Charles F. and Joanne Knight Alzheimer Disease Research Center, Department of Neurology, Washington University School of Medicine, St. Louis, MO, USA.
, Mountain View, USA.
Behav Res Methods. 2023 Sep;55(6):2800-2812. doi: 10.3758/s13428-022-01925-1. Epub 2022 Aug 11.
Studies using remote cognitive testing must make a critical decision: whether to allow participants to use their own devices or to provide participants with a study-specific device. Bring-your-own-device (BYOD) studies have several advantages including increased accessibility, potential for larger sample sizes, and reduced participant burden. However, BYOD studies offer little control over device performance characteristics that could potentially influence results. In particular, response times measured by each device not only include the participant's true response time, but also latencies of the device itself. The present study investigated two prominent sources of device latencies that pose significant risks to data quality: device display output latency and touchscreen input latency. We comprehensively tested 26 popular smartphones ranging in price from < $100 to $1000+ running either Android or iOS to determine if hardware and operating system differences led to appreciable device latency variability. To accomplish this, a custom-built device called the Latency and Timing Assessment Robot (LaTARbot) measured device display output and capacitive touchscreen input latencies. We found considerable variability across smartphones in display and touch latencies which, if unaccounted for, could be misattributed as individual or group differences in response times. Specifically, total device (sum of display and touch) latencies ranged from 35 to 140 ms. We offer recommendations to researchers to increase the precision of data collection and analysis in the context of remote BYOD studies.
是允许参与者使用自己的设备，还是为参与者提供专门用于研究的设备。自带设备（Bring-your-own-device，BYOD）研究具有几个优势，包括提高了可及性、潜在的更大样本量以及降低了参与者负担。然而，BYOD 研究几乎无法控制可能影响结果的设备性能特征。特别是，每个设备测量的响应时间不仅包括参与者的真实响应时间，还包括设备本身的延迟。本研究调查了两个对数据质量构成重大风险的显著设备延迟源：设备显示输出延迟和触摸屏输入延迟。我们全面测试了 26 款价格从 < 100 美元到 1000 多美元不等的流行智能手机，以确定硬件和操作系统差异是否导致了明显的设备延迟可变性。为了实现这一目标，我们使用一种名为 Latency and Timing Assessment Robot（LaTARbot）的定制设备来测量设备显示输出和电容式触摸屏输入延迟。我们发现智能手机在显示和触摸延迟方面存在相当大的差异，如果不加以考虑，这些差异可能会被错误地归因于响应时间的个体或群体差异。具体来说，总设备（显示和触摸之和）延迟范围为 35 到 140 毫秒。我们为研究人员提供了一些建议，以提高远程 BYOD 研究中数据收集和分析的精度。
