Gait & Posture 41 (2015) 46–51
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Texting and walking: Effect of environmental setting and task prioritization on dual-task interference in healthy young adults Prudence Plummer a,*, Sarah Apple b, Colleen Dowd b, Eliza Keith b a b
Division of Physical Therapy, Department of Allied Health Sciences, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States Department of Physical Therapy, Northeastern University, Boston, MA, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Received 14 March 2014 Received in revised form 9 July 2014 Accepted 11 August 2014
Recent studies have shown that young adults significantly reduce their gait speed and weave more when texting while walking. Previous research has not examined the simultaneous dual-task effects on texting performance, therefore, the attention prioritization strategy used by young adults while texting and walking is not currently known. Moreover, it is not known whether laboratory-based studies accurately reflect texting and walking performance in the real world. This study compared dual-task interference during texting and walking between laboratory and real-world settings, and examined the ability of young adults to flexibly prioritize their attention between the two tasks in each environment. Texting and walking were assessed in single-task and three dual-task conditions (no-priority, gait-priority, texting-priority) in the lab and a University Student Center, in 32 healthy young adults. Dual-task effects on gait speed, texting speed, and texting accuracy were significant, but did not significantly differ between the two environments. Young adults were able to flexibly prioritize their attention between texting and walking, according to specific instruction, and this ability was not influenced by environmental setting. In the absence of instructions, young adults prioritized the texting task in the low-distraction environment, but displayed more equal focus between tasks in the real world. The finding that young adults do not significantly modify their texting and walking behavior in highdistraction environments lends weight to growing concerns about cell phone use and pedestrian safety. ß 2014 Elsevier B.V. All rights reserved.
Keywords: Gait Cognitive-motor interference Attention Prioritization Texting
1. Introduction Although laboratory-based studies have demonstrated that young adults are relatively resilient to dual-task interference during walking [1,2], emerging research suggests that the safety of young adults may be compromised during distracted walking in the real world, especially when walking while texting or talking on a cell phone [3–6]. Compared to undistracted pedestrians, individuals talking on a cell phone notice significantly fewer objects in their surrounding environment [4,7]. Reduced situational awareness, or inattention blindness [7], may be contributing to the increasing number of accidents and injuries reported during cell phone use while walking [8,9]. Furthermore, young adults using a cell phone demonstrate more risky behavior when crossing
* Corresponding author at: Division of Physical Therapy, Department of Allied Health Sciences, The University of North Carolina at Chapel Hill, 3020 Bondurant Hall, Campus Box #7135, Chapel Hill, NC 27599, United States. Tel.: +1 919 843 8658; fax: +1 919 966 3678. E-mail address: [email protected] (P. Plummer). http://dx.doi.org/10.1016/j.gaitpost.2014.08.007 0966-6362/ß 2014 Elsevier B.V. All rights reserved.
a street (e.g., more hits by virtual vehicles) than those not distracted by a cell phone conversation [4,6,10,11] or texting [5]. Texting while walking may increase safety risks and produce greater decrements in gait than talking while walking due to the visual attention and added motor demands required for reading and typing, in addition to the cognitive processes required for the communication interchanges [5]. Indeed, research demonstrates larger decreases in gait velocity and greater lateral deviation in young adults walking and texting compared to those walking and talking [3] or reading text [12]. Nonetheless, young adults talking on a cell phone while walking on a university campus slow down, change direction, and weave more, than those not using cell phones or other electronic devices [7]. Existing research provides insight into gait characteristics while texting, but none of the studies has reported the simultaneous dual-task effects on texting performance. Therefore, it is presently not known how young adults prioritize their attention during texting and walking. To accurately interpret dual-task interference it is imperative to measure single and dual-task performance in both tasks [13]. Thus, the purpose of this study was to comprehensively examine dual-task interference on texting and
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
walking in healthy young adults. The specific aims were to (1) compare dual-task interference during texting and walking between the laboratory and a real-world setting, and (2) examine the ability of young adults to flexibly prioritize attention while texting and walking in each environment. To address our aims, we adopted the paradigm of Kelly et al. [14] with three critical differences: instead of ‘‘usual’’ and ‘‘challenging’’ lab-based gait tasks we compared walking in the lab with walking in the ‘‘real world;’’ we used texting as the non-gait secondary task, arguably the most relevant dual-task for young adults; finally, we used a ‘‘no-priority’’ dual-task condition instead of ‘‘equal focus’’ to investigate how young adults spontaneously prioritize attention in each environment. We expected dual-task interference to be greater in the real world than in the laboratory due to increased attentional demands required to safely navigate an open environment, and that the ability to flexibly shift attention between the two tasks would be reduced in the real world. 2. Methods 2.1. Participants Thirty-two healthy young adults were recruited from the University community. Participants had to be 18–30 years old, fluent in English, regular users of a touch-screen Smart Phone, report familiarity with text-messaging, and have normal or corrected-to-normal vision. Individuals were ineligible if they reported a history of medical illness or hospitalization in the last 6 months, diagnosis of neurological disease, vestibular dysfunction, pain or other condition limiting walking or the ability to text on a mobile phone. Two brief questionnaires evaluated typical cell phone usage and texting habits. The study was approved by the local Institutional Review Board University and all participants provided written informed consent. 2.2. Procedures Participants were assessed in single and dual-task conditions in the research laboratory and a real-world setting (University Student Center). In each environment, all participants performed each task: (1) texting while standing (single-task texting), (2) walking at preferred speed (single-task walking), (3) walking while texting, without specific instruction to prioritize either task (dualtask no-priority), (4) walking while texting with instruction to focus on walking (dual-task gait-priority), and (5) walking while texting with instruction to focus on texting (dual-task textingpriority). Order of the environment was counterbalanced. The dual-task no-priority condition was performed before the gaitpriority and texting-priority conditions to minimize the effect of instructions on no-priority performance [14]. Order of gait-priority and texting-priority conditions was quasi-randomized. Singletasks were performed before dual-tasks. Each task was repeated twice in each setting and the average of the two trials was used for analysis. A freely available iPhone application, ‘‘My Speed,’’ was used for the texting task. Participants were instructed to type the phrase that appeared on the screen as quickly and as accurately as possible into the textbox below the phrase. The texting keyboard was the typical iPhone (QWERTY) keyboard. The software did not permit any errors, nor did it perform autocorrect or autofill operations. Thus, participants were required to type every character and correct any errors. Participants were alerted to an error by a change in text color from black to red, and vibration of the phone. At the completion of the task, the software displayed the texting speed (characters per minute), error rate (%), and duration (s), which were recorded by the experimenters.
47
Participants first underwent a familiarization period with the texting program sitting in the lab (minimum of 5 texting trials, until stability in speed and accuracy were observed). All participants used the same iPhone for the experiment. The gait task involved continuous straight-line walking along a 30-m walkway in each environment. Spatiotemporal gait data were acquired using a 5-node a body-worn sensor system (BioSensics, Cambridge, MA), comprising 5 inertial measurement units attached via Velcro straps to the anterior surface of each shin and thigh, and posteriorly on the low back. The system uses a twolink inversed pendulum model based on the participant’s height to determine spatiotemporal gait parameters [15]. Reliability and validity have been established in several publications [16–18]. The laboratory environment was a quiet corridor immediately outside the research lab. It had firm, tiled flooring and was neighbored by faculty office suites with infrequent foot-traffic. The real-world environment was an indoor walkway in the University Student Center, with firm, low-pile carpeted flooring. In contrast to the lab setting, the area was a busy pedestrian thoroughfare, with ATMs and a cafeteria on one side and a bookstore and restrooms on the other side; participants traversed through sliding doors at each end of the defined walkway. For the dual-task conditions, participants were instructed to text and walk, and to stop walking as soon as they completed the phrase (ensuring that gait data represented dual-tasking). In the no-priority dual-task condition, participants were not given any instruction regarding which task to prioritize. In the gait-priority condition, participants were instructed to focus mainly on their walking so that they were walking as they did when they were not texting. Conversely, in the texting-priority condition, participants were instructed to focus mainly on texting so that they were texting as fast and as accurately as they did when they were not walking. Performances were videotaped and later coded for density of pedestrian traffic (number of people that walked passed or directly across the path of the participant), collisions or near collisions (number of contacts/near contacts with another person), path adjustment (number of times the participant deviated to avoid a collision, not including spontaneous weaving), and situational awareness (number of times the person looked up from the phone while walking). 2.3. Statistical analysis and sample size The effect of environmental setting on dual-task interference (Aim 1) was first analyzed by applying a repeated measures ANOVA with Environment (lab, real world) and Task (single-task, dual-task no-priority) to gait speed (m/s), texting speed (characters per minute) and texting accuracy (%). We also compared the relative dual-task effects (DTE, percent change in performance in the dualtask condition relative to the single-task condition) between the lab and the real world on gait speed (DTEg) and texting performance (DTEt) using paired t-tests. We summed texting speed and accuracy DTE to compute an overall texting DTE, thereby accounting for speed-accuracy tradeoffs within the texting task [19]. Negative DTE values indicate performance deterioration, or a dual-task cost, while positive values indicate an improvement, or dual-task benefit. Instructed prioritization effects (Aim 2) were examined using an Environment (lab, real world) Instructions (no-priority, gait-priority, texting-priority) repeated measures ANOVA for gait speed, texting speed, texting accuracy, DTEg, and DTEt. Analyses were performed using SPSS 16.0 (SPSS Inc., Chicago, IL). The study was powered to detect a large standardized effect size for the interaction effect on gait speed between Environment and Task (Aim 1), since small or moderate effect sizes would generally be of limited practical importance for healthy young adults. Based on Type I error rate of 5% and Type II error rate of 20%
48
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
Table 1 Participant characteristics and media use (n = 31). Mean (SD) or % Gender female Age (years) Education (years) Ethnicity White/non-Hispanic African American Hispanic Asian Montreal Cognitive Assessment (max. 30) Cell phone type iPhone Other touch-screen phone Length of time with current phone (months) Texts sent per day – weekday, weekend 0–9 10–30 31–50 51–70 71–90 91–110 >110 % of cell phone use (talking/texting) while walking None 75% Experienced collision or near collision while texting and walking
61.3% 22.5 16.2
(2.1) (1.6)
83.9% 3.2% 3.2% 9.7% 28.4
(2.0)
90.3% 9.7% 14.6
(12.4)
0% 45.2% 12.9% 25.8% 3.2% 6.5% 6.5%
0% 22.6% 32.3% 29.0% 6.5% 3.2% 6.5%
0% 32.3% 29.0% 29.0% 9.7% 32.3%
0% 22.6% 51.6% 25.8% 0%
p < .001, h2p ¼ :82]. Overall, gait speed in the real world (M = 1.22 m/s, SD = .15) was faster than in the lab (M = 1.18 m/s, SD = .14) [Environment main effect: F(1,30) = 6.31, p = .02, h2p ¼ :17]; however, the absolute effect size is small and is unlikely to be clinically meaningful. There was no Task Environment interaction for gait speed [F(1,30) = .96, p = .33, h2p ¼ :03] (Fig. 1A). DTEg (no-priority) was not significantly different between the lab and the real world [t(30) = 1.29, p = .21, d = .23]. There were significant effects of Task on texting speed [F(1,30) = 30.59, p < .001, h2p ¼ :51] and accuracy [F(1,30) = 22.67, p < .001, h2p ¼ :43] but no main effects of Environment or interaction effects for either (Fig. 1). Although no-priority DTEt was larger in the real world (M = 17.8%, SD = 22.1) compared to the lab (M = 7.9%, SD = 19.0), the difference was not statistically significant [t(30) = 1.85, p = .08, d = .36]. 3.2. Flexible prioritization of attention in laboratory versus real world
(i.e., power = .80), 26 participants were required. To maximize power to detect effects in Aim 2, we increased the sample size by 20%. 3. Results One participant consistently walked too fast to complete the texting task within the available walkway, and therefore finished texting while standing still. Since his data did not represent true dual-task performance, he was excluded from the analysis. Descriptive characteristics of the 31 participants included in the analysis are presented in Table 1. Pedestrian traffic in the real world was greater than in the lab setting (p < .001); participants encountered an average of 10.9 (SD 6.1) pedestrians in each trial in the Student Center compared to only .5 (SD 1.0) in the lab corridor. None of the participants were required to deviate from their straight-line walking path in the lab, but path adjustment to avoid collisions was required in 44% of gait trials in the real world, with 35% of those involving more than one adjustment (max. 4). Indeed, based on our video analysis, participants looked up from the phone more often in the real world (1.7 2.0 times per trial) than in the lab (.2 .6 times per trial; p < .001). 3.1. Dual-task effects in laboratory versus real world Gait speed while texting (M = 1.09 m/s, SD = .17) was significantly slower than single-task walking (M = 1.32 m/s, SD = .12) [Task main effect: F(1,30) = 133.19,
Instructed prioritization influenced gait speed [F(2,60) = 67.87, p < .001,
h2p ¼ :69]. Gait-priority gait speed was faster than no-priority and texting-priority gait speeds (p < .001; Table 2). Conversely, texting-priority gait speed was slower than no-priority (p < .001; Fig. 2). Instructed prioritization also influenced DTEg [F(2,60) = 68.72, p < .001, h2p ¼ :70], with smaller gait-priority DTEg compared to no-priority and texting-priority DTEg (p < .001); texting-priority DTEg was larger than no-priority DTEg (p < .001). Environment did not influence DTEg [F(1,30) = 1.35, p = .25, h2p ¼ :04], nor was there an interaction [F(2,60) = 1.40, p = .25, h2p ¼ :05]. Instructions did not influence texting speed [F(2,60) = 2.40, p = .09, h2p ¼ :07] (Fig. 2), but texting speed averaged across the three dual-task conditions was slower in the real world than the lab [Environment main effect: F(1,30) = 7.31, p = .01, h2p ¼ :20], with no interaction [F(2,60) = .05, p = .96, h2p < :01]. Texting accuracy was influenced by Instructions [F(2,60) = 3.60, p = .03, h2p ¼ :11], with lower accuracy when participants were instructed to prioritize gait (M = 89.2%, SD = 7.1) compared to texting (M = 92.0%, SD = 5.1; p = .04). Neither gait-priority texting accuracy nor texting-priority accuracy were significantly different from nopriority accuracy (M = 91.1%, SD = 4.9). Instructed prioritization also influenced DTEt [F(2,60) = 3.58, p = .03, h2p ¼ :11]; but Environment did not [F(1,30) = 3.81, p = .06, h2p ¼ :11], with no interaction [F(2,60) = .10, p = .91, h2p < :01] (Fig. 3). To further evaluate the patterns of prioritization between texting and walking in each environment, we plotted the gait and texting DTE against each other. Fig. 3A illustrates the instructed-prioritization effect in both environments and highlights that Environment affected DTEt more than DTEg (i.e., further displacement of DTE between the circle and diamond-shaped symbols on the x-axis than the y-axis). Of note, the 45-degree angle between the two zero lines in Fig. 3A represents equal gait and texting DTE. Based on this, the default prioritization of participants (i.e., nopriority) in the lab was to prioritize texting: DTEt (7.9%) < DTEg (18.5%). In the real world, default prioritization reflected more equal focus: DTEt (17.8%) > DTEg (16.3%).
4. Discussion Our hypothesis that dual-task interference would be greater in the real world than in the laboratory setting was only partially supported by our data. There was no significant difference in the absolute and relative no-priority dual-task effects between the two environments. Thus, without specific instructions about which task to prioritize, young adults perform similarly in high and lowdistraction environments. However, there was a significant effect of Environment on texting speed when averaged across the three dual-task conditions, with slower mean dual-task texting speed in
Fig. 1. Gait speed (A), texting speed (B), and texting accuracy (C) shown as a function of Task (single-task, dual-task no-priority) and Environment (lab, real world). The difference between single and dual-task performance is representative of the dual-task effect. Error bars indicate standard error.
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
49
Table 2 Mean values (SD) absolute and relative dual-task measures of gait and texting performance in each environment, task condition, and instructed prioritization. Real world
Laboratory Single-task
Gait Gait speed (m/s) Gait speed DTE (%) Texting Speed (char/min) Speed DTE (%) Accuracy (%) Accuracy DTE (%) Texting DTE (%)
Dual-task
Single-task
No-priority
Gait-priority
Texting-priority
1.30 (.12)
1.06 (.18) 18.5 (10.9)
1.19 (.15) 9.2 (7.6)
.94 (.17) 27.9 (10.0)
238 (50)
222 (45) 5.5 (15.1) 91.6 (4.2) 2.4 (4.8) 7.9 (19.0)
215 (61) 9.2 (19.3) 89.2 (6.8) 4.9 (7.5) 14.1 (25.5)
229 (53) 2.3 (19.1) 92.7 (4.3) 1.2 (4.7) 3.5 (22.5)
93.9 (3.8)
the real world. The effect of Environment on overall texting dualtask effect was not statistically significant, but the effect size was large [20]: 11% of the variance in DTEt was accounted for by Environment. Since the study was powered for the dependent variable of gait speed, it is possible that there was insufficient power to detect environment effects in texting performance, especially in light of the greater variability in texting performance than gait (as illustrated Fig. 3A and Table 2). The current study is the first to examine how young adults spontaneously allocate their attention while walking and texting. Our data suggest that in the lab young adults prioritized texting, perhaps because there were minimal risks to safety in this setting. This is consistent with previous research in which young adults were observed to allocate most of their attention to the texting task instead of walking [21]. The gait and texting no-priority dual-task effects in the real world signify more equal distribution of attention in that environment. Relative to the lab, dual-task costs on texting increased, while DTEg remained relatively stable. The increase in texting-related dual-task costs in the real world is likely due to the increase in number of times participants looked up from the phone in the real world compared to the lab. Kelly et al. [14] reported that in more challenging gait tasks (narrow-based walking), young adults prioritized walking over the cognitive task, whereas during usual walking, young adults prioritized the cognitive task, despite instructions to prioritize both tasks equally. Our results partially support those of Kelly et al. [14] if one considers walking in the real world to be more challenging than walking in the lab. The reason that we may not have found evidence of walking prioritization in the more challenging situation is that, unlike narrow-based walking, it did not directly increase postural control demands.
Dual-task No-priority
Gait-priority
Texting-priority
1.33 (.13)
1.12 (.15) 16.3 (8.8)
1.21 (.15) 9.4 (7.9)
.98 (.17) 26.2 (10.0)
242 (50)
208 (50) 13.2 (17.2) 90.5 (5.5) 4.6 (6.1) 17.8 (22.1)
202 (53) 15.9 (17.7) 89.1 (7.3) 6.1 (8.2) 22.0 (24.5)
217 (55) 9.6 (17.2) 91.2 (5.9) 3.9 (6.9) 13.4 (22.8)
95.0 (3.1)
Consistent with previous research [14,19,22], we found that young adults can flexibly shift their attention during dual-task walking. We examined whether this ability was influenced by environmental setting. Young adults increased/decreased their gait speed and texting accuracy according to instructed focus, but texting speed did not significantly differ across the variablepriority conditions. In other words, in the texting-priority dualtask condition, participants further slowed their walking to improve their typing accuracy, while maintaining their typing speed. The effects on typing accuracy and not speed might underscore a habitual reliance on autocorrect functions. There was no difference in the effect of instructed prioritization between environments. This is in contrast to Kelly et al. [14] who found that young adults were better able to modulate dual-task gait speed according to instructed focus during usual walking than during more challenging walking. The current findings suggest that walking in the real world may not demand considerably greater attentional resources than walking in the lab, at least for the realworld environment we used, and for healthy young adults. Although the current study provides evidence that young adults can flexibly shift attentional focus, and that this ability is not hampered by the pervasive distractions of a real-world environment, it is unclear whether young adults make the appropriate choices regarding task prioritization in everyday life. It is plausible that, in reality, young adults may choose to stop texting and focus on gait in a very crowded environment, or when safety is more likely to be compromised (e.g., crossing the street). Participants may have felt compelled to continue walking and texting due to the methodological approach and awareness of the experimental situation. Another limitation is that our texting task was not entirely realistic. The abstract phrases (as opposed to self-generated text messages) and
Fig. 2. Gait speed (A), texting speed (B), and texting accuracy (C) shown as a function of instructed prioritization (dual-task no-priority, dual-task gait-priority, dual-task texting-priority) and environment (lab, real world). Error bars indicate standard error.
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
50
Fig. 3. (A) Plot of gait speed and texting dual-task effects (DTE) illustrating prioritization patterns in each dual-task condition in each environment. Circles represent DTE in the lab and diamonds represent DTE in the real world. (B) Gait DTE as a function of environment and priority condition. (C) Texting DTE as a function of environment and priority condition. In all charts, error bars indicate standard error.
absence of autocorrect and autofill operations may have made the texting task relatively more difficult. However, we consider the controlled nature of the texting task to be a minor limitation that is outweighed by the benefits of acquiring objective measures of texting performance.
Acknowledgements The authors gratefully acknowledge K. Simberg and E. Bitterman for assistance with data collection, and B. Najafi and G. Grewal for assistance with data processing.
5. Conclusion This study examined dual-task interference using the most ecologically valid dual-task for healthy young adults, and compared lab-based performance with performance in the actual, everyday environment. Our findings corroborate recent reports that texting produces large dual-task decrements in gait performance [3,12]. Moreover, the current study provides new evidence that dual-task interference associated with texting while walking, and the ability to flexibly prioritize attention between the tasks, are not significantly altered by ubiquitous distractions in the real world. Cell phone use while walking has become a pedestrian hazard [8,9]. The finding that young adults do not spontaneously modify their texting and walking behavior between low-distraction and high-distraction environments substantiates pedestrian safety concerns. Conflicts of interest statement There are no conflicts of interest for any of the authors.
References [1] Plummer-D’Amato P, Brancato B, Dantowitz M, Birken S, Bonke C, Furey E. Effects of gait and cognitive task difficulty on cognitive-motor interference in aging. J Aging Res 2012. 8 pp.. [2] Plummer-D’Amato P, Altmann LJP, Reilly K. Dual-task effects of spontaneous speech and executive function on gait in aging: exaggerated effects in slow walkers. Gait Posture 2011;33:233–7. [3] Lamberg E, Muratori L. Cell phones change the way we walk. Gait Posture 2012;35:688–90. [4] Nasar J, Hecht P, Wener R. Mobile telephones, distracted attention, and pedestrian safety. Accid Anal Prev 2008;40:69–75. [5] Schwebel D, Stavrinos D, Byington K, Davis T, O’Neal E, de Jong D. Distraction and pedestrian safety: how talking on the phone, texting, and listening to music impact crossing the street. Accid Anal Prev 2012;45:266–71. [6] Stavrinos D, Byington K, Schwebel D. Distracted walking: cell phones increase injury risk for college pedestrians. J Safety Res 2011;42:101–7. [7] Hyman Jr I, Boss M, Wise B, McKenzie K, Caggiano J. Did you see the unicycling clown? Inattentional blindness while walking and talking on a cell phone. Appl Cogn Psychol 2010;24:597–607. [8] Richtel M. Driven to distraction. Forget gum. Walking and using phone is risky. New York Times: A12010. New York, NY: The New York Times Company.
P. Plummer et al. / Gait & Posture 41 (2015) 46–51 [9] Nasar JL, Troyer D. Pedestrian injuries due to mobile phone use in public places. Accid Anal Prevent 2013;57:91–5. [10] Hatfield J, Murphy S. The effects of mobile phone use on pedestrian crossing behaviour at signalised and unsignalised intersections. Accid Anal Prevent 2007;39:197–205. [11] Neider MB, McCarley JS, Crowell JA, Kaczmarski H, Kramer AF. Pedestrians, vehicles, and cell phones. Accid Anal Prevent 2010;42:589–94. [12] Schabrun S, van den Hoorn W, Moorcroft A, Greenland C, Hodges P. Texting and walking: strategies for postural control and implications for safety. PLOS ONE 2014;9:e84312. [13] Plummer P, Eskes G, Wallace S, Giuffrida C, Fraas M, Campbell G, et al. Cognitive-motor interference during functional mobility after stroke: state of the science and implications for future research. Arch Phys Med Rehabil 2013;94:2565–74. [14] Kelly V, Eusterbrock AJ, Shumway-Cook A. Factors influencing dynamic prioritization during dual-task walking in healthy young adults. Gait Posture 2013;37:131–4. [15] Najafi B, Aminian K, Loew F, Blanc Y, Robert P. Measurement of stand-sit and sit-stand transitions using a miniature gyroscope and its application in fall risk evaluation in the elderly. IEEE Trans Biomed Eng 2002;49:843–51.
51
[16] Aminian K, Najafi B, Bula C, Leyvraz P, Robert P. Spatiotemporal parameters of gait measured by an ambulatory system using miniature gyroscopes. J Biomech 2002;35:689–99. [17] Aminian K, Trevisan C, Najafi B, Dejnabadi H, Frigo C, Pavan E, et al. Evaluation of an ambulatory system for gait analysis in hip osteoarthritis and after total hip replacement. Gait Posture 2004;20:102–7. [18] Najafi B, Helbostad J, Moe-Nilssen R, Zijlstra W, Aminian K. Does walking strategy in older people change as a function of walking distance? Gait Posture 2009;29:261–6. [19] Kelly V, Janke A, Shumway-Cook A. Effects of instructed focus and task difficulty on concurrent walking and cognitive task performance in healthy young adults. Exp Brain Res 2010;207:65–73. [20] Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Erlbaum, NJ: Hillsdale; 1988. [21] Lopresti-Goodman SM, Rivera A, Dressel C. Practicing safe text: the impact of texting on walking behavior. Appl Cogn Psychol 2012;26:644–8. [22] Yogev-Seligmann G, Rotem-Galili Y, Mirelman A, Dickstein R, Giladi N, Hausdorff JM. How does explicit prioritization alter walking during dual-task performance? Effects of age and sex on gait speed and variability. Phys Ther 2010;90:177–86.
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Texting and walking: Effect of environmental setting and task prioritization on dual-task interference in healthy young adults Prudence Plummer a,*, Sarah Apple b, Colleen Dowd b, Eliza Keith b a b
Division of Physical Therapy, Department of Allied Health Sciences, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States Department of Physical Therapy, Northeastern University, Boston, MA, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Received 14 March 2014 Received in revised form 9 July 2014 Accepted 11 August 2014
Recent studies have shown that young adults significantly reduce their gait speed and weave more when texting while walking. Previous research has not examined the simultaneous dual-task effects on texting performance, therefore, the attention prioritization strategy used by young adults while texting and walking is not currently known. Moreover, it is not known whether laboratory-based studies accurately reflect texting and walking performance in the real world. This study compared dual-task interference during texting and walking between laboratory and real-world settings, and examined the ability of young adults to flexibly prioritize their attention between the two tasks in each environment. Texting and walking were assessed in single-task and three dual-task conditions (no-priority, gait-priority, texting-priority) in the lab and a University Student Center, in 32 healthy young adults. Dual-task effects on gait speed, texting speed, and texting accuracy were significant, but did not significantly differ between the two environments. Young adults were able to flexibly prioritize their attention between texting and walking, according to specific instruction, and this ability was not influenced by environmental setting. In the absence of instructions, young adults prioritized the texting task in the low-distraction environment, but displayed more equal focus between tasks in the real world. The finding that young adults do not significantly modify their texting and walking behavior in highdistraction environments lends weight to growing concerns about cell phone use and pedestrian safety. ß 2014 Elsevier B.V. All rights reserved.
Keywords: Gait Cognitive-motor interference Attention Prioritization Texting
1. Introduction Although laboratory-based studies have demonstrated that young adults are relatively resilient to dual-task interference during walking [1,2], emerging research suggests that the safety of young adults may be compromised during distracted walking in the real world, especially when walking while texting or talking on a cell phone [3–6]. Compared to undistracted pedestrians, individuals talking on a cell phone notice significantly fewer objects in their surrounding environment [4,7]. Reduced situational awareness, or inattention blindness [7], may be contributing to the increasing number of accidents and injuries reported during cell phone use while walking [8,9]. Furthermore, young adults using a cell phone demonstrate more risky behavior when crossing
* Corresponding author at: Division of Physical Therapy, Department of Allied Health Sciences, The University of North Carolina at Chapel Hill, 3020 Bondurant Hall, Campus Box #7135, Chapel Hill, NC 27599, United States. Tel.: +1 919 843 8658; fax: +1 919 966 3678. E-mail address: [email protected] (P. Plummer). http://dx.doi.org/10.1016/j.gaitpost.2014.08.007 0966-6362/ß 2014 Elsevier B.V. All rights reserved.
a street (e.g., more hits by virtual vehicles) than those not distracted by a cell phone conversation [4,6,10,11] or texting [5]. Texting while walking may increase safety risks and produce greater decrements in gait than talking while walking due to the visual attention and added motor demands required for reading and typing, in addition to the cognitive processes required for the communication interchanges [5]. Indeed, research demonstrates larger decreases in gait velocity and greater lateral deviation in young adults walking and texting compared to those walking and talking [3] or reading text [12]. Nonetheless, young adults talking on a cell phone while walking on a university campus slow down, change direction, and weave more, than those not using cell phones or other electronic devices [7]. Existing research provides insight into gait characteristics while texting, but none of the studies has reported the simultaneous dual-task effects on texting performance. Therefore, it is presently not known how young adults prioritize their attention during texting and walking. To accurately interpret dual-task interference it is imperative to measure single and dual-task performance in both tasks [13]. Thus, the purpose of this study was to comprehensively examine dual-task interference on texting and
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
walking in healthy young adults. The specific aims were to (1) compare dual-task interference during texting and walking between the laboratory and a real-world setting, and (2) examine the ability of young adults to flexibly prioritize attention while texting and walking in each environment. To address our aims, we adopted the paradigm of Kelly et al. [14] with three critical differences: instead of ‘‘usual’’ and ‘‘challenging’’ lab-based gait tasks we compared walking in the lab with walking in the ‘‘real world;’’ we used texting as the non-gait secondary task, arguably the most relevant dual-task for young adults; finally, we used a ‘‘no-priority’’ dual-task condition instead of ‘‘equal focus’’ to investigate how young adults spontaneously prioritize attention in each environment. We expected dual-task interference to be greater in the real world than in the laboratory due to increased attentional demands required to safely navigate an open environment, and that the ability to flexibly shift attention between the two tasks would be reduced in the real world. 2. Methods 2.1. Participants Thirty-two healthy young adults were recruited from the University community. Participants had to be 18–30 years old, fluent in English, regular users of a touch-screen Smart Phone, report familiarity with text-messaging, and have normal or corrected-to-normal vision. Individuals were ineligible if they reported a history of medical illness or hospitalization in the last 6 months, diagnosis of neurological disease, vestibular dysfunction, pain or other condition limiting walking or the ability to text on a mobile phone. Two brief questionnaires evaluated typical cell phone usage and texting habits. The study was approved by the local Institutional Review Board University and all participants provided written informed consent. 2.2. Procedures Participants were assessed in single and dual-task conditions in the research laboratory and a real-world setting (University Student Center). In each environment, all participants performed each task: (1) texting while standing (single-task texting), (2) walking at preferred speed (single-task walking), (3) walking while texting, without specific instruction to prioritize either task (dualtask no-priority), (4) walking while texting with instruction to focus on walking (dual-task gait-priority), and (5) walking while texting with instruction to focus on texting (dual-task textingpriority). Order of the environment was counterbalanced. The dual-task no-priority condition was performed before the gaitpriority and texting-priority conditions to minimize the effect of instructions on no-priority performance [14]. Order of gait-priority and texting-priority conditions was quasi-randomized. Singletasks were performed before dual-tasks. Each task was repeated twice in each setting and the average of the two trials was used for analysis. A freely available iPhone application, ‘‘My Speed,’’ was used for the texting task. Participants were instructed to type the phrase that appeared on the screen as quickly and as accurately as possible into the textbox below the phrase. The texting keyboard was the typical iPhone (QWERTY) keyboard. The software did not permit any errors, nor did it perform autocorrect or autofill operations. Thus, participants were required to type every character and correct any errors. Participants were alerted to an error by a change in text color from black to red, and vibration of the phone. At the completion of the task, the software displayed the texting speed (characters per minute), error rate (%), and duration (s), which were recorded by the experimenters.
47
Participants first underwent a familiarization period with the texting program sitting in the lab (minimum of 5 texting trials, until stability in speed and accuracy were observed). All participants used the same iPhone for the experiment. The gait task involved continuous straight-line walking along a 30-m walkway in each environment. Spatiotemporal gait data were acquired using a 5-node a body-worn sensor system (BioSensics, Cambridge, MA), comprising 5 inertial measurement units attached via Velcro straps to the anterior surface of each shin and thigh, and posteriorly on the low back. The system uses a twolink inversed pendulum model based on the participant’s height to determine spatiotemporal gait parameters [15]. Reliability and validity have been established in several publications [16–18]. The laboratory environment was a quiet corridor immediately outside the research lab. It had firm, tiled flooring and was neighbored by faculty office suites with infrequent foot-traffic. The real-world environment was an indoor walkway in the University Student Center, with firm, low-pile carpeted flooring. In contrast to the lab setting, the area was a busy pedestrian thoroughfare, with ATMs and a cafeteria on one side and a bookstore and restrooms on the other side; participants traversed through sliding doors at each end of the defined walkway. For the dual-task conditions, participants were instructed to text and walk, and to stop walking as soon as they completed the phrase (ensuring that gait data represented dual-tasking). In the no-priority dual-task condition, participants were not given any instruction regarding which task to prioritize. In the gait-priority condition, participants were instructed to focus mainly on their walking so that they were walking as they did when they were not texting. Conversely, in the texting-priority condition, participants were instructed to focus mainly on texting so that they were texting as fast and as accurately as they did when they were not walking. Performances were videotaped and later coded for density of pedestrian traffic (number of people that walked passed or directly across the path of the participant), collisions or near collisions (number of contacts/near contacts with another person), path adjustment (number of times the participant deviated to avoid a collision, not including spontaneous weaving), and situational awareness (number of times the person looked up from the phone while walking). 2.3. Statistical analysis and sample size The effect of environmental setting on dual-task interference (Aim 1) was first analyzed by applying a repeated measures ANOVA with Environment (lab, real world) and Task (single-task, dual-task no-priority) to gait speed (m/s), texting speed (characters per minute) and texting accuracy (%). We also compared the relative dual-task effects (DTE, percent change in performance in the dualtask condition relative to the single-task condition) between the lab and the real world on gait speed (DTEg) and texting performance (DTEt) using paired t-tests. We summed texting speed and accuracy DTE to compute an overall texting DTE, thereby accounting for speed-accuracy tradeoffs within the texting task [19]. Negative DTE values indicate performance deterioration, or a dual-task cost, while positive values indicate an improvement, or dual-task benefit. Instructed prioritization effects (Aim 2) were examined using an Environment (lab, real world) Instructions (no-priority, gait-priority, texting-priority) repeated measures ANOVA for gait speed, texting speed, texting accuracy, DTEg, and DTEt. Analyses were performed using SPSS 16.0 (SPSS Inc., Chicago, IL). The study was powered to detect a large standardized effect size for the interaction effect on gait speed between Environment and Task (Aim 1), since small or moderate effect sizes would generally be of limited practical importance for healthy young adults. Based on Type I error rate of 5% and Type II error rate of 20%
48
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
Table 1 Participant characteristics and media use (n = 31). Mean (SD) or % Gender female Age (years) Education (years) Ethnicity White/non-Hispanic African American Hispanic Asian Montreal Cognitive Assessment (max. 30) Cell phone type iPhone Other touch-screen phone Length of time with current phone (months) Texts sent per day – weekday, weekend 0–9 10–30 31–50 51–70 71–90 91–110 >110 % of cell phone use (talking/texting) while walking None 75% Experienced collision or near collision while texting and walking
61.3% 22.5 16.2
(2.1) (1.6)
83.9% 3.2% 3.2% 9.7% 28.4
(2.0)
90.3% 9.7% 14.6
(12.4)
0% 45.2% 12.9% 25.8% 3.2% 6.5% 6.5%
0% 22.6% 32.3% 29.0% 6.5% 3.2% 6.5%
0% 32.3% 29.0% 29.0% 9.7% 32.3%
0% 22.6% 51.6% 25.8% 0%
p < .001, h2p ¼ :82]. Overall, gait speed in the real world (M = 1.22 m/s, SD = .15) was faster than in the lab (M = 1.18 m/s, SD = .14) [Environment main effect: F(1,30) = 6.31, p = .02, h2p ¼ :17]; however, the absolute effect size is small and is unlikely to be clinically meaningful. There was no Task Environment interaction for gait speed [F(1,30) = .96, p = .33, h2p ¼ :03] (Fig. 1A). DTEg (no-priority) was not significantly different between the lab and the real world [t(30) = 1.29, p = .21, d = .23]. There were significant effects of Task on texting speed [F(1,30) = 30.59, p < .001, h2p ¼ :51] and accuracy [F(1,30) = 22.67, p < .001, h2p ¼ :43] but no main effects of Environment or interaction effects for either (Fig. 1). Although no-priority DTEt was larger in the real world (M = 17.8%, SD = 22.1) compared to the lab (M = 7.9%, SD = 19.0), the difference was not statistically significant [t(30) = 1.85, p = .08, d = .36]. 3.2. Flexible prioritization of attention in laboratory versus real world
(i.e., power = .80), 26 participants were required. To maximize power to detect effects in Aim 2, we increased the sample size by 20%. 3. Results One participant consistently walked too fast to complete the texting task within the available walkway, and therefore finished texting while standing still. Since his data did not represent true dual-task performance, he was excluded from the analysis. Descriptive characteristics of the 31 participants included in the analysis are presented in Table 1. Pedestrian traffic in the real world was greater than in the lab setting (p < .001); participants encountered an average of 10.9 (SD 6.1) pedestrians in each trial in the Student Center compared to only .5 (SD 1.0) in the lab corridor. None of the participants were required to deviate from their straight-line walking path in the lab, but path adjustment to avoid collisions was required in 44% of gait trials in the real world, with 35% of those involving more than one adjustment (max. 4). Indeed, based on our video analysis, participants looked up from the phone more often in the real world (1.7 2.0 times per trial) than in the lab (.2 .6 times per trial; p < .001). 3.1. Dual-task effects in laboratory versus real world Gait speed while texting (M = 1.09 m/s, SD = .17) was significantly slower than single-task walking (M = 1.32 m/s, SD = .12) [Task main effect: F(1,30) = 133.19,
Instructed prioritization influenced gait speed [F(2,60) = 67.87, p < .001,
h2p ¼ :69]. Gait-priority gait speed was faster than no-priority and texting-priority gait speeds (p < .001; Table 2). Conversely, texting-priority gait speed was slower than no-priority (p < .001; Fig. 2). Instructed prioritization also influenced DTEg [F(2,60) = 68.72, p < .001, h2p ¼ :70], with smaller gait-priority DTEg compared to no-priority and texting-priority DTEg (p < .001); texting-priority DTEg was larger than no-priority DTEg (p < .001). Environment did not influence DTEg [F(1,30) = 1.35, p = .25, h2p ¼ :04], nor was there an interaction [F(2,60) = 1.40, p = .25, h2p ¼ :05]. Instructions did not influence texting speed [F(2,60) = 2.40, p = .09, h2p ¼ :07] (Fig. 2), but texting speed averaged across the three dual-task conditions was slower in the real world than the lab [Environment main effect: F(1,30) = 7.31, p = .01, h2p ¼ :20], with no interaction [F(2,60) = .05, p = .96, h2p < :01]. Texting accuracy was influenced by Instructions [F(2,60) = 3.60, p = .03, h2p ¼ :11], with lower accuracy when participants were instructed to prioritize gait (M = 89.2%, SD = 7.1) compared to texting (M = 92.0%, SD = 5.1; p = .04). Neither gait-priority texting accuracy nor texting-priority accuracy were significantly different from nopriority accuracy (M = 91.1%, SD = 4.9). Instructed prioritization also influenced DTEt [F(2,60) = 3.58, p = .03, h2p ¼ :11]; but Environment did not [F(1,30) = 3.81, p = .06, h2p ¼ :11], with no interaction [F(2,60) = .10, p = .91, h2p < :01] (Fig. 3). To further evaluate the patterns of prioritization between texting and walking in each environment, we plotted the gait and texting DTE against each other. Fig. 3A illustrates the instructed-prioritization effect in both environments and highlights that Environment affected DTEt more than DTEg (i.e., further displacement of DTE between the circle and diamond-shaped symbols on the x-axis than the y-axis). Of note, the 45-degree angle between the two zero lines in Fig. 3A represents equal gait and texting DTE. Based on this, the default prioritization of participants (i.e., nopriority) in the lab was to prioritize texting: DTEt (7.9%) < DTEg (18.5%). In the real world, default prioritization reflected more equal focus: DTEt (17.8%) > DTEg (16.3%).
4. Discussion Our hypothesis that dual-task interference would be greater in the real world than in the laboratory setting was only partially supported by our data. There was no significant difference in the absolute and relative no-priority dual-task effects between the two environments. Thus, without specific instructions about which task to prioritize, young adults perform similarly in high and lowdistraction environments. However, there was a significant effect of Environment on texting speed when averaged across the three dual-task conditions, with slower mean dual-task texting speed in
Fig. 1. Gait speed (A), texting speed (B), and texting accuracy (C) shown as a function of Task (single-task, dual-task no-priority) and Environment (lab, real world). The difference between single and dual-task performance is representative of the dual-task effect. Error bars indicate standard error.
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
49
Table 2 Mean values (SD) absolute and relative dual-task measures of gait and texting performance in each environment, task condition, and instructed prioritization. Real world
Laboratory Single-task
Gait Gait speed (m/s) Gait speed DTE (%) Texting Speed (char/min) Speed DTE (%) Accuracy (%) Accuracy DTE (%) Texting DTE (%)
Dual-task
Single-task
No-priority
Gait-priority
Texting-priority
1.30 (.12)
1.06 (.18) 18.5 (10.9)
1.19 (.15) 9.2 (7.6)
.94 (.17) 27.9 (10.0)
238 (50)
222 (45) 5.5 (15.1) 91.6 (4.2) 2.4 (4.8) 7.9 (19.0)
215 (61) 9.2 (19.3) 89.2 (6.8) 4.9 (7.5) 14.1 (25.5)
229 (53) 2.3 (19.1) 92.7 (4.3) 1.2 (4.7) 3.5 (22.5)
93.9 (3.8)
the real world. The effect of Environment on overall texting dualtask effect was not statistically significant, but the effect size was large [20]: 11% of the variance in DTEt was accounted for by Environment. Since the study was powered for the dependent variable of gait speed, it is possible that there was insufficient power to detect environment effects in texting performance, especially in light of the greater variability in texting performance than gait (as illustrated Fig. 3A and Table 2). The current study is the first to examine how young adults spontaneously allocate their attention while walking and texting. Our data suggest that in the lab young adults prioritized texting, perhaps because there were minimal risks to safety in this setting. This is consistent with previous research in which young adults were observed to allocate most of their attention to the texting task instead of walking [21]. The gait and texting no-priority dual-task effects in the real world signify more equal distribution of attention in that environment. Relative to the lab, dual-task costs on texting increased, while DTEg remained relatively stable. The increase in texting-related dual-task costs in the real world is likely due to the increase in number of times participants looked up from the phone in the real world compared to the lab. Kelly et al. [14] reported that in more challenging gait tasks (narrow-based walking), young adults prioritized walking over the cognitive task, whereas during usual walking, young adults prioritized the cognitive task, despite instructions to prioritize both tasks equally. Our results partially support those of Kelly et al. [14] if one considers walking in the real world to be more challenging than walking in the lab. The reason that we may not have found evidence of walking prioritization in the more challenging situation is that, unlike narrow-based walking, it did not directly increase postural control demands.
Dual-task No-priority
Gait-priority
Texting-priority
1.33 (.13)
1.12 (.15) 16.3 (8.8)
1.21 (.15) 9.4 (7.9)
.98 (.17) 26.2 (10.0)
242 (50)
208 (50) 13.2 (17.2) 90.5 (5.5) 4.6 (6.1) 17.8 (22.1)
202 (53) 15.9 (17.7) 89.1 (7.3) 6.1 (8.2) 22.0 (24.5)
217 (55) 9.6 (17.2) 91.2 (5.9) 3.9 (6.9) 13.4 (22.8)
95.0 (3.1)
Consistent with previous research [14,19,22], we found that young adults can flexibly shift their attention during dual-task walking. We examined whether this ability was influenced by environmental setting. Young adults increased/decreased their gait speed and texting accuracy according to instructed focus, but texting speed did not significantly differ across the variablepriority conditions. In other words, in the texting-priority dualtask condition, participants further slowed their walking to improve their typing accuracy, while maintaining their typing speed. The effects on typing accuracy and not speed might underscore a habitual reliance on autocorrect functions. There was no difference in the effect of instructed prioritization between environments. This is in contrast to Kelly et al. [14] who found that young adults were better able to modulate dual-task gait speed according to instructed focus during usual walking than during more challenging walking. The current findings suggest that walking in the real world may not demand considerably greater attentional resources than walking in the lab, at least for the realworld environment we used, and for healthy young adults. Although the current study provides evidence that young adults can flexibly shift attentional focus, and that this ability is not hampered by the pervasive distractions of a real-world environment, it is unclear whether young adults make the appropriate choices regarding task prioritization in everyday life. It is plausible that, in reality, young adults may choose to stop texting and focus on gait in a very crowded environment, or when safety is more likely to be compromised (e.g., crossing the street). Participants may have felt compelled to continue walking and texting due to the methodological approach and awareness of the experimental situation. Another limitation is that our texting task was not entirely realistic. The abstract phrases (as opposed to self-generated text messages) and
Fig. 2. Gait speed (A), texting speed (B), and texting accuracy (C) shown as a function of instructed prioritization (dual-task no-priority, dual-task gait-priority, dual-task texting-priority) and environment (lab, real world). Error bars indicate standard error.
P. Plummer et al. / Gait & Posture 41 (2015) 46–51
50
Fig. 3. (A) Plot of gait speed and texting dual-task effects (DTE) illustrating prioritization patterns in each dual-task condition in each environment. Circles represent DTE in the lab and diamonds represent DTE in the real world. (B) Gait DTE as a function of environment and priority condition. (C) Texting DTE as a function of environment and priority condition. In all charts, error bars indicate standard error.
absence of autocorrect and autofill operations may have made the texting task relatively more difficult. However, we consider the controlled nature of the texting task to be a minor limitation that is outweighed by the benefits of acquiring objective measures of texting performance.
Acknowledgements The authors gratefully acknowledge K. Simberg and E. Bitterman for assistance with data collection, and B. Najafi and G. Grewal for assistance with data processing.
5. Conclusion This study examined dual-task interference using the most ecologically valid dual-task for healthy young adults, and compared lab-based performance with performance in the actual, everyday environment. Our findings corroborate recent reports that texting produces large dual-task decrements in gait performance [3,12]. Moreover, the current study provides new evidence that dual-task interference associated with texting while walking, and the ability to flexibly prioritize attention between the tasks, are not significantly altered by ubiquitous distractions in the real world. Cell phone use while walking has become a pedestrian hazard [8,9]. The finding that young adults do not spontaneously modify their texting and walking behavior between low-distraction and high-distraction environments substantiates pedestrian safety concerns. Conflicts of interest statement There are no conflicts of interest for any of the authors.
References [1] Plummer-D’Amato P, Brancato B, Dantowitz M, Birken S, Bonke C, Furey E. Effects of gait and cognitive task difficulty on cognitive-motor interference in aging. J Aging Res 2012. 8 pp.. [2] Plummer-D’Amato P, Altmann LJP, Reilly K. Dual-task effects of spontaneous speech and executive function on gait in aging: exaggerated effects in slow walkers. Gait Posture 2011;33:233–7. [3] Lamberg E, Muratori L. Cell phones change the way we walk. Gait Posture 2012;35:688–90. [4] Nasar J, Hecht P, Wener R. Mobile telephones, distracted attention, and pedestrian safety. Accid Anal Prev 2008;40:69–75. [5] Schwebel D, Stavrinos D, Byington K, Davis T, O’Neal E, de Jong D. Distraction and pedestrian safety: how talking on the phone, texting, and listening to music impact crossing the street. Accid Anal Prev 2012;45:266–71. [6] Stavrinos D, Byington K, Schwebel D. Distracted walking: cell phones increase injury risk for college pedestrians. J Safety Res 2011;42:101–7. [7] Hyman Jr I, Boss M, Wise B, McKenzie K, Caggiano J. Did you see the unicycling clown? Inattentional blindness while walking and talking on a cell phone. Appl Cogn Psychol 2010;24:597–607. [8] Richtel M. Driven to distraction. Forget gum. Walking and using phone is risky. New York Times: A12010. New York, NY: The New York Times Company.
P. Plummer et al. / Gait & Posture 41 (2015) 46–51 [9] Nasar JL, Troyer D. Pedestrian injuries due to mobile phone use in public places. Accid Anal Prevent 2013;57:91–5. [10] Hatfield J, Murphy S. The effects of mobile phone use on pedestrian crossing behaviour at signalised and unsignalised intersections. Accid Anal Prevent 2007;39:197–205. [11] Neider MB, McCarley JS, Crowell JA, Kaczmarski H, Kramer AF. Pedestrians, vehicles, and cell phones. Accid Anal Prevent 2010;42:589–94. [12] Schabrun S, van den Hoorn W, Moorcroft A, Greenland C, Hodges P. Texting and walking: strategies for postural control and implications for safety. PLOS ONE 2014;9:e84312. [13] Plummer P, Eskes G, Wallace S, Giuffrida C, Fraas M, Campbell G, et al. Cognitive-motor interference during functional mobility after stroke: state of the science and implications for future research. Arch Phys Med Rehabil 2013;94:2565–74. [14] Kelly V, Eusterbrock AJ, Shumway-Cook A. Factors influencing dynamic prioritization during dual-task walking in healthy young adults. Gait Posture 2013;37:131–4. [15] Najafi B, Aminian K, Loew F, Blanc Y, Robert P. Measurement of stand-sit and sit-stand transitions using a miniature gyroscope and its application in fall risk evaluation in the elderly. IEEE Trans Biomed Eng 2002;49:843–51.
51
[16] Aminian K, Najafi B, Bula C, Leyvraz P, Robert P. Spatiotemporal parameters of gait measured by an ambulatory system using miniature gyroscopes. J Biomech 2002;35:689–99. [17] Aminian K, Trevisan C, Najafi B, Dejnabadi H, Frigo C, Pavan E, et al. Evaluation of an ambulatory system for gait analysis in hip osteoarthritis and after total hip replacement. Gait Posture 2004;20:102–7. [18] Najafi B, Helbostad J, Moe-Nilssen R, Zijlstra W, Aminian K. Does walking strategy in older people change as a function of walking distance? Gait Posture 2009;29:261–6. [19] Kelly V, Janke A, Shumway-Cook A. Effects of instructed focus and task difficulty on concurrent walking and cognitive task performance in healthy young adults. Exp Brain Res 2010;207:65–73. [20] Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Erlbaum, NJ: Hillsdale; 1988. [21] Lopresti-Goodman SM, Rivera A, Dressel C. Practicing safe text: the impact of texting on walking behavior. Appl Cogn Psychol 2012;26:644–8. [22] Yogev-Seligmann G, Rotem-Galili Y, Mirelman A, Dickstein R, Giladi N, Hausdorff JM. How does explicit prioritization alter walking during dual-task performance? Effects of age and sex on gait speed and variability. Phys Ther 2010;90:177–86.
