ANN ARBOR—Challenges of ridesharing among low income populations, personalization of health improvement technology, and possible explanations for why we aren't overwhelmed with choice when shopping online are among the research papers University of Michigan School of Information faculty, alumni and current students are sharing this week at the Association for Computing Machinery Special Interest Group on Human-Computer Interaction Conference on Human Factors in Computing Systems in Denver.
No such thing as too much chocolate: Evidence against choice overload in e-commerce
Does this sound familiar: You head down the cereal aisle to find something for the breakfast table only to be overwhelmed by all of the choices, so you buy nothing at all? Or you take home Cheerios and later wish you had purchased Raisin Bran, instead?
Previous research has established something called choice overload when we're doing in-person shopping but U-M School of Information researchers find it basically doesn't exist in the world of ecommerce. The researchers randomly assigned 611 participants a gourmet online chocolate store with 12, 24, 40, 50, 60 or 72 different options and found the number of choices presented did not impact purchase satisfaction.
Researchers say the reason choice overload does not occur online may be because shoppers are used to many options at one time. Additionally, because online shopping has risk associated with not being able to inspect the product, more choices may seem to reduce this risk. The fact that choice overload doesn't exist online is important for ecommerce sites like eBay, Amazon and Target that need not be constrained by consumers' decision-making abilities, said study author Sarita Schoenebeck, U-M assistant professor of information.
Uncovering the values and constraints of real-time ridesharing for low-resource populations
Ridesharing services like Uber and Lyft are considered low-cost transportation alternatives for people without cars or who don't have access to public transportation, but the structure of these services, based on internet technology and the need for a credit card, make them inaccessible to socially disadvantaged low-income people.
U-M School of Information researchers, led by assistant professor Tawanna Dillahunt, studied over time and several rides the experiences of 13 individuals living in transportation-scarce and low-income areas of Detroit—where 40 percent of residents do not own a car—and found positive and challenging aspects of ridesharing. On the plus side, participants had positive interactions with drivers that led to new information about the city, job leads and encouragement. Downsides included mistrust of the companies because of their lack of visibility and physical presence in communities, and discomfort with using smartphones to schedule rides due to the challenges of digital literacy and an inability to stay connected to the internet.
Researchers suggest services that want to serve this population should consider collaborating with local nonprofits to increase awareness, comfort and trust in using these technologies and providing flexible payment methods. The researchers said public kiosks at community businesses could eliminate the need for smartphones while accommodating multiple forms of payment.
Self-experimentation for behavior change: Design and formative evaluation of two approaches
A person who needs to lose weight for better health can find a plethora of apps to help track calories, nutrients and exercise. But what if a program could know, based on self-identification by the user, that during the hours of 7-9 p.m. he mindlessly munches while watching TV?
Personalized support is lacking in most human-computer interaction design, so Matthew Kay, assistant professor of information, and colleagues investigated the design of two complementary strategies for helping users create their own personalized behavior-change plans through self-experimentation.
One strategy emphasized interactive instructional materials and the other introduced context-aware computing to enable user creation of "just in time" home-based interventions. Instead of diet, the subjects reported having sleep problems. Users were able to customize interventions to suggest times to go to bed, stop using tech, control light in the room, cut off caffeine and food consumption, and other very specific actions that helped them improve sleep patterns over the seven-week study period.
Kay said the team's use of Bayesian statistics to combine information from a small feasibility study of 27 people with expert domain knowledge allowed the researchers to make better use of the evidence available.