Part 1: Recruiting and Retaining Volunteers

This part presents research conducted by Brad’s Observing Civic Engagement lab focusing specifically on its Systematic Social Observation (SSO) method. The core argument is that understanding what actually happens inside community-based organizations requires something beyond traditional methods: survey research is broad but shallow; ethnography is deep but narrow; SSO achieves both breadth and depth. The presentation describes a custom SSO App with roughly 1,000 variables used to observe civic life in action across 4 communities, capturing data from 150 organizations across 4,000 observations. Visual slides highlight civic infrastructure, convening locations, and the observers themselves. Key findings presented under the “Recruiting and Retaining Volunteers” theme include insights organized around bridging without bonding, handshakes vs. hugs (relational dynamics), gathering vs. organizing, and collective decision-making.
‍

Part 2: Finding Funders — Classifying Foundations

This part shifts to Brad’s work through Smart Charity and has three core components:

1. Constructing a Foundation-Grantee Network Dataset — Drawing on IRS Form 990
data, the dataset covers approximately 250,000 foundations, 1 million nonprofits, and 2.5
million grants. Four critical enabling factors are cited: a 2015 federal court ruling against
the IRS, the 2019 Taxpayers First Act, 2022 advances in generative AI, and substantial
federal and philanthropic funding.

‍
2. Developing Customizable Organization Classifiers — The project classifies organizations
by both activity and identity. Using entrepreneurship as a worked example, the classifier
analyzes 990 text (mission statements, grant purposes, and program descriptions) using
large language models—including Lemma, Label2Vec, RoBERTa, and BART—combined
into an ensemble-based classification system.

‍
3. Identifying Foundations Most Likely to Fund Your Nonprofit — A foundation is
classified as “entrepreneur-related” (e-related) based on three criteria: having an e-related
mission, having made a grant with an e-related purpose, or having made a grant to a
nonprofit with an e-related mission. The data yields 7,674 total e-related foundations.
‍

Part 2 culminates with a Funder Finder App that operationalizes this classification work for
nonprofits seeking targeted funder matches.