A Big Data Approach to Mapping the Effects of Transformative Events
How are people transformed by life’s major events? Marriage and divorce, childbirth and death, illness and natural disaster — all can upend people’s lives, altering their outlook and perhaps even their personalities. But not everyone responds to the same upheavals in the same ways. Following a health scare, one person might slide into depression while another might find new sources of joy and gratitude. So how do these major events impact our social worlds and the ways that we cope with change? And what are the psychological characteristics that shape the different ways in which people respond to such events?
Computer scientist Rada Mihalcea from the University of Michigan and social psychologist James Pennebaker from the University of Texas at Austin have teamed up to answer these questions using the latest in big data methods. Supported by a $1.7 million grant from the John Templeton Foundation, the project team will apply computational linguistics and machine learning tools to study social media posts and online writing samples from hundreds of thousands of people to understand changes associated with major events. The project team will make use of publicly available sources as well as private communications that people have agreed to share for research purposes.
What people choose to share online has already proven to be a rich source of naturally occurring behavior from which researchers can make reliable inferences about underlying personality differences. The initial phase of Mihalcea and Pennebaker’s project involves developing and refining algorithms that can identify when certain major events — for example, marriage, birth of a first child, divorce, cancer diagnosis, depression, natural disasters, acts of terrorism — have occurred in people’s lives. The research team will then look at changes in what people write about before and after these events. From these changes, the research team aims to extract signals of change in behaviors including gratitude, kindness, perseverance, and creativity. For example, someone who begins to use religion-related words (like God, pray, church, awe, or faith) with much higher frequency following a major personal event would suggest a change in that person’s spiritual interests and behaviors. The robust data set generated by the project will allow the team — as well as future researchers — to explore how age, sex, personality, culture, and other variables impact how people change in response to upheavals.
“A major challenge in studying how people change in response to major life events — especially unexpected ones — is in how to collect data from people before the events occur,” says Nicholas Gibson, the John Templeton Foundation’s Senior Program Officer for Human Sciences. “Prior approaches are mainly characterized by small sample sizes, lack of prospective data, self-report responses to researcher-driven questions, and lack of longitudinal follow-up. Applying cutting-edge natural language processing methods to people’s everyday writing provides an innovative way of addressing all of these issues, and creates an opportunity for a huge advance in our understanding of how and why people change in the ways they do over the course of their lives.”