Source: CORNELL UNIVERSITY submitted to NRP
SOCIAL INFLUENCE AND HEALTH: A SOCIAL NETWORK APPROACH TO HEALTHY LIVING
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
1000494
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Oct 1, 2013
Project End Date
Sep 30, 2015
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
CORNELL UNIVERSITY
(N/A)
ITHACA,NY 14853
Performing Department
Communication
Non Technical Summary
People are more successful at weight loss when they lose weight with friends or family. Our ties have a powerful influence on health, but few studies show how networks contribute to changes in health behavior. Social media websites provide new opportunities for people to communicate about health and allow scholars to understand social influence in ways that were formerly unthinkable. The proposed project will use a mobile social media app, VERA, developed at Cornell's Interaction Design Lab to study social influence of health for college students, who are transitioning to adulthood and beginning to develop a personal repertoire of health behavior. A series of experiments will show how social influence of healthy and unhealthy behaviors occurs in social networks and how stress and emotions affect this process. Understanding this process is important for public health programs and healthcare professionals dedicated to helping people adopt a lasting healthy lifestyle by providing insights about leveraging social relationships to help people make healthy behavior changes.
Animal Health Component
80%
Research Effort Categories
Basic
20%
Applied
80%
Developmental
0%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
72460203070100%
Knowledge Area
724 - Healthy Lifestyle;

Subject Of Investigation
6020 - The family and its members;

Field Of Science
3070 - Psychology;
Goals / Objectives
Exploring social influence processes behind health behavior change and adoption is important to understanding how people develop patterns of healthy and unhealthy behavior and how social networks contribute to the adoption of these behaviors. The objectives of this proposal are to: Use social media as a platform for studying how social influence affects health behaviors developed by college students (e.g., diet, exercise). In particular, this research seeks to better understand how peers' healthy and unhealthy behaviors influence the health behaviors that people report using the VERA social mobile phone application developed in our lab at Cornell University. Understand how emotions and stress affect social influence processes for health behavior. This research will be important for designing interventions that incorporate people's social relationships to encourage healthy behaviors while finding ways to minimize the negative impact that peers' unhealthy habits have on behavior.
Project Methods
To study social influence of health behaviors we developed a mobile health application called VERA+, which integrates EMA methods with a social networking tool to understand how health behaviors spread in networks. VERA+ allows for tracking of health behaviors while integrating this data across users in a network, illuminating patterns of behavior between users. VERA+ users photograph, comment and rate healthiness of daily health behaviors (e.g., food choices, exercise) and share this information with peers. Importantly, given the sensitive nature of health behaviors, VERA+ allows users to share as anonymous users. Two field experiments were conducted to answer our hypothesis and research question. The first explored social influence of healthy behaviors and the second explored social influence of unhealthy behaviors. The experimental design, procedure and measures were identical for both studies. Differences in the experimental treatment conditions used for each study are described below. Experimental design Participants in both studies will be placed in a series of small, anonymous social networks (between 10 and 12 users), which included 4 to 6 fellow participants and 5 confederate participants. Participants will be told that all other users were fellow undergraduate participants from their university and confederate users appeared to be fellow participants. In actuality, confederate users will be created using a simulated profile and posts from a corpus of VERA+ posts created by participants in previous, exploratory studies. Posts will be cleaned and sent at the appropriate time of day to ensure that the health behaviors that confederates posted were realistic given time of day, year and their status as undergraduate students. Experimental treatment conditions In Study 1, participants will be randomly assigned to one of two experimental treatment conditions: 1) healthy confederates (n= 50), or 2) control confederates (n= 50). In the healthy confederates' condition, confederate participants will share posts conveying mostly healthy behaviors. Confederate posts will be pulled from previous studies and reviewed for each confederate user to ensure that they appeared to come from the same individual. Confederate users will post healthy behaviors 75 percent of the time, with healthy posts defined as those rated by former users as 2 to 3 on a -3 to 3 scale of healthiness where -3 was very unhealthy and 3 was very healthy. The remaining 25 percent of posts will be mildly unhealthy to healthy posts (rated between -1 and 1). These mild posts were included to ensure that confederate users appeared to be actual users. In the control condition, confederates will post mildly unhealthy to healthy posts from former users (rated between -1 and 1) 100 percent of the time. In Study 2, networks will mirror those used in the Study 1 with one important difference. Confederate users will share posts that conveyed unhealthy, rather than healthy, behaviors (n= 33). Confederates will post unhealthy behaviors (rated -2 or -3) 75 percent of the time and mildly unhealthy to healthy behaviors (rated -1 to 1) 25 percent of the time. Control networks (n=28) will mirror those in Study 1. In both experiments, participants will download VERA and be added to one of a series of anonymous social networks, which consist of 50% confederates (i.e., computer-generated users who appear to be participants) and 50% participants. Participants will post about health behaviors for 3 weeks. They will: photograph a behavior, rate its healthiness, rate stress level and emotion (Pollak, Adams & Gay, 2011), and share behavior with the network. Participants can interact with network members by viewing photos, commenting on and "liking" behaviors.

Progress 10/01/13 to 09/30/15

Outputs
Target Audience:We enrolled approximately 110 18-20 year olds in our longitudinal studies of nutrition, social support, and nutrition monitoring on the cell phonesand 500 for our online studies. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?Two graduate students and a postdoc worked on designing the mobile health app as well as data collection and analysis of the use of social media for health interventions. Six undergraduate researchers also helped us collect data from participants, sent reminders to participants and helped analyze data from the 3 studies. How have the results been disseminated to communities of interest?We have presented the results at several conferences. There are 3 publications (see publications section) and our mobile apps are in the public domain and being used by other researchers. Publications and links to the apps are available on our website. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? We developed and tested mobile apps thatencourage healthy decisions by drawing on open-ended social awareness - making users aware of both others' and their own decisions. Building on usage data and rich interviews from previous projects, Vera+ affords users greater control over sharing and consuming health decisions and offers deeper understanding via improved data collection and complete activity logging. We learned that social support provides an important role in health systems. Short and bursty interactions provide greater esteem support which, in turn, builds self-confidence and respect. Users of the app tended to follow up with a positive health activity (from tracking info data) after receiving short positive messages. Short positive interactions and social support also helps reduce an individual's uncertainty, lack of perceived personal control, or stress. As millions of people turn to social media for health information, better understanding the factors that guide health related judgments and perceptions in context is imperative. We also conducted two Web experiments where we found widespread weight bias and related stereotypes to influence nutrition judgments in social media spaces. In Experiment 1, meals were judged as lower in nutritional quality when the person who recommended them was depicted as obese rather than of normal weight, an effect mediated by stereotypic beliefs about the source as a generally unhealthy person. Experiment 2 replicated this effect, which remained significant when controlling for objective nutritional information (calories and fat content). Results highlight the spillover effects of weight: bias that extend beyond the person to color impressions of objects (nutrition and food choices).

Publications

  • Type: Journal Articles Status: Awaiting Publication Year Published: 2016 Citation: Guillory, J., Adams, P. & Gay, G. (2016) Social support in a mobile health application. (under review). Pervasive Health. Schuldt, J., Guillory, J., & Gay. G. (2016). Prejudice and the plate: Effects of weight bias in nutrition judgments. Health Communication. 31: 182-192.


Progress 10/01/13 to 09/30/14

Outputs
Target Audience: The target audience is college students interested in improving their healthy behaviors. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? The project has provided the opportunity to mentor and train postdocs and graduate researchers, as well as provided opportunities for students to share the work at professional conferences. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? We plan to continue to explore this exciting area of research, as well as find a publication venue for the social influence confederate study described above.

Impacts
What was accomplished under these goals? We have conducted studies exploring social influence processes behind health behavior change. We explored socially supportive messages in VERA, a mobile application for sharing health decisions and behaviors. The short and bursty interactions in social awareness streams afford a particular style of social support, for which we offer the label staccato social support. Results indicate that, in comparison to previous work, staccato social support is characterized by a greater prevalence of esteem support, which builds respect and confidence. We further note the presence of 'following up', a positive behavior that contributes to supportive interactions, likely via social pressure and accountability. These findings suggest design recommendations to developers of mobile social support systems and contribute to understanding technologically mediated social support for health. We alsoinvestigated the social influence of healthy and unhealthy behaviors using the VERA mobile application designed to have people track their health behaviors and share them within a social network. In two field experiments, participants spent 3 weeks using a mobile application to photograph, rate the healthiness of daily health behaviors, and share this information with a small network of other users. In order to study social influence processes, confederate users (i.e., fake users posting a compilation of behavior posts from previous exploratory studies using the same mobile application) represented approximately half of the other users in participants' networks. Confederate users in Study 1 shared behaviors perceived to be very healthy by users in previous studies and confederate users in Study 2 shared behaviors perceived to be very unhealthy by users in previous studies. Findings indicate that people interacting with confederates who shared very healthy behaviors posted healthier behaviors to the network over time whereas the healthiness of behaviors shared did not change for people who interacted in networks with confederates who shared very unhealthy behaviors. Group and cultural norms as well as self-presentation are discussed as theoretical mechanisms driving social influence in these studies.

Publications

  • Type: Journal Articles Status: Accepted Year Published: 2015 Citation: Schuldt, J., Guillory, J., & Gay, G. K. (2015). Prejudice and the plate: Effects of weight bias in nutrition judgments. Health Communication.
  • Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Adams, P., Baumer, E., and Gay, G. (2014). Staccato social support in mobile health applications. Proceedings of CHI 2014.