffic illegally within the view of the camera. Dual use was observed for n= 92 scooters (3%), with 67

338

occurrences on Lime scooters, 19 on Tier, 4 on Voi, and 2 occurrences on Jump e-scooter models. Only

339

n= 8 riders (0.3%) were observed to use a helmet, while non-helmet use was observed in n= 2670

340

instances (not-identified n= 386 (13%)).

341

Since every observed e-scooter model has two brake levers which can actuate one or two wheels,

342

brake readiness is first presented in relation to the levers on each e-scooters, regardless of the lever-

343

to-wheel coupling. Figure 8 shows the observed lever-based brake readiness for all six observed e-

344

scooter models. Since only observed e-scooters with complete available brake-data are analyzed, the

345

sample size (n=2082) is smaller than that of all observed e-scooters (n=2972), as for n=890 e-scooters

346

(30%) at least one variable for brake-readiness detection is missing. For all registered variables, the

347

rate of non-identification increased during evening hours (Figure 9).

348

349

Figure 8. Lever-based brake readiness observed on e-scooter models of the six providers.

350

For three e-scooter models (Bird, Circ, and Jump), the majority of users is not brake-ready, i.e. users

351

have not positioned their hands for a quick actuation of a brake lever. For the other three e-scooter

352

models (Lime, Tier, and Voi), the majority of users is brake-ready with one brake lever. The highest

353

lever-based brake-readiness for two brakes was observed for Jump e-scooters, where 23% of riders

354

have both brakes ready, followed by Tier (15%), and Lime (10%). The lowest average number of brake

355

levers readied is observed for Circ e-scooter (0.2 levers readied per e-scooter), followed by Bird (0.6

356

levers), Voi and Lime (both 0.7 levers); Tier (0.7 levers), and Jump (0.8 levers). For an assessment of

357

minimum brake readiness, all e-scooters with at least one brake readied are grouped (i.e. “one brake”

358

and “two brakes” observations in Figure 8 are added). Minimum brake readiness differs significantly

359

between observed e-scooter providers (

2(5)=18.23; p<.01). Fisher’s exact test with Bonferroni

360

correction for multiple comparisons reveals significant differences between minimum brake readiness

361

of Circ scooters in comparison with Jump, Lime, Tier, and Voi scooters (all p<.0033).

362

363

10%

20%

30%

40%

50%

60%

70%

80%

90%

Bird Circ Jump Lime Tier Voi

(n=18) (n=21) (n=60) (n=1447) (n=313) (n=223)

Brake-readiness

E-scooter provider

no brake

one brake

two brakes

364

Figure 9. Number of observed e-scooters throughout the day, split for e-scooters where all variables

365

from Table 2 were registered, and those where at least one variable from Table 2 was not identified.

366

To look at brake readiness in more detail, Figure 10 shows the distribution of brake lever usage for all

367

riders with brake-readiness for one brake lever (n=1043). This brake readiness is of special interest, as

368

riders chose to ready one lever instead of the other, while riders readying no brake or both brakes

369

cannot be observed to prefer on lever over the other. Among all shared e-scooter riders that were

370

observed to ready one brake, the majority readies the left hand lever brake (n=821), while the right

371

hand lever brake and the foot brake are readied less often (n=222, see Figure 10). This difference is

372

significant, i.e. left hand brake readying is significantly higher than 50% (z=18.52; p<.001). Observations

373

were grouped to investigate whether riders on e-scooters with two hand brakes (Bird, Circ, Jump, Tier)

374

show differences in brake readiness compared to e-scooters with one hand and one foot brake (Lime,

375

Voi). For this analysis, left hand brake readiness was compared to “other lever” brake readiness (foot

376

brake or right hand brake) between the two types of brake system. A significant difference was found

377

in the share of left hand lever brake readying between e-scooters with two hand brakes, and e-scooters

378

with combined hand and foot levers (

2(1)=19.86; p<.001). Riders on two hand brake e-scooters had a

379

more balanced ratio of left hand vs other lever brake readying (66% left hand vs 34% right hand),

380

compared to riders on hand and foot brake scooters (81% left hand, 19% foot brake), which had higher

381

brake readying for the left hand brake compared to other lever brake readying.

382

100

150

200

250

300

350

400

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Number of e-scooters observed

Time of day

all

variables complete

variables incomplete

383

Figure 10. Observed distribution of brake levers for e-scooters where one brake lever is ready.

384

Apart from lever-based brake readiness (Figure 8 and Figure 10), wheel-based brake-readiness can be

385

assessed by mapping the available brake levers to the front and the back wheel brake of scooters (using

386

the information from Table 1). The resulting distribution is presented in Figure 11 for riders with a

387

brake readiness of one lever, as these riders have (knowingly or unknowingly) chosen to use one brake

388

lever which brakes an individual wheel over the other one. Descriptively, no overall pattern of front vs.

389

back wheel braking can be observed. For Bird, Jump, and Tier scooters, a tendency for back wheel

390

braking (left hand lever actuated) can be observed. For lime scooters, a strong tendency for front wheel

391

braking can be observed (likewise actuated with the left-hand lever). For the Circ e-scooter model, all

392

braking is front wheel braking, as the scooter model does not have a back wheel brake and both levers

393

actuate the front wheel brake. Similarly, for the Voi e-scooter model, all braking is back wheel braking.

394

Despite the fact that for Circ and Voi e-scooters, the same wheel is actuated with different levers,

395

Figure 8 shows that 5% (Circ) and 9% (Voi) of their respective users ready two levers for potential

396

braking, i.e. they ready two levers to brake the same wheel. This indicates that these riders are

397

unaware of the lever-to-wheel coupling of their e-scooter.

398

For those e-scooter models that allow braking of the front and back wheel (Bird, Jump, Lime, Tier), we

399

investigated whether there is a significant difference between providers and front wheel brake

400

readiness in riders who ready one brake. Since the expected value of one cell in the contingency table

401

was smaller than 5, Fisher’s exact test was used. The test revealed a significant difference between

402

providers in the share of brake readying of the front wheel (p<.001). To test which providers differ in

403

the observed readying of the front brake, Fisher’s exact test was used to compare individual providers,

404

with Bonferroni correction for multiple comparison. Front wheel brake readying of observed lime

405

scooters differed significantly from front wheel brake readying of Bird, Jump, and Tier scooters (all

406

p<.0083).

407

25%

50%

75%

100%

Bird Circ Jump Lime Tier Voi

(n=8) (n=3) (n=20) (n=746) (n=143) (n=123)

Brake lever readiness (%)

E-scooter provider

foot brake

right brake

left brake

408

Figure 11. Observed distribution of wheel-based brake readiness where one brake lever is ready.

409

For dual use (n=92), 53 cases were observed in which the driver is in the front position with the

410

passenger in the back of the scooter, and 39 cases were observed where the driver stands in the back

411

of the scooter, reaching around the passenger to control the scooter. In addition to being illegal under

412

the German Law, dual use can impact the ability to use a foot brake, if the driver is positioned in the

413

front of the e-scooter, as drivers’ access to the foot brake is blocked by the passenger (Figure 4). There

414

were 44 occurrences of dual use where the foot brake was blocked by a passenger (1.5% of

415

observations). For Lime scooters, there are 42 instances of dual use with the driver in the front, and

416

25 instances with the driver in the back. For Voi, two instances of dual use were observed with the

417

driver in the front position, and 2 instances with the driver in the back position.

418

4.2. Questionnaire

419

All questionnaire data is presented in Table 3, Table 4, Table 5, and Table 6 showing individual

420

questions, answering options, and percentage results of answers. The original order of items can be

421

reconstructed by the numbering of the items. Results of the questions 1, 2, 5, 6, 7 and 10 are presented

422

in section 3.2.1 – the characteristics of participants.

423

4.2.1. Driving history, and self-reported feeling of safety

424

Results regarding driving history, helmet use and self-reported feeling of safety are presented in Table

425

3. Respondents’ most frequently used e-scooter provider was Lime (60%), followed by Tier (24%), Voi

426

(6%), Circ (3%), Bird (2%), and Jump (1%), which is broadly comparable to the e-scooter provider

427

distribution in our observational study. As similar distribution was found for the provider used during

428

the last ride and the providers used in the past. The majority of respondents (62%) had only used one

429

shared e-scooter provider, while 39% had used more than one shared e-scooter provider in the past.

430

On helmet use, nearly all respondents report to never use a helmet on an e-scooter. However, around

431

half of helmet non-users indicated that they would potentially use a helmet if it was provided by the

432

25%

50%

75%

100%

Bird Circ Jump Lime Tier Voi

(n=8) (n=3) (n=20) (n=746) (n=143) (n=123)

Wheel brake readiness (%)

back wheel

front wheel

e-scooter provider, 32% indicate potential helmet use if it was mandatory by law, and 33% report

433

neither of the two measures would encourage them to use a helmet. Almost half of respondents

434

indicate that their e-scooter use would decrease if there was a mandatory helmet use law.

435

One-tenth of respondents reported to have experienced a fall or a collision with another road user

436

while using an e-scooter in the past. Crashes were mainly ascribed to a bad road surface, distraction,

437

loss of control over the e-scooter, or going too fast.

438

Asked to rate how safe they generally feel when riding an e-scooter on a scale from 1 (very unsafe) to

439

7 (very safe), the average ratings of respondents was m=3.95 (SD=1.5). For comparison, respondents

440

rated their perceived safety while riding a bicycle as m=5.61 (SD=1.1) which was significantly higher

441

than their perceived safety on an e-scooter (t(155)= -11.68; p<.001). When respondents were asked to

442

choose the safest road infrastructure, bicycle lanes performed best, followed by sidewalks and streets.

443

The question regarding mostly used road infrastructure showed a high usage of bicycle lanes followed

444

by streets and sidewalks.

445

446

Table 3. Survey questions and answers for driving history, helmet use and self-reported feeling of

447

safety (% of answers for n=156 respondents). All items are single choice unless indicated otherwise.

448

Question no.

Answering options

3. Which e-scooter sharing

companies have you used

before? (multiple answers)

3.8%

17.9%

75.6%

39.1%

6.4%

16.0%

3.8%

Bird

Circ

Lime

Tier

Uber/Jump

Voi

other

4. Which e-scooter sharing

company do you use most often?

60%

24%

Bird

Circ

Lime

Tier

Uber/Jump

Voi

other

11. Which e-scooter sharing

company did you use during your

last ride?

1.3%

3.2%

62.2%

21.2%

1.9%

6.4%

3.8%

Bird

Circ

Lime

Tier

Uber/

Jump

Voi

Other/don’t

remember

8. On which road infrastructure

have you ridden an e-scooter?

(multiple answers)

85.3%

69.2%

76.9%

bicycle lane

sidewalk

street

9. Where do you ride the e-

scooter the most?

58.3%

15.4%

26.3%

bicycle lane

sidewalk

street

15. Where do you feel the safest

when riding an e-scooter?

76.9%

17.9%

5.1%

bicycle lane

sidewalk

street

16. Do you wear a helmet when

riding an e-scooter?

98.1%

0.6%

0.0%

0.6%

never

rarely

sometimes

often

always

17. If you don't always wear a

helmet, what would encourage

you to wear a helmet more

often? (multiple answers)

32.1%

51.3%

32.7%

mandatory helmet law

helmet provided by sharing

company

neither

18. Would your use of e-scooters

decrease, if wearing a helmet

was required by law?

20.5%

33.3%

46.2%

I don't know

yes

23. How safe do you feel on an

e-scooter?

5.1%

14.1%

21.8%

18.6%

25.0%

10.9%

4.5%

(very

unsafe)

(very safe)

24. How safe do you feel on a

bicycle?

0.0%

1.3%

2.6%

12.2%

23.1%

39.7%

21.2%

1 (very

unsafe)

7 (very safe)

25. Did you ever fall or collide

with another road user while

using an e-scooter?

9.6%

90.4%

yes

26. If yes, what was the reason

for the accident? (multiple

answers)

13.3%

0.0%

40.0%

0.0%

20.0%

loss of control

over the e-

scooter

I was going

too fast

brake(s) of e-

scooter too

weak

road surface

was in a bad

condition

other road

users were

reckless

I was

distracted by

my phone

449

4.2.2. Knowledge about brake-system

450

Table 4 shows the results of the questions regarding the knowledge about the brake-system of e-

451

scooters. Participants were asked to think back to their last e-scooter ride and indicate if the e-

452

scooter had one brake (i.e. for one wheel) or two brakes (for two wheels). An answering option for

453

two brakes that both decelerate one wheel was erroneously not included, hence respondents who

454

used e-scooters with such a brake system (Circ: n=3%; Voi: n=6%) were excluded from the following

455

analysis, as were respondents that could not remember which e-scooter provider they used last

456

(n=4%). Of all remaining respondents, 34% correctly identified that their last used e-scooter had two

457

brakes (one for the front and one for the back wheel), while 26% of respondents falsely assumed that

458

their e-scooter model had just one brake, and 40% did not know if their e-scooter had one or two

459

brakes. Asked which brakes they normally use, 31% named the rear brake, 26% named the front

460

brake, 16% reported to usually use both the front and rear brake, and 27% answered that they did

461

not know which brake they normally use. Asked how they would intuitively brake the back wheel of

462

an e-scooter, 22% of respondents would use the left brake lever on the handle bar, 45% would use

463

the right hand brake lever, and 33% would use a back wheel footbrake.

464

Table 4. Survey questions and answers for knowledge about brake-system (% of answers for n=156

465

respondents). All items are single choice unless indicated otherwise.

466

Question no.

Answering options

19. Please think back to your last

ride with an e-scooter. How many

brakes did this particular model

have and which wheels were

decelerated?

6.4%

16.0%

4.5%

34.6%

38.5%

1 brake, applies

braking force to

the front wheel

1 brake, applies

braking force to

the back wheel

1 brake, applies

braking force to

both wheels

2 brakes, one for

the front wheel

and one for the

back wheel

I don't know

20. Which brake(s) do you

normally use?

26.9%

16.0%

30.8%

26.3%

I don't know

front and rear brake

rear brake

front brake

21. Assuming you are using an e-

scooter equipped with a brake for

the rear wheel, how would you

intuitively use it?

21.8%

44.9%

33.3%

left brake lever on handlebar

right brake lever on handlebar

using my foot to press down

on the brake over the rear

wheel

467

4.2.3. Knowledge and behavior related to traffic laws

468

Table 5 shows the questionnaire results for knowledge and behavior related to traffic laws. Of all

469

respondents, 42% reported to have used an e-scooter with two people in the past. Asked if they had

470

used an e-scooter under the influence of alcohol in the past, 39% reported to have ridden under the

471

influence of alcohol. Regarding infrastructure usage, nearly two thirds of riders report to never have

472

driven an e-scooter against the direction of traffic, 23% admit to have done so rarely, 10% sometimes,

473

3% often, and 3% always. Asked how they signal a turn, 46% use their hands, 5% signal a turn by

474

extending their legs, and 49% report not to signal turns. One quarter of respondents could correctly

475

identify the legal age limit for e-scooter use in Germany. Three quarters correctly answered that no

476

driver’s license is needed for e-scooter use. Asked how many people are allowed on an e-scooter at

477

the same time, 84% of respondents correctly identified the limit of one person per e-scooter. On turn

478

signaling, only 19% correctly answered that Germany has a law on turn signaling on e-scooter by hand.

479

Asked whether there is a legal alcohol limit, 20% named a limit of 0.0 ‰ BAC, 46% named 0.5 ‰ BAC,

480

1% named 1.0 ‰ BAC, and 10% named 1.6 ‰ BAC. One fifth of respondents reported not to know the

481

limit, and 4% indicated to think that the alcohol limit is not regulated for e-scooters. As data on driver’s

482

license ownership was not collected in this study, only the answers of an alcohol limit over 0.5 ‰ BAC,

483

no limit, and lack of knowledge are counted as incorrect, leading to a total of 35% incorrect answers

484

on the legal alcohol limit for e-scooters.

485

In two questions (no. 27 and 33), respondents were presented with multiple infrastructure options

486

and asked to name those ones that they could legally use if all those options were available. For

487

question no. 27, the single correct answer was the use of the bicycle lane, which was correctly

488

identified as the sole correct answer by only 17% of respondents (although 90% included the bicycle

489

lane as one of multiple answers). For question no. 33, no bicycle lane was presented as an option,

490

hence e-scooters are required to use the street. More than half of the participants correctly identified

491

the street as the sole correct answer, while 86% included it as one of multiple answers.

492

493

Table 5. Survey questions and answers for knowledge and behavior related to traffic laws (% of

494

answers for n=156 respondents). All items are single choice unless indicated otherwise.

495

Question no.

Answering options

12. Have you ever used a single e-

scooter with two people?

57.7%

42.3%

yes

13. Have you ridden an e-scooter

under the influence of alcohol

before?

61.5%

38.5%

yes

14. Have you ridden an e-scooter

in the wrong direction before?

61.5%

23.1%

10.3%

2.6%

never

rarely

sometimes

often

always

22. How do you signal a turn?

46.2%

5.1%

48.7%

using my hands

extending my legs

not at all

28. How old do you have to be to

use an e-scooter on a public

German road?

1.3%

25.6%

25.0%

20.5%

0.0%

9.0%

18.6%

not

regulated

I don't know

29. Do you need a driver's license

to ride an e-scooter on public

roads in Germany?

5.1%

0.6%

3.8%

76.3%

14.1%

yes, a regular

driver's license

for cars

yes, a driver's

license for e-

scooters

yes, a driver's

license for

bicycles

I don't know

30. How many people are allowed

to simultaneously ride on a single

e-scooter on a public German

road?

84.0%

1.3%

1.9%

5.8%

7.1%

not regulated

I don't know

31. Does Germany have a law on

how to signal a turn when riding

an e-scooter?

19.2%

3.2%

21.8%

55.8%

yes, using your hands

yes, by extending your

legs

not regulated

I don't know

32. Is there a legal alcohol limit for

riding an e-scooter in Germany?

19.9%

45.5%

1.3%

10.3%

3.8%

19.2%

0.0 Blood

Alcohol

Content

0.5 BAC (same

as with cars in

Germany)

1.0 BAC

1.6 BAC (same

as with bikes

in Germany)

not regulated

I don't know

27. Where are you allowed to ride

e-scooters in public traffic in

Germany, if the following

infrastructure is available? (more

than one answer possible)

90.4%

19.2%

8.3%

10.3%

76.9%

1.3%

bicycle lane

bus lane

pedestrian

area

sidewalk

street

none of these

options

33. Where are you allowed to ride

e-scooters in public traffic in

Germany, if only the following

infrastructure is available? (more

than one answer possible)

22.4%

7.7%

12.2%

85.9%

10.3%

bus lane

pedestrian area

sidewalk

street

none of these

options

496

4.2.4. Gender and safety related behaviors

497

To assess whether the gender of riders is related to differences in reported safety related behavior, we

498

split survey data for riders that identified as female (n=46) or male (n=107). Resulting answers are

499

presented in Table 6, where the Chi-square test was used to compare questions with dichotomous

500

answers, and the Mann-Whitney U test was used to compare Likert-scale answers, due to non-normal

501

distributions in the subsamples of male and female riders. The comparison of female and male riders

502

in their self-reported safety related behavior did not reveal significant differences.

503

504

Table 6. Survey questions on safety related behavior for female and male riders.

505

Question no.

Female

Male

Test statistics

12. Have you ever used a single e-

scooter with two people?

Yes

58.7%

41.3%

56.1%

43.9%

(

2= 0.09, df = 1,

p=.76, φ=.02)

13. Have you ridden an e-scooter

under the influence of alcohol

before?

Yes

69.6%

30.4%

57.9%

42.1%

(

2= 1.83, df = 1,

p=.18, φ=.11)

(22.) Do you signal a turn?†

Yes

43.5%

56.5%

52.3%

47.7%

(

2= 1.01, df = 1,

p=.32, φ=.08)

14. Have you ridden an e-scooter in

the wrong direction before?

(1=never … 5=always)

Mean (SD)

1.39 (0.71)

1.73 (1.03)

U=2888.5, p=.51,

r=0.16

16. Do you wear a helmet when

riding an e-scooter?

(1=never … 5=always)

Mean (SD)

1.15 (0.73)

1.01 (0.10)

U=2376, p=.16, r=-

0.11

23. How safe do you feel on an e-

scooter?

(1=very unsafe … 7= very safe)

Mean (SD)

3.74 (1.44)

4.04 (1.57)

U=2724.5, p=.29,

r=0.09

† “Yes”-answers include hand and foot signaling from question no. 22

506

5. Discussion

507

5.1. Brake related hypotheses

508

In this study, the safety related knowledge and behavior of e-scooter riders in Berlin was investigated

509

in a combined observational and questionnaire survey. In our first hypothesis, we expected that riders

510

are unable to correctly identify the type of braking system of the shared e-scooter they had last used.

511

The results of our questionnaire survey indicate that this is correct, as only one third of respondents

512

was able to correctly identify the braking system of the shared e-scooter they had last used. While

513

these results could be a consequence of little experience with shared e-scooters (as more than 60% of

514

users had used a shared e-scooter only three times or less) and a long time interval since their last use,

515

they also indicate a lack of a simple mental model for e-scooter braking systems.

516

In our second hypothesis, we expected that right hand and foot brake levers would be readied less

517

frequently than the left hand brake lever by riders. Our data indicates that this is true, as the left hand

518

brake lever is readied significantly more often than the other available lever. For scooter models with

519

different braking systems (all hand lever vs. hand lever combined with foot brake), the foot brake was

520

readied significantly less often than the right hand lever. A possible reason for the preference of the

521

left hand brake lever over the right hand lever is the positioning of the acceleration lever on shared e-

522

scooters. For all e-scooter models, the lever for acceleration needs to be constantly actuated with the

523

thumb of the right hand, potentially impeding the readying of any available right hand lever brake. As

524

a similar complication in comparison to the left hand brake, the readying of the foot brake necessitates

525

a shift in riders’ body position, a prerequisite that is more effortful than the readying of the left hand

526

brake. Further, our observational results suggest that readying the foot brake is more effortful than

527

readying the right hand lever brake. In addition, our observational results suggest that shared e-

528

scooter riders do not base their brake readying decision on considerations on front-wheel vs. back-

529

wheel braking, as the location of the brake lever is the main influence on brake readying (Figure 10 &

530

Figure 11). Our observation of riders readying two brakes that actuate the same wheel (on Circ & Voi

531

scooters) reinforces this indication.

532

533

5.2. Additional challenges for the safety of riders

534

In addition to these braking-related hypotheses, our study revealed additional challenges for the safe

535

operation of e-scooters in Germany. The observational study revealed a small share of illegal dual use

536

of e-scooters (3%), which blocked drivers’ access to the foot brake in 1.5% of observations, limiting the

537

number of available brakes levers to one. This small share of observed dual use (registered as point

538

prevalence, i.e. at a single time point) conforms with a large share (42%) of self-reported dual use in

539

the past (life-time prevalence). Observed and self-reported helmet use was critically low, while self-

540

reported e-scooter riding under the influence of alcohol was high.

541

A considerable number of riders is unaware of existing legal regulations on e-scooters regarding the

542

age and alcohol limits for e-scooter use, turn signaling, and permissible infrastructure. On actual turn

543

signaling, close to 50% of respondents report not to signal turns, which could be related to findings of

544

riders feeling less safe when hand signaling on an e-scooter (Löcken, Brunner, Kates, & Riener, 2020).

545

The lack of overall knowledge about e-scooter regulation, in addition to the acknowledgement of past

546

illegal behavior may contribute to our finding that riding an e-scooter is rated as significantly less safe

547

than riding a bicycle. The share of riders who report having had a fall or a collision (10%) while using

548

an e-scooter is an indication that riders’ assessment of the risk related to e-scooter riding could be

549

accurate. The relatively high number of reported falls and collisions is even more alarming when

550

factoring in the short amount of time that shared e-scooters had been allowed in Germany at the time

551

of the questionnaire survey, and the very limited exposure to e-scooter riding that was present in the

552

survey sample. This finding is in line with a study on e-scooter related injuries in Austin, Texas (Austin

553

Public Health, 2019), which found that one third of 125 interviewed injured e-scooter users were first

554

time riders.

555

556

5.3. Implications for ergonomic design and regulation of shared e-scooters

557

For the design of e-scooter braking systems, our findings have direct implications to brake lever

558

placement and lever-to-wheel coupling. Our observational results indicate that shared e-scooter riders

559

do not chose to prepare a brake lever based on considerations of which wheel to brake, but solely on

560

the placement of the brake levers on the e-scooter. The preference for readying the left hand brake

561

lever indicates a higher usability of this brake lever in comparison to the right hand lever and the foot

562

brake. The most likely reason for this preference lies in the placement of the right thumb actuated

563

throttle lever which needs to be continuously actuated, and a comparatively high effort to ready the

564

foot brake. This knowledge can be used by e-scooter providers and manufacturers to design their

565

braking system more intuitively. In light of the higher efficiency of front wheel braking, it seems

566

advisable to couple the left hand brake lever with the front wheel of e-scooter models (as Circ and

567

Lime already do) and not to the back wheel (as Bird, Jump, Tier, and Voi do). However, further research

568

is needed to investigate the relation of front- and/or back wheel braking and e-scooter stability.

569

The indications of lack of knowledge of lever-to-wheel coupling of riders calls into question the practice

570

of coupling two separate brake levers to the same wheel (as Circ and Voi do). While this “same wheel

571

dual braking” complies with the letter of the law of e-scooter regulation in Germany (eKFV), it prevents

572

riders from decelerating both wheels of the e-scooter, reducing the overall potentially applicable brake

573

power. In addition, brake force application to both wheels, actuated through on lever (preferably on

574

the left side of the handlebar) could be used to increase potential brake force available to riders. For

575

the legislative regulation of braking systems, it seems worth investigating how brake levers actuated

576

by the right hand or through a footbrake stand in compliance with the general German road safety

577

regulation (StVO), which requires an “adequate brake that can be easily operated while driving”. While

578

experimental studies need to investigate the share of use of the right hand and foot brake, our results

579

indicate that these brake lever types will not be easily and quickly actuated in emergency braking

580

situations. To support the knowledge of shared e-scooter riders about the braking systems of a given

581

e-scooter, it seems advisable to add consistent color- and haptic coding of front and back wheel brake

582

levers. E.g. regulators could mandate that the back wheel actuating brake lever should be colored

583

darker and be tactilely coarser than the brighter and smoother front wheel actuating lever.

584

585

5.4. Limitations

586

There are a number of limitations to this study. In the observational study, brake readiness was

587

registered, but not actual braking. While we argue that brake readiness translates to actual braking

588

with the readied brake levers, an observation of individual e-scooters over a longer time span is needed

589

to show what share of brake readiness at a given brake lever translates to actual braking at that

590

individual lever. This validation of our observational approach is needed especially for the actuation of

591

the foot brake, where readying of the brake is not as apparent as for the hand lever brakes. For the

592

analysis of the video data, a number of variables could not be registered due to blurry video and riders

593

being partly out of frame (Figure 7). In addition, the number of e-scooters without complete data for

594

all variables increased during evening hours (Figure 9), potentially obscuring more dangerous

595

behaviors at evening hours, and prohibiting an analysis of the influence of time of day on riders

596

behavior. Future studies should use more light-sensitive (or infrared) cameras to minimize motion blur.

597

As the sample in the questionnaire survey was relatively small and young, future studies should aim

598

for larger sample sizes with a broader age-range, to produce results that are more representative,

599

especially in the light of the relation between age and traffic rule violations and crash rates (Alver,

600

Demirel, & Mutlu, 2014). As riders were surveyed mostly between noon and the early evening, future

601

studies should expand survey times to later hours, to collect a more comprehensive sample of e-

602

scooter users. Riders surveyed in our study had comparatively little experience in e-scooter use, as

603

shared e-scooters had just been introduced in Germany. While e-scooter use experience will further

604

increase in Germany and future studies will potentially not have this issue, they should nonetheless

605

aim to collect data from riders that use e-scooters more frequently, to check whether frequency of use

606

influences knowledge about braking systems and applicable regulation. In addition, future studies

607

should explore if different e-scooter providers are used by different types of riders. While the cost

608

structure and general marketing of providers in Germany did not initially target different user groups

609

this might change once providers try to differentiate themselves from their competitors. To assess

610

other potential influences on left- versus right-hand brake lever usage, handedness of riders should be

611

assessed in future studies.

612

613

5.5. Conclusion

614

In conclusion, this study revealed a number of factors in the ergonomic design of shared e-scooter

615

braking systems which can influence the safe use of e-scooters in the road environment. Legislative

616

bodies and e-scooter providers need to consider these findings to increase the safeness of e-scooter

617

use. In addition to these ergonomics challenges, our questionnaire survey revealed a critical lack of

618

knowledge in e-scooter users. Public education campaigns coupled with better information provision

619

through e-scooter providers on applicable laws and regulation are necessary to increase users’

620

knowledge on the safe use of e-scooters on the road.

621

622

623

References

624

Aizpuru, M., Farley, K. X., Rojas, J. C., Crawford, R. S., Moore, T. J., & Wagner, E. R. (2019). Motorized scooter

625

injuries in the era of scooter-shares: A review of the national electronic surveillance system. The American

626

Journal of Emergency Medicine, 37(6), 1133–1138. https://doi.org/10.1016/j.ajem.2019.03.049

627

Alver, Y., Demirel, M. C., & Mutlu, M. M. (2014). Interaction between socio-demographic characteristics: Traffic

628

rule violations and traffic crash history for young drivers. Accident; Analysis and Prevention, 72, 95–104.

629

https://doi.org/10.1016/j.aap.2014.06.015

630

Arellano, J. F., & Fang, K. (2019). Sunday Drivers, or Too Fast and Too Furious? Transport Findings. Advance online

631

publication. https://doi.org/10.32866/001c.11210

632

Austin Public Health (2019). Dockless Electric Scooter-Related Injuries Study: Austin, Texas. September -

633

Nobemer 2018. Retrieved from

634

https://www.austintexas.gov/sites/default/files/files/Health/Epidemiology/APH_Dockless_Electric_Scooter

635

_Study_5-2-19.pdf

636

Badeau, A., Carman, C., Newman, M., Steenblik, J., Carlson, M., & Madsen, T. (2019). Emergency department

637

visits for electric scooter-related injuries after introduction of an urban rental program. The American Journal

638

of Emergency Medicine, 37(8), 1531–1533. https://doi.org/10.1016/j.ajem.2019.05.003

639

Bai, S., & Jiao, J. (2020). Dockless E-scooter usage patterns and urban built Environments: A comparison study of

640

Austin, TX, and Minneapolis, MN. Travel Behaviour and Society, 20, 264–272.

641

https://doi.org/10.1016/j.tbs.2020.04.005

642

Beck, R. F. (2004). Mountain bicycle acceleration and braking factors. In Proceedings of the Canadian

643

Multidisciplinary Road Safety Conference XIV, Ottawa, Ontario. Retrieved from

644

https://pdfs.semanticscholar.org/863c/e69fbd86fc36a039d8f3c3e771926a04cf7a.pdf

645

Bekhit, M. N. Z., Le Fevre, J., & Bergin, C. J. (2020). Regional healthcare costs and burden of injury associated with

646

electric scooters. Injury, 51(2), 271–277. https://doi.org/10.1016/j.injury.2019.10.026

647

Bhise, V. D. (2012). Ergonomics in the automotive design process. Boca Raton: CRC Press. Retrieved from

648

http://site.ebrary.com/lib/academiccompletetitles/home.action

649

Bierbach, M., Adolph, T., Frey, A., Kollmus, B., Bartels, O., Hoffmann, H., & Halbach, A.-L. (2018). Untersuchung

650

zu Elektrokleinstfahrzeugen (Berichte der Bundesanstalt für Straßenwesen, Fahrzeugtechnik No. F 125).

651

Bergisch Gladbach. Retrieved from Bundesanstalt für Straßenwesen website:

652

https://www.bast.de/BASt_2017/DE/Publikationen/Berichte/unterreihe-f/2019-2018/f125.html

653

Blomberg, S. N. F., Rosenkrantz, O. C. M., Lippert, F., & Collatz Christensen, H. (2019). Injury from electric scooters

654

in Copenhagen: A retrospective cohort study. BMJ Open, 9(12), e033988. https://doi.org/10.1136/bmjopen-

655

2019-033988

656

DIN Deutsches Institut für Normung e.V. (2009). Safety of machinery - Ergonomics requirements for the design of

657

displays and control actuators - Part 1: General principles for human interactions with displays and control

658

actuators. (DIN EN 894-1:1997+A1:2008): Beuth.

659

European Union (2016, May 4). Regulation on the protection of natural persons with regard to the processing of

660

personal data and on the free movement of such data, and repealin. General Data Protection Regulation. OJ

661

L 119. Retrieved from https://eur-lex.europa.eu/legal-

662

content/EN/TXT/PDF/?uri=CELEX:32016R0679&from=EN

663

Friard, O., & Gamba, M. (2016). BORIS : a free, versatile open‐source event‐logging software for video/audio

664

coding and live observations. Methods in Ecology and Evolution, 7(11), 1325–1330.

665

https://doi.org/10.1111/2041-210X.12584

666

Garman, C., Como, S. G., Campbell, I. C., Wishart, J., O'Brien, K., & McLean, S. (2020). Micro-Mobility Vehicle

667

Dynamics and Rider Kinematics during Electric Scooter Riding. In SAE Technical Paper Series (2020-01-0935).

668

Commonwealth Drive, Warrendale, PA, United States: SAE International. https://doi.org/10.4271/2020-01-

669

0935

670

Gössling, S. (2020). Integrating e-scooters in urban transportation: Problems, policies, and the prospect of system

671

change. Transportation Research Part D: Transport and Environment, 79, 102230.

672

https://doi.org/10.1016/j.trd.2020.102230

673

Hawkins, F. H. (2006). Human factors in flight (2. ed.): Routledge.

674

Haworth, N. L., & Schramm, A. (2019). Illegal and risky riding of electric scooters in Brisbane. The Medical Journal

675

of Australia, 211(9), 412–413. https://doi.org/10.5694/mja2.50275

676

Huertas-Leyva, P., Dozza, M., & Baldanzini, N. (2019). E-bikers' braking behavior: Results from a naturalistic

677

cycling study. Traffic Injury Prevention, 20(sup3), 62–67. https://doi.org/10.1080/15389588.2019.1643015

678

Kraftfahrtbundesamt (2019). Allgemeine Betriebserlaubnis (ABE) für Fahrzeuge gemäß der Verordnung über die

679

Teilnahme von Elektrokleinstfahrzeugen am Straßenverkehr (Elektrokleinstfahrzeuge-Verordnung - eKFV).

680

Retrieved from

681

https://www.kba.de/DE/Typgenehmigung/Typgenehmigungen/Typgenehmigungserteilung/ABE_Elektroklei

682

nstfahrzeuge/ABE_Elektrokleinstfahrzeuge_node.html

683

Löcken, A., Brunner, P., Kates, R., & Riener, A. (2020). Impact of Hand Signals on Safety: Two Controlled Studies

684

With Novice E-Scooter Riders. In 12th International Conference on Automotive User Interfaces and Interactive

685

Vehicular Applications (pp. 132-140).

686

Mayer, E., Breuss, J., Robatsch, K., Salamon, B., & Soteropoulos, A. (2020). E-Scooter: Was bedeutet das neue

687

Fortbewegungsmittel für die Verkehrssicherheit. Zeitschrift Für Verkehrssicherheit, 66(3), 153–164.

688

Mordfin, L. (1975). The CPSC Road Test of Bicycle Braking Performance - Kinetic and Error Analyses (No. NBSIR

689

75-786). Retrieved from https://nvlpubs.nist.gov/nistpubs/Legacy/IR/nbsir75-786.pdf

690

Namiri, N. K., Lui, H., Tangney, T., Allen, I. E., Cohen, A. J., & Breyer, B. N. (2020). Electric Scooter Injuries and

691

Hospital Admissions in the United States, 2014-2018. JAMA Surgery. Advance online publication.

692

https://doi.org/10.1001/jamasurg.2019.5423

693

Oppenheim, I., & Shinar, D. (2011). Human Factors and Ergonomics. In B. E. Porter (Ed.), Handbook of Traffic

694

Psychology (1st ed., pp. 193–211). Amsterdam: Elsevier/Academic Press. https://doi.org/10.1016/B978-0-12-

695

381984-0.10015-3

696

Puzio, T. J., Murphy, P. B., Gazzetta, J., Dineen, H. A., Savage, S. A., Streib, E. W., & Zarzaur, B. L. (2020). The

697

electric scooter: A surging new mode of transportation that comes with risk to riders. Traffic Injury Prevention,

698

21(2), 175–178. https://doi.org/10.1080/15389588.2019.1709176

699

Störmann, P., Klug, A., Nau, C., Verboket, R. D., Leiblein, M., Müller, D., . . . Lustenberger, T. (2020).

700

Characteristics and Injury Patterns in Electric-Scooter Related Accidents-A Prospective Two-Center Report

701

from Germany. Journal of Clinical Medicine, 9(5), 1569. https://doi.org/10.3390/jcm9051569

702

Tack, A., Klein, A., & Bock, B. (2020). E-Scooter in Deutschland: Ein datenbasierter Debattenbeitrag. Retrieved

703

from http://scooters.civity.de

704

Todd, J., Krauss, D., Zimmermann, J., & Dunning, A. (2019). Behavior of Electric Scooter Operators in Naturalistic

705

Environments. In SAE Technical Paper Series. Commonwealth Drive, Warrendale, PA, United States: SAE

706

International. https://doi.org/10.4271/2019-01-1007

707

Trivedi, B., Kesterke, M. J., Bhattacharjee, R., Weber, W., Mynar, K., & Reddy, L. V. (2019). Craniofacial Injuries

708

Seen With the Introduction of Bicycle-Share Electric Scooters in an Urban Setting. Journal of Oral and

709

Maxillofacial Surgery, 77(11), 2292–2297. https://doi.org/10.1016/j.joms.2019.07.014

710

Trivedi, T. K., Liu, C., Antonio, A. L. M., Wheaton, N., Kreger, V., Yap, A., . . . Elmore, J. G. (2019). Injuries Associated

711

With Standing Electric Scooter Use. JAMA Network Open, 2(1), e187381.

712

https://doi.org/10.1001/jamanetworkopen.2018.7381

713

Uluk, D., Lindner, T., Palmowski, Y., Garritzmann, C., Göncz, E., Dahne, M., . . . Gerlach, U. A. (2020). E-Scooter:

714

erste Erkenntnisse über Unfallursachen und Verletzungsmuster. Notfall + Rettungsmedizin, 23(4), 293–298.

715

https://doi.org/10.1007/s10049-019-00678-3

716

Wilson, D. G., Schmidt, T., & Papadopoulos, J. (2020). Bicycling science (4. ed.). Cambridge (MA): The MIT Press.

717

Wolff, C. (2017). Grundlegendes zum Bremsvorgang. In B. Breuer & K. H. Bill (Eds.), ATZ / MTZ-Fachbuch.

718

Bremsenhandbuch: Grundlagen, Komponenten, Systeme, Fahrdynamik (5th ed., pp. 15–25). Wiesbaden:

719

Springer Vieweg. https://doi.org/10.1007/978-3-658-15489-9_2

720

721