East Coast Ballroom
WITHDRAWN - Early Detection of Health Risk Factors: Leveraging Online Data from Crowdsourcing Platforms (307902)
*Samantha Robinson, University of ArkansasKeywords: Amazon Mechanical Turk (MTurk), TurkPrime, MicroBatch, Crowdsourcing, Perceived Stress, Substance Abuse, Obesity, Mental Illness
Despite continued media coverage and public interest in those epidemics that dominate current health policy initiatives (e.g., opioid abuse, obesity, mental illness, etc.), stress is an often overlooked health risk factor. Stress, having an adverse effect on overall health, has been linked to substance abuse, sleep deprivation, obesity, and depression. Consequently, it is imperative to detect at-risk communities, regions where observed levels of stress are excessive. Early detection of stress as a health risk factor on a regional level is difficult due to the cost prohibitive nature of data collection. However, given the widespread use of crowdsourced online samples in consumer and social science research recruited from popular platforms such as Amazon Mechanical Turk (MTurk), we propose that the leveraging of survey data from such platforms (when done carefully and appropriately) can provide policy makers with cross-sectional snapshots of regional health risk efficiently and at low cost. Using panel data collected from MTurk via TurkPrime with micro-batching enabled, current work explores trends in the stress of both adults and their children at the country and regional level.