Modeling social support scores using phone use patterns

• Published in: Proceedings of the Association for Information Science and Technology Vol. 55; no. 1; pp. 133 - 142
• Main Authors: Ghosh, Isha, Singh, Vivek K.
• Format: Journal Article
• Language: English
• Published: 2018
• Subjects: Data Science; Phone Mediated Communication; Social Informatics
• ISSN: 2373-9231; 2373-9231
• DOI: 10.1002/pra2.2018.14505501015

ABSTRACT
Social support is a fundamental concept of interactional behavior and has been consistently linked with higher levels of emotional and physical well‐being. Social support has been shown to be an effective coping mechanism in managing stress and depression and to provide positive experiences that directly enhance overall well‐being and quality of life. The measurement and quantification of social support has generated a great deal of interest in the behavioral science community over the past 10 years. The dominant approach for quantifying individual social support remains self‐reported surveys and generator‐methods, which are costly, attention‐consuming, and fraught with biases. Given the important role played by mobile phones in mediating human social lives, this study explores the use of phone metadata (call and SMS logs) to automatically infer an individual's social support. Based on Sherbourne and Stewart's Modified Social Support survey (MSSS) as ground truth and ten‐week phone data collection for 55 participants, we report that multiple call and SMS based social features are intrinsically associated with social support and perform better than DEMOGRAPHY‐BASED models.