Pamela Paxton, a professor of sociology at The University of Texas at Austin, is taking a fresh approach to studying donations and volunteering: looking at the data from electronically-filed Form 990s that U.S. nonprofit organizations submit to the Internal Revenue Service (IRS). Two years ago, her current research would have been impossible.
The IRS began releasing free and searchable electronically-filed Form 990 data to the public via Amazon Web Services in 2016, following a successful lawsuit by Public.Resource.org and pressure from the Aspen Institute and nonprofit research groups such as GuideStar, Urban Institute, Foundation Center, Indiana University and Johns Hopkins University. Since then, Paxton’s undertakings reveal that the data are key to answering critical questions social scientists aim to address and understand.
“Do nonprofits in a community increase well-being? What increases volunteering rates?” Paxton asked. “We can use the data to understand how nonprofits matter, and where and when they thrive.”
Exploring nonprofit mission statements is one way Paxton’s team works to do this. Using computerized text analysis techniques, Paxton’s team is making nonprofit mission statements usable by coding emotional language, extracting topics, and tracking changes in wording over time. They are connecting features of nonprofit mission statements, such as language that elicits positive or negative emotion, to the donations those nonprofits receive, and the number of volunteers they attract.
“We rarely have the opportunity to work with detailed information about nonprofits’ work in a community, but with mission statements, and open data, we can get very specific and unique information,” Paxton explained. “It’s just phenomenal for academic research. Now, the sky’s the limit.”
The electronically filed 990s allow Paxton’s team to compare volunteer rates reported from Form 990 data to the volunteer rates that traditional social science surveys report. In doing so, her team can explore how bias skews surveys.
“People with pro-social tendencies are already more likely to answer surveys,” Paxton said. “We need to be exploring alternative sources of information – such as administrative records, social media, and more. Open data allows for that to become a reality.”
According to Paxton, considering such alternative sources of information has become crucial, especially at a time when conducting traditional social science surveys is increasingly difficult and expensive. A recent special reportfrom the National Academy of Science reveals that even the most reputable surveys are experiencing declines in participation.
As for future research prospects, with millions of electronically-filed nonprofit tax records on file and 67 percent of all 990s now being filed electronically, social science researchers have a larger number of variables to study and review. Meanwhile, to promote greater understanding of the nonprofit sector as a whole, the Aspen Institute’s Program on Philanthropy and Social Innovation and its partners continue to press for mandatory electronic filing of nonprofit tax forms and the release of that data by the IRS – a proposal that unanimously passed the U.S. House of Representatives, H.R. 5443, in April 2018.
