Wearable Bluetooth Sensors for Capturing Relational Variables and Temporal Variability in Relationships: A Construct Validation Study

• 成果类型: Article
• 署名作者: Matusik, James G.; Heidl, Ralph; Hollenbeck, John R.; Yu, Andrew; Lee, Hun Whee; Howe, Michael
• 署名单位: Michigan State University; University of Oregon; Iowa State University
• 刊物名称: JOURNAL OF APPLIED PSYCHOLOGY
• ISSN/ISSBN: 0021-9010
• DOI: 10.1037/apl0000334
• 发表日期: 2019
• 页码: 357-387
• 关键词: wearable sensors; Bluetooth; convergent validity; predictive validity; Network dynamics

摘要

The advent of wearable sensor technologies has the potential to transform organizational research by offering the unprecedented opportunity to collect continuous, objective, highly granular data over extended time periods. Recent evidence has demonstrated the potential utility of Bluetooth-enabled sensors, specifically, in identifying emergent networks via colocation signals in highly controlled contexts with known distances and groups. Although there is proof of concept that wearable Bluetooth sensors may be able to contribute to organizational research in highly controlled contexts, to date there has been no explicit psychometric construct validation effort dedicated to these sensors in field settings. Thus, the two studies described here represent the first attempt to formally evaluate longitudinal Bluetooth data streams generated in field settings, testing their ability to (a) show convergent validity with respect to traditional self-reports of relational data; (b) display discriminant validity with respect to qualitative differences in the nature of alternative relationships (i.e., advice vs. friendship); (c) document predictive validity with respect to performance; (d) decompose variance in network-related measures into meaningful within-and between-unit variability over time; and (e) complement retrospective self-reports of time spent with different groups where there is a ground truth criterion. Our results provide insights into the validity of Bluetooth signals with respect to capturing variables traditionally studied in organizational science and highlight how the continuous data collection capabilities made possible by wearable sensors can advance research far beyond that of the static perspectives imposed by traditional data collection strategies.