Social Media

Since it is expensive to estimate daily sentiment across the United States using a survey, publicly available updates on social media provide a low-cost alternative. Using a large set of geolocated tweets requested through Twitter API, this project aims to measure daily variation in expressed sentiment across the United States as a proxy for mental health status in response to daylight savings time (DST). A specific research project in these directions is as follow:

Manuscript in preparation

Chin, S., Zahran, S., & Mushinski, D. Daylight Time Saving Transitions and Expressed Sentiment: Evidence from Twitter

Daylight savings time (DTS) represents a public good with costs and benefits. We provide the first comprehensive examination of the welfare effects of the spring and autumn transitions for the U.S. Using geocoded Twitter data and a regression discontinuity design, we estimate the effect of the transitions on expressed sentiment. Our results show that individuals experience deteriorations in expressed sentiment in the first week after the spring transition. We find no effect of the autumn transition. We attribute the negative effect of the spring transition to the reduction in the time endowment and the process of adjusting to the disruption in circadian rhythms. We conclude that the higher the shadow price of time, the more difficult is adjustment. Presumably, an increase in flexibility to reallocate time could reduce the welfare loss for individuals with binding time constraints.