Tech Science Seminar


Tuesday, February 5, 2019, 3:30pm to 5:00pm


CGIS Knafel K354

Stressed? Buy This! - Predicting Shopping Behavior with Stress Information Measured by Wearable Devices and the Potential for Unfair Advertising Practices

Stress-shopping, or “retail therapy,” is a common cultural trope. However, with the rise in popularity of wearables and specifically wearables designed to monitor a user’s stress level, can advertisers use this stress information to market to those in need of a little retail therapy? In this experiment, we explore whether stress-related health data should be protected from potentially predatory advertisers. Stress-measuring wearables are marketed to monitor both physical and psychological stress, but their predominant audience is athletes looking at physical stress. Therefore, we wanted to see if these wearables could reliably detect psychological stress by having participants (n=27) wear a device and then induce stress, checking to see if this stress was recorded by the wearable. Next, we wanted to determine if this information would be useful to advertisers by determining if stressed shoppers were more impulsive. We measure impulsivity here by having participants complete a shopping task before and after experiencing the induced stress and examining the number of options that shoppers researched, time they took to purchase, and price of items selected.   

We found that the wearables did reliably and significantly predict a stress response in participants. We also found that participants do spend significantly less time making a decision when they are stressed as compared to control, one measure of impulsivity. Because of these results and our examination of the wearable’s privacy policy, which is ambiguous as to the protection of data, we recommend that this stress data be treated with care and consumers be advised that their data is sensitive.

Stress and wearables paper

Speaker: My name is Michel Li and I am currently a senior (in Dunster House!) concentrating in Government. As someone who recognizes the revolutionary force that technology is, but who is less than technologically proficient myself, I have really enjoyed the classes offered under the new Technology Science track – which allow students to study the impact of technologies and think through ways to resolve any conflicts between technology and society. On campus, I am a member of the Asian American Christian Fellowship and like to babysit. After graduating, I am open to many opportunities, but am currently looking at jobs that directly provide services to youth.    

Polling Place Locations: Do You Travel Farther?

An addition of 0.2 miles to a voter’s travel to the polls may make them less likely to go and cast their ballot. In fact, research shows that those small differences can make big impacts on voter turnout. Does this mean that possibly some groups travel farther than others? That democrats travel farther to the ballot boxes than Republicans? Poor farther than rich? White voters farther than Asian voters? To study this further, I analyzed North Carolina’s 2016 presidential election and quantified the distance that voters have to travel to get to their polling locations using datasets from the 2016 presidential election: 1) voter file data for all North Carolina designated precinct 2) polling place location and precinct information.

Polling place distance paper

Speaker: Jennifer Lee is a junior in Leverett House studying computer science and government. She is currently interested in the intersection of technology, policy, and social impact. Previously, she has studied the security of voting systems and privacy policies of mobile health and fitness applications. On campus, she serves in executive boards of the Harvard Votes Challenge, Asian Student Arts Project, the Seneca, and is involved with the Harvard Financial Aid Initiative to increase socioeconomic diversity on campus.

See also: Seminars