Research

Ongoing Projects

We are actively working on projects that examine how collaborative networks and coordinated service systems address complex social challenges across health & human services, homelessness prevention, and veteran support. Through applied, comparative, and data-driven research, these studies explore how networks improve outcomes for the communities they serve.

SanDiego 211/CIE Preventing Homelessness: An Analysis of HEAP and CIE referrals
In partnership with 211 San Diego's Community Information Exchange, a platform that enables organizations to share client data and make bi-directional referrals, this project analyzes referral patterns under the Homeless Emergency Aid Program to understand how coordinated cross-sector referrals can prevent homelessness.
Health & Human Services and Environmental Purpose-Oriented Networks: A Global Comparison
Purpose-oriented networks bring organizations together around a shared goal that no single one could achieve alone. This project involves an in-depth comparison of how these networks operate, govern themselves, and produce results across countries.
What Do Networks Actually Do? A Study of Toolkits
In collaboration with researchers at the University of Illinois at Chicago, NNSI is analyzing roughly 70 playbooks, frameworks, and templates that guide collaborative networks. The goal is to map what activities these toolkits actually recommend, and how those recommendations vary by social issue, network type, and the kind of organization producing the toolkit.
Evidence Summary: What Do We Know about Health and Human Service Referral Networks as an Intervention?
Health and human service referral networks are increasingly promoted as a way to address health-related social needs, but the evidence base is scattered across disciplines. This project synthesizes what we currently know about referral networks as an intervention — what works, for whom, and under what conditions — to give practitioners, funders, and policymakers a clearer foundation.
Beyond Service Volume: Validating the Trajectories of Care Metric for Long-term Veteran Referral Success
Most veteran referral networks measure success by how many services they deliver, but volume tells you little about whether clients are actually better off over time. This project develops and validates the Trajectories of Care metric — a client-centered, cost-effective way to evaluate how veterans' needs evolve as they move through networks like AmericaServes and Combined Arms.
Predicting Request Resolution in Coordinated Care Networks for Military-Affiliated Populations
Working with AmericaServes data, this project uses predictive analytics to forecast which referral requests are likely to resolve successfully — and which are likely to stall. The aim is to give coordination centers earlier signals about where to intervene, so military-affiliated clients spend less time stuck in the system.

Recent Publications

Authors: Michelle Shumate, Yeha Kim, Marwa Tahboub

The Handbook of Infrastructure Communication

First published: April 2026

When many leaders think of infrastructure, they think about roads, water, electricity, and, increasingly, broadband Internet. This chapter introduces social infrastructure as systems that organize and manage complex health and human service systems of referrals, eligibility criteria, and financial compensation for services. Social infrastructure is grounded in the institutional approach to comparative social welfare policy and relies on coordinated communication across agencies, specializations, and clients to be successful. It includes financial, human resource, resource directory, and technological components that can be arranged in various ways. This framework aims to provide a common language to compare social infrastructure investments across municipalities, states, and countries. In this chapter, we describe several innovations in social infrastructure in the United States. These illustrative cases demonstrate the heuristic value of the social infrastructure conceptualization because they give order to a complex and changing policy landscape.

Link to Article

Authors: Rong Wang, Michelle Shumate, Katherine R. Cooper

Nonprofit and Voluntary Sector Quarterly

First published: March 2026

Nonprofits and public organizations may benefit from using evaluation data to improve programs and managerial decisions. Guided by organizational learning and institutional theory, this study investigates how partners, coalitions, and technical assistance organizations influence these organizations’ professional use of evaluation data within collaborative networks. Data were collected from 449 organizations embedded in an ego-centric network of partnerships and purpose-oriented coalitions on education reform. Results from the hierarchical linear models demonstrated that the only significant source of learning was organizations’ relationships with peer organizations in the ego-centric partnership network. This effect remained significant even after controlling for all the other factors. We discuss our findings and offer implications on integrating learning efforts at the coalition network and ego-centric network levels when motivating nonprofits and public organizations to conduct performance evaluations.

Link to Article

Authors: Michelle Shumate, Joshua-Paul Miles, Marwa Tahboub, Yeha Kim

The Routledge Handbook of Communication and Resilience

First published: September 2025

Health and human service networks are essential to the social care infrastructure of communities and regions. However, maintaining network operations during periods of disruption can be challenging. During shocks (e.g., pandemics, significant migration, and significant budget cuts), the demand for services and the types of needs may shift dramatically. This chapter provides an overview of the research on health and human service networks and resilience. It describes how framing and communication influence resilience processes and outcomes. Finally, it ends with directions for future research on health and human service networks and resilience.

Link to Article

Authors: Yeha Kim, Michelle Shumate, Peixin Hua

Health Communication

First published: May 2025

Health-related social needs are increasingly recognized as significantly influencing healthcare outcomes. Material hardship, defined as the inability to meet basic needs for physiological functioning, such as food and shelter, affects many Americans. This study explores the dynamics of help-seeking behavior, focusing on the perceived effectiveness of different help-seeking channels and the primary barriers associated with each. As part of a larger research project, we interviewed 27 participants aged 18 to 70 to compare help-seekers’ experiences in referral programs (n = 14) with those without such support (n = 13). Our findings indicate that people turn to formal channels for complex needs, expecting tangible solutions. They resort to informal channels for less complicated issues with lower expectations. Because expectations are lower, people often describe informal support as helpful, even when the support fails to adequately address their material hardship. In contrast, individuals find help from organizations ineffective when inadequate or too slow. Stigma significantly impedes help-seeking across both channels. Systemic barriers pose the most significant barrier within formal channels, and social isolation is the most prevalent barrier within informal channels. This research provides insights into the challenges help-seekers face by identifying and understanding the barriers associated with formal and informal help-seeking channels.

Link to Article

Authors: Michelle Shumate, Marwa Tahboub, Qiwen Zhang

Health Services Research

First published: April 2025

Background: With a growing emphasis on building successful health and social care networks, there is a pressing need for effective evaluation metrics. However, past research on quantifying service volume and evaluating referral processing metrics is insufficient for understanding client outcomes over time.

Objective: Our paper introduces the trajectories of care (ToC) metric, a cost–effective approach for referral networks to evaluate service episodes and client outcomes. ToC refers to the dynamic pathways that clients follow while navigating services through the referral network system over time. We answer four questions: (1) What are the typical ToC for veterans? (2) What demographic and dynamic case factors influence ToC? (3) Is there evidence of validity for the metric? (4) Are process metrics related to trajectories?

Methods: Using a mixed–method design, we validate this metric through a series of analyses of client data from AmericaServes and Combined Arms. We categorize service types as basic needs, stressors, and nonessential needs and trajectories of service needs as either positive, neutral, or negative, providing insight into the factors that influence a client’s ToC.

Results: We generated a new measurement evaluating referral networks’ performance and examined its effectiveness compared to existing metrics. We tested three types of validity (concurrent, expert, and discriminant); using Combined Arms client WHO–5 scores, we determined a relationship between clients’ trajectory of care and mental well–being. We compared our results with the practical knowledge of intake specialists at AmericaServes. We then ran an analysis to discover the relations between clients’ success rates and trajectories. The ToC metric is a powerful tool for making referral decisions.

Conclusion: Our paper supports the ToC metric and advocates for reassessing referral networks to holistically account for the complex nature of co–occurring needs that impact veterans’ health and overall quality of life.

Link to Article

Authors: Katherine R. Cooper, Rong Wang, Jack L. Harris, Joshua Paul Miles, Michelle Shumate

Management Communication Quarterly

First published: March 2024

This study examines the evolving role of conveners in sustaining purpose-oriented interorganizational networks, particularly during periods of organizational change. While prior research has largely focused on the early stages of network formation, less attention has been given to how conveners manage change over time and contribute to network resilience. Using qualitative analysis, this study draws on two waves of interview data collected from conveners of 26 cross-sector collaborative networks focused on education reform in the United States. The first round of interviews was analyzed using attribute and provisional coding to examine network composition, activities, and motivations for formation, while the second round explored changes experienced after initial convening and convener responses through provisional and magnitude coding. Findings identify varying magnitudes of network change—incremental, radical, and intermediate—that shaped how conveners responded and adapted. The study also identifies two distinct founding orientations among networks: problem-oriented and solution-oriented. Networks founded with a problem orientation demonstrated greater resilience than those founded with a solution orientation, partly due to the broader repertoire of activities employed by their conveners. Overall, the findings highlight the central role of conveners not only in establishing purpose-oriented networks but also in guiding them through change. The study contributes the concept of founding orientations and offers theoretical and practical insights into building resilient collaborative networks.

Link to Article

Authors: Rong Wang, Michelle Shumate

Nonprofit Management and Leadership

First published: December 2023

Increasingly, nonprofits and corporations publicly communicate about their partnerships. Guided by Information Integration Theory, this paper examines how information about a nonprofit’s relationship with a corporation relates to individuals’ intention to donate and volunteer. This research used a two-study experimental design. Study 1 (N = 966) examined how partnership explanations and evaluation were related to the two outcomes. Study 2 (N = 970) further examined whether specific information about partnerships, including type, duration, and communication source, was integrated with existing knowledge to relate to the outcomes. Partnership evaluation consistently related to stakeholders’ intention to support nonprofits, and it mediated the effect of partnership explanations on the intention to volunteer. Furthermore, partnership type was significantly related to the two outcomes, while duration and source of communication were not.

Link to Article

Authors: Michelle Shumate, Shaun M. Dougherty, Joshua-Paul Miles, Anne-Marie Boyer, Rong Wang, Zachary M. Gibson, Katherine R. Cooper

Journal of Public Administration Research and Theory

First published: March 2023

Increasingly, scholars and practitioners are interested in evaluating the effectiveness of interorganizational networks. We use a configuration approach to study network effectiveness. This research is a mixed-method study of 26 education networks in the United States. We measure network effectiveness by comparing 4th-grade literacy, 8th-grade literacy, and high-school graduation rates. We compare these scores with all school districts in the state using interrupted time series or parametric difference-in-differences approaches. Then, drawing from qualitative data from interviews and archives, we investigate the network governance, environmental characteristics, and theories of change associated with greater student achievement. We find three configurations associated with network effectiveness using fuzzy-set qualitative comparative analysis. One configuration combines decentralized governance with a project theory of change in the context of resource munificence. A second configuration associated with network effectiveness is to combine learning and systems alignment theories of change with smaller network size and resource munificence. The final configuration combines decentralized governance, a learning theory of change, less resource munificence, and larger network size and does not use a systems alignment theory of change. The results support the configurational approach, which suggests multiple configurations of factors in combination may result in network effectiveness.

Link to Article