THE DIFFUSION OF CIVIC TECHNOLOGY:

HOW TECHNOLOGY AND DEVOLUTION ARE RESHAPING CIVIC LIFE IN URBAN AMERICA

by

Eli Turkel

A dissertation submitted to the Faculty of the University of Delaware in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Urban Affairs and Public Policy.

Spring, 2020

© 2020, Turkel All Rights Reserved

THE DIFFUSION OF CIVIC TECHNOLOGY:

HOW TECHNOLOGY AND DEVOLUTION ARE RESHAPING CIVIC LIFE IN URBAN AMERICA

by

Eli Turkel

Approved: ______Maria P. Aristigueta, D.P.A. Director of the Joseph R. Biden, Jr. School of Public Policy and Administration

Approved: ______John A. Pelesko, Ph.D. Dean of the College of Arts and Sciences

Approved: ______Douglas J. Doren, Ph.D. Interim Vice Provost for Graduate and Professional Education and Dean of the Graduate College

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______John McNutt, Ph.D. Professor in charge of dissertation

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______Jonathan Justice, Ph.D. Member of dissertation committee

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______Tibor Toth, Ph.D. Member of dissertation committee

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______John Stephens, Ph.D. Member of dissertation committee

ACKNOWLEDGMENTS

I would like to thank the faculty and staff of the Biden School for all of their guidance and support over the last four years. Additionally, I want to thank my dissertation committee for all of their guidance and support, including Dr. Johnathan

Justice, Dr. John Stephens, and Dr.Tibor Toth. I met Dr. Stephens at the Code for America Brigade Congress in the fall of 2017. I look forward to seeing Dr. Stephens at future Code for America events. To my chair, Dr. John McNutt, thank you for all of the lunches and talks we had throughout this process. You taught me a tremendous amount about the research process. I also want to thank the Code for America and Open Data Delaware communities. I am grateful to be a part of this movement and to work with all of you. Finally, I would like to thank my family for all of their support during this process and throughout my life. Earning a Ph.D. is something I have aspired to for a long time. The love and guidance of my family allowed me to reach this goal.

iv

TABLE OF CONTENTS

LIST OF TABLES ...... viii LIST OF FIGURES ...... ix ABSTRACT ...... x

Chapter

1 INTRODUCTION ...... 1

1.1 Research Questions ...... 3

2 LITERATURE REVIEW ...... 5

2.1 Civic Technology As Civic Participation ...... 5

2.1.1 Definitions of Participation ...... 7 2.1.2 Wither Participation? ...... 9

2.2 Historical Antecedents of Civic Technology ...... 10

2.2.1 Community Networks...... 11 2.2.2 Community Technology ...... 13

2.3 Components of Civic Technology ...... 15

2.3.1 Open Civic Data ...... 17 2.3.2 Technology ...... 18 2.3.3 Collaborative Processes ...... 19

2.3.3.1 Civic Hacking and Hackathons ...... 20 2.3.3.2 Code For America ...... 23 2.3.3.3 Code For America Brigades ...... 26

2.4 Civic Technology As a Social Movement ...... 28 2.5 Related Social Movements ...... 33

2.5.1 Open Government ...... 34 2.5.2 Tech4Good ...... 38

v

2.5.3 Data4Good ...... 39 2.5.4 Evidence Based Practice ...... 39 2.5.5 Online Protest Movemments...... 40 2.5.6 Participatory budgeting...... 41

2.6 Conclusion ...... 42 2.7 The Consequences Of Diffusion ...... 42

2.7.1 The Diffusion Context ...... 44

3 THEORETICAL FRAMEWORK...... 47

3.1 Propositions ...... 50

3.1.1 Traditional Social Capital ...... 51 3.1.2 Creative Class ...... 52

3.2 Central Hypotheses ...... 52 3.3 Covariants ...... 54

3.3.1 Distress Score ...... 54 3.3.2 Networked Social Capital ...... 55 3.3.3 Broadband ...... 55 3.3.4 Population ...... 56 3.3.5 Education ...... 56

4 METHODOLOGY ...... 59

4.1 Subjects...... 59 4.2 Design Structure ...... 62

4.2.1 Variables and Data Sources ...... 62 4.2.2 Dependent Variable Code For America Brigades ...... 62 4.2.3 Independent Variables ...... 64 4.2.4 Traditional Social Capital ...... 64 4.2.5 Creative Class ...... 67 4.2.6 Covariates...... 69

4.2.6.1 Education ...... 69 4.2.6.2 Population ...... 69 4.2.6.3 Distressed Community Index ...... 70 4.2.6.4 Networked Social Capital Organizations ...... 70 4.2.6.5 Broadband...... 72

vi

4.3 Data Analysis...... 72 4.4 Limitations ...... 74

5 RESULTS ...... 76

6 DISCUSSION...... 83

6.1 Findings ...... 85 6.2 Conclusions ...... 88

REFERENCES ...... 92

Appendix

A CODE BOOK ...... 106 B CREATIVE CLASS CODES ...... 107

vii

LIST OF TABLES

Table 1. Social Capital Original Variables ...... 66

Table 2. Creative Class Original Variables ...... 69

Table 3. Networked Non-profit Organizations ...... 70

Table 4. Broadband Original Variables ...... 72

Table 5. Descriptive Statistics & Bivariate Correlations ...... 77

viii

LIST OF FIGURES

Figure 1. Cleveland Free-Net Home page...... 12

Figure 2. Cleveland Free-Net main menu...... 13

Figure 3. Components of Civic Technology...... 16

Figure 4. Open Data Delaware, Data Jam 2017...... 21

Figure 5. GetCalFresh...... 25

Figure 6. MobiliDE Launch Page ...... 27

Figure 7. Delaware Open Data Portal ...... 38

Figure 8. Hypothesis matrix, adapted from McNutt and Justice, 2016 ...... 53

Figure 9. Census Data Hierarchy...... 61

Figure 10. Code for America Brigades as of August, 2018 ...... 63

Figure 11. Social Capital by County ...... 65

Figure 12. Creative Class by County, USDA ...... 68

Figure 13. Receiver operating characteristic (ROC) Curve...... 81

ix

ABSTRACT

Over the last decade, urban communities in the United States and around the world have seen the diffusion of civic technology. Civic technology, or “technology for the public good”, takes various forms. In this case, civic technology refers to volunteer activities that use open civic data, technology, and innovative practices. Within the United States, Code for America is a particularly active agent in the work to diffuse civic technology. Code for America has a volunteer arm known as Code for America Brigades. Using Code for America Brigades a proxy for civic technology this dissertation asks the question, who is adopting civic technology and why? Since de Tocqueville, academic literature on American democracy has argued that civic association is a central feature of the success of American society. Though there is evidence that civic participation in American civil society is in decline, civic technology represents a new form of civic association. Another feature of American life in recent decades is economic polarization at multiple scales, including regional.

Using urban U.S. counties as the study population this dissertation asks a series of questions regarding the characteristics of urban counties that have and have not adopted Code for America Brigades. The dissertation finds that the diffusion of Code for America Brigades is not a function of population, but rather a function of socioeconomic organization. That is, Code for America Brigades are not strongly associated with traditional forms of social capital nor are they associated with economically distressed areas. Rather, the diffusion of Code for America Brigades is

x

associated with economically prosperous regions with high concentrations of creative class workers. The dissertation concludes that the creative class is a group with a shared incentive structure is supportive of mobilization in Code for America Brigades.

Young, talented, tech-savvy, career-oriented individuals seek out the types of professional development and career advancement opportunities that affiliation with Code for America provides. In order to expand the civic technology movement, it is incumbent on Code for America and related organizations to build network structures that incentivize the involvement of organizations that are mobilized by differentiated incentive structures.

xi

INTRODUCTION

In his classic work, Democracy in America, Alexis de Tocqueville (1966) observed that among all nations the principle of association is strongest among the Americans. The principle of association, de Tocqueville argued, animates religious, social, and civic life in the United States. That is, in all these various aspects of community life, Americans look to each other to share beliefs and ideas that form the basis of building institutions and making change. In recent years, the American civic association has changed. In a wide variety of examples, technology now plays a central role in American community building. One example is civic technology (McNutt, 2018; Schuler, 1996) which represents an important development in American communities and communities throughout the world.

Practitioners such as Micah Sifry, Matt Stempeck, and Erin Simpson, along with scholars such as Graeff (2018), Hou (2018), McNutt (2016), and Stephens (2018) agree that the term refers to a field of practice focused on producing technologies geared toward improving civic life. In collaboration with Micah Sifry and Erin Simpson, Matt Stempack defines civic technology as “the use of technology for the public good” (Stempack, 2016, 1). These overarching definitions do not convey principles of the civic technology movement, but it does point to two clear aspects of the movement: It is focused on technology, and it is focused on providing a societal benefit. The literature review section of the dissertation will explain how the

1

movement evolved and will convey varied approaches to fulfilling Stempack’s overarching definition. After discussing conceptual issues related to what constitutes the field of civic technology, the dissertation will theorize about the diffusion of the civic technology movement. The technological revolution of the second half of the twentieth century significantly contributed to reshaping American communities. As numerous authors argue, American communities during the middle decades of the twentieth century were evenly distributed in terms of education, wealth, and partisanship. That is no longer the case (Berry & Glaeser, 2005; Bishop & Cushing, 2008). As American communities polarize, patterns of association change as well. While there are communities with associative patterns similar to those of the mid-twentieth century, now there are communities with vastly different patterns of association (Rupasingha, Goetz & Freshwater, 2006). Measuring existing patterns of association requires new methodologies and data sources than those used to measure the civic and political life of post-World War II America. A crucial aspect of the debate about the decline of community in the United States was the rise of the post-industrial economy. Beginning in the mid-1970s, the growth of the information sector of the economy, combined with globalization, has devastated many communities with an industrial economic base. There were concurrent developments in other communities that were far more hopeful. Richard Florida (2012) refers to one pool of post-industrial economy workers as the Creative Class – clustered in well-known prosperous metropolitan enclaves such as Silicon Valley, Austin, and Washington D.C. Whether these individuals and organizations were drawn to these places by their creative milieu (Porter, 1990),

2

tolerant cultures and diverse demographics (Florida, 2012), flexible work cultures

(Saxenian, 1996), or sheer resources (Markusen, 1991, 1996) is not the focus on this research. Rather, the focus is on how the reshuffling of the American workforce is now affecting civic life. In particular, this dissertation tests knowledge built by the diffusion of innovation framework in order to investigate emerging patterns of experimental forms of civic life referred to as civic technology. The overarching research question posed by this dissertation is, who is adopting civic technology, and why?

It is clear that civic technology is one element of a new urban setting. It is therefore important to better understand this growing phenomenon. In that spirit, the following research questions are proposed:

1.1 Research Questions

RQ1 Do higher levels of traditional social capital predict higher levels of civic

technology use in urban United States counties?

RQ2 Do higher levels of creative class employment predict higher levels of civic

technology use in urban United States counties?

RQ3 Do higher levels of advanced education predict higher levels of civic

technology use in urban United States counties?

RQ4 Do larger populations predict higher levels of civic technology use in urban

United States counties?

RQ5 Do higher levels of networked non-profit arrangements predict higher levels

of civic technology use in urban United States counties?

3

RQ6 Do higher levels of broadband Internet predict higher levels of civic

technology use in urban United States counties?

RQ7 Do higher levels of economic distress predict lower levels of civic

technology use in urban United States counties?

These specific research questions investigate the characteristics of metropolitan areas in the United States that are adopting civic technology as well as those that are not. These research questions are significant because civic technology is a new form of association that seeks to contribute to the reinvigoration of civic life.

The diffusion of innovation literature assumes that over time new innovations are fully adopted, but in practice, the literature demonstrates that innovations are rarely, if ever, fully adopted. Recent work on civic technology and open data (McNutt et al., 2016), participatory budgeting (Gilman, 2016), and network non-profit arrangements

(McGuinness et al., 2018) demonstrate recent democratic innovations are still in the early adopters' stage of diffusion. Through seeking to understand the dynamics of the current diffusion process of Code for America Brigades this dissertation contributes to the practical work of spreading this innovative practice.

4

LITERATURE REVIEW

Civic technology is an emerging practice and research is beginning to appear. This literature review defines civic technology as a new type of civic participation in the life of communities. As such, the literature review will first review the literature on civic participation. Subsequently, it will argue that civic technology is based on productive activities such as writing software, producing data visualizations, and organizing events. Finally, it will explore the implications of civic technology for governance, participation, and social movements (Brady, Verba, & Schlozman, 1995; Stephens, 2018; Zukin, Keeter, Andolina, Jenkins, & Carpini, 2006; McNutt, 2018).

2.1 Civic Technology As Civic Participation

Civic technology includes a wide variety of technology-based collaborative activities at various levels of community and governance. Variations in civic technology activity include the actors involved and the process through which technologies are produced. At times, civic technology activities are conducted by private, for-profit organizations, while at other times they are conducted by communities of volunteers. There is also variation in the civic technology literature regarding the development of civic technologies and the activities of the civic technology community. For instance, Hou (2018), Graeff (2018a, 2018b), Hendler (2016), and Patel and Sotsky (2013) write about the technology itself as an intervention in public affairs. Hendler argues that civic technology is “any technology

5

that is used to empower citizens or help make government more accessible, efficient, and effective” ( 2016, p. 7). Others focus on the activities of civic technology. McNutt et. al. (2016, p. 154) argue that the dimensions of civic technology are open civic data and transparency, civic and service applications and advanced technology solutions, and organizational innovations such as hackathons, Code for America fellowships, and similar arrangements. Stephens explains civic technology as “the use of open data by the people outside government to create new software applications or presentations of the data for public benefit” (Stephens, 2017, p. 3). Finally, Living Cities defines civic technology as “the use of digital technologies and social media for service provision, civic engagement, and data analysis” (OpenPlans, 2012, p. 2). The literature makes clear that civic technology involves the development of information and communication technologies. As stated above, civic technologies are sometimes developed by private organizations, and at other times, civic technologies are developed by volunteer communities. What is also clear is that whether private organizations or community volunteers produce civic technologies, there is a deep concern in the civic technology community with the process by which technologies are produced. This dissertation argues that inclusivity in civic life is the main concern of the civic technology movement. The civic technology movement argues that the people who technology is meant to serve should be involved in the process of its production.

The dissertation focuses more heavily on the activities of the civic technology community as a form of civic participation. However, the lines between private development and community development are not clearly defined. A focus on companies and organizations involved in the development of civic technologies is

6

important to clearly explain the process by which civic technologies are made. The next section reviews the literature on civic participation. Subsequently, the argument for civic technology as civic participation will be made.

2.1.1 Definitions of Participation

Definitions of civic participation fall into either of the following two categories: power-centered approaches (Arnstein, 1969; Connor, 1988; Fung, 2006) to civic participation and evolutionary approaches to civic participation (Putnam, 2000;

Verba & Nie, 1991; Zukin, Keeter, Andolina, Jenkins, and Carpini, 2006). The main difference between these two categories is that the power-centered approach claims to understand how citizen power is exercised, while the evolutionary approach does not.

Rather, the evolutionary approach claims to understand the various ways in which citizens participate in civil society, who participates, and why certain socioeconomic groups participate at different rates. Early models of political participation, such as Sherry Arnstein’s “Ladder of

Citizen Participation” (1969) are unidimensional because they only capture a single mode of involvement in civil society. Arnstein’s (1969) “Ladder of Citizen Participation” focuses on urban environments, arguing that various arrangements of citizen involvement are located on a scale of empowerment. On the lower end of the scale, meetings are designed to manipulate or tokenize citizens who seek to exert influence over policy processes and decisions. On the upper end of the scale, meaningful citizen involvement cedes control over community decisions to citizen discretion (Arnstein, 1969). In the early 1970s, Verba and Nie’s (1991) study rejected unidimensional conceptions of participation. Rather than conceptualizing civic and political

7

participation solely as attending meetings, they argued that modes of participation in

American civil society vary and include: voting, election campaign activity, contacting public officials, and cooperative activity (Bowyer & Kahne, 2016; Verba & Nie, 1991). Verba, Schlozman, and Brady’s (1995) definition of participation includes activities and excludes unconventional politics such as protest marches or consumer advocacy (Bowyer & Kahne, 2016). Further, Verba, Schlozman, and Brady (1995) define political participation as actions designed to directly affect government behavior.

Zukin, Keeter, Andolina, Jenkins, and Carpini (2006) differ on this point by distinguishing between civic and political participation. They define civic participation as “organized voluntary activity focused on problem-solving and helping others”

(Zukin, Keeter, Andolina, Jenkins, & Carpini., 2006, p. 7). While the inclusion of activity outside of established political systems can lead to a conception of civic and political participation that encompasses nearly any behavior, only using participatory actions sanctioned by the formal political system excludes innovative action when tremendous amounts of such activity are occurring. Civic technology is representative of civic participation that is in line with Zukin Zukin, Keeter, Andolina, Jenkins, & Carpini's (2006) definition. The overarching focus of civic technology on developing information and communications technologies does not always represent a direct appeal to the government to change behavior. The civic technology community’s focus on building technologies for community benefit is oriented towards problem-solving and helping others.

8

2.1.2 Wither Participation?

The definition and measurement of participation are important because what counts as participation shapes whether studies determine if participation is increasing or decreasing. For instance, Robert Putnam’s Bowling Alone (2000) demonstrated the decline of the United States’ post-World War II era social capital. Participation and association in political, civic, religious, and work-life have all significantly eroded. Voting, writing letters, contacting elected officials, making speeches, and attending meetings have all declined significantly since the 1970s. Putnam (2000) argues that the reason participation in these activities is in decline is due to societal changes, such as suburbanization, television, family structure, and generational replacement. At the same time, Putnam explains that modes of participation are changing. Affiliation through mailing lists and donations to tertiary associations such as non-profits and political organizations increased. Participation in virtual communities, self-help groups, and non-profit organizations is growing. Following Putnam (2000), scholars continued to monitor the evolution of participation in American civic life (Norris, 2013; Zukin Zukin, Keeter, Andolina, Jenkins, & Carpini 2006). Zukin, Keeter, Andolina, Jenkins, and Carpini (2006) asked the question of whether Putnam’s (2000) social capital measure accounted for the civic activity of younger generations. They find that when civic and political engagement is broken down into segments, it is less clear that younger people are less involved than their older counterparts. Whereas Putnam divides engagement between cooperative activity and expressive activity, Zukin, Keeter, Andolina, Jenkins, and Carpini (2006) divide engagement by civic engagement, political engagement, public voice, and cognitive engagement. They find that older generations are more likely to engage in traditional politics than their younger counterparts, but that younger

9

generations tend to be more involved civically. Older generations tend to vote more frequently, are more politically knowledgeable, and report that it is important to follow current events. Younger cohorts, however, are more likely to volunteer with civic organizations.

Zukin and his colleagues contextualize participation by noting that the United States is characterized by a decentralized administrative state that, since the 1980s, has seen a move towards devolution (Barnekov et al., 1981; Sellers, 2002; Zukin, Keeter, Andolina, Jenkins, & Carpini, 2006). They argue that the shift toward privatization makes the belief of many younger Americans that business has more influence over their lives than government understandable. For this reason, Zukin and colleagues (Zukin, Keeter, Andolina, Jenkins, & Carpini, 2006) examine consumer activism as a dimension of participation. This study takes an evolutionary approach to civic participation. It argues that Putnam’s (2000) work documenting the decline of the United States’ post-World War

II era civil society is a snapshot in time. Rather than viewing Putnam’s work as documenting the decline of American civil society, the argument here is that Putnam’s (2000) work contributes to documenting a transition in American society. The transition of American society is away from formal political channels and towards mission-driven, non-profit organizations as well as virtual forms of participation. In this formulation, civic technology represents an extension of the emerging society Putnam documented.

2.2 Historical Antecedents of Civic Technology

As Zukin, Keeter, Andolina, Jenkins, and Carpini (2006) and Norris (2013) point out and, as the community technology and community networks movements

10

illustrate, civic participation started to change in the late twentieth century. As participation in post-World War II era civil society waned, new forms of community focused on consciousness, principles and values, and action arose (Schuler, 1996). Early examples of technology-focused communities included community networks and community technology movements. In the face of the decline of the idyllic (though marred by racism, homophobia, and gender inequality) 1950s American community, Schuler argues that a new form of community that embraces technology as a core component is forming out of necessity.

2.2.1 Community Networks

In the late 1980s, associations known as community networks appeared for the first time in the United States (Schuler, 1996). These associations were often organized around websites that stored and presented community information. Some of them, such as The Cleveland Free Net, mimicked the organization of a downtown district (see Figures 1 and 2). As shown in Figure 1, users identified themselves as members or visitors. Once inside the Cleveland Free-Net, the main menu appeared that listed services: Press 4 to see information regarding the courthouse, Press 11 to access the library, and so on. The Cleveland Free-Net was community-run and fiscally sponsored by Case Western Reserve University. Similar projects had varying organizational structures. Community Memory of Berkeley, California was run by Cal-Berkeley, while Public Electronic Network (PEN) of Santa Monica was sponsored by the city government (Schuler, 1996). Other projects tackled specific community problems, such as education or economic development. Liberty Net in Philadelphia included information from the Greater Philadelphia Chamber of Commerce,

11

University City Science Center, and the Philadelphia Unemployment Project (Schuler,

1996).

Figure 1. Cleveland Free-Net Home page. (E. Smith, 2018)

12

Figure 2. Cleveland Free-Net main menu. (E. Smith, 2018)

2.2.2 Community Technology

The community technology movement has its origins in the effort to overcome social inequalities, especially concerning knowledge and access to computing. For a time, the movement received government support in order to fulfill this mission. This stands in contrast to the community networks movement, which sought to provide public service websites that connected individuals to other community members and resources without the support of the government. These two movements demonstrate that some communities have the resources to forge new community endeavors while others require support due to historic inequalities (Maloney, Smith, & Stoker, 2000).

13

In 1968 the National Urban League established a mainframe computing training program in Los Angeles, California. The program was the start of the community technology movement (Davies et al., 2003). Since that time, the community technology movement has sought to close the digital divide by providing information technology training. The community technology movement carried out its work in stand-alone centers, partnered with public libraries or non-profits such as the YWCA, and turned public spaces such as schools, libraries, community centers, churches, and housing projects into local networks of community technology centers

(Davies et al., 2003). In 1992 a Harlem-based Community Technology Center (CTC), Playing2Win, was awarded a National Science Foundation grant to construct a regional network of community technology centers across the northeast. This grant was the starting point of a period of growth for the community technology movement. The CTC program funded nearly 400 CTCs and contributed to the expansion of more than 150 CTCs

(Borgstrom et al., 2005). A key to this growth was the Technology Opportunities Program (TOP), a grant program in the U.S. Department of Commerce, National Telecommunication and Information Administration.

From 1994-2004, TOP made 610 grants in all 50 states, Washington, D.C, the U.S. Virgin Islands, and Puerto Rico, granting approximately $233.5 million

in federal funds, matched by local communities that provided $313.7 million.

TOP projects cut across many different sectors including community and economic development; lifelong learning and the arts; health; public safety;

14

and social, governmental, and other public services (Borgstrom et al., 2005,

p. 104). In addition to TOP, in 1995 the Office of Multifamily Housing in the U.S. Department of Housing and Urban Development (HUD) launched the

Neighborhood Networks (NN) initiative, which encouraged the development of CTCs in HUD properties. By 2001, NN had established more than 800 centers in operation nationwide (Davies et al., 2003). These programs were discontinued in 2004. However, the momentum built by the community technology movement during the 1990s built infrastructure and taught valuable lessons about how to bridge the gap between community technology and community development.

2.3 Components of Civic Technology

Justice, McNutt, Melitski, Ahn, David, Siddiqui, and Ronquillo (2018) argue that the components of civic technology are open civic data, technology, and innovative practices (Justice et al., 2018, p. 91). Certainly, each component of civic technology constitutes a specific activity. Bringing the components together, as civic technology does, has larger implications for American governance that continue the evolutionary trend of privatization (Justice, McNutt, Melitski, Ahn, David, Siddiqui, and Ronquillo., 2018).

Retrenchment of federal funds from large grant programs and devolution of decision-making over administrative responsibility to state and local governments, and ultimately to non-profit organizations, has made the boundaries among the business, non-profit, and government sectors more porous (Justice McNutt, Melitski, Ahn,

15

David, Siddiqui, and Ronquillo, 2018). The argument advanced here is that civic technology has implications for the relationship among political advocacy organizations, non-profits, and governmental institutions (Justice et al., 2018). Below, the details of each component of civic technology will be explained and the implications of the convergence of these components will be discussed.

Open Civic Data

Civic Technology

Innovative Technology Practices

Figure 3. Components of Civic Technology. McNutt, J. G. (2018, p. 91).

16

2.3.1 Open Civic Data

Open civic data refers to sets of data made available by a government that provides a complete and primary accounting of a government program (Opengovdata, 2007; also see Lathrop & Ruma, 2010). This does not include data that is personally sensitive or secretive, nor does it necessarily include physical artifacts. The data must be timely, accessible, and machine-readable. Open data does not refer to archival data sets, but rather data sets of current government activity. Further, the data must be on the Internet and in up-to-date formats and protocols. It must be non-discriminatory, non-proprietary, and license-free. The data is free for anyone to download anonymously, no entity has exclusive control, nor do they attempt to control the use of the data in any way (Opengovdata, 2007).

Open data is one aspect of the move toward open government. A worldwide movement, open government seeks to encourage transparency and accountability (Lathrop & Ruma, 2010) Open civic data, a component of open data, is often described as the fuel of civic technology (Whitaker, 2015) because it plays an instrumental role in enabling civic technologists to analyze, visualize, and redesign government programs and services. Open civic data represents an outward occurrence of horizontal transparency where individuals and organizations can observe the behavior of other organizations (Heald, 2006). Open civic data entails the recognition that data is vital to solving public problems. The sources of data that can contribute to solving public problems can come from public, private, or non-profit organizations in order to provide valuable public services. Data can either be disclosed or undisclosed, personal or non-personal. The mechanism for accessing data can be a data cooperative, research partnerships,

17

intelligence products, Application Programming Interfaces (APIs), or trusted corporate data intermediaries (Verhulst, Young & Srinivasan, 2016.). Indeed, the principles of open civic data are ethical standards for the management and dissemination of government records. Part of what makes civic technology a movement is that groups and organizations provide tools to assist with the diffusion of these ethical standards (Civic Tech Field Guide, n.d.; Data Ethics Canvas, 2019). Crowdsourced documents such as the Civic Tech Field Guide or the Data Ethics Canvas, created by the Open Data Institute, provide case studies and workflows to enable governments of all sizes to institute open government data programs.

2.3.2 Technology

Hyper powerful, free computing software such as Slack, R, Python, Google Drive, Google Maps, Mapbox, or Hadoop, enables anyone with a computer and an Internet connection to access and analyze data in a way that used to be reserved for professionals. The civic technology movement is enabled by technology, and the movement critically examines technology’s societal role, in particular as technology pertains to government services (McNutt, 2018). As Patel and Sotsky (2013) explain, the civic technology sector is composed of products that enable community life and analyze community problems, including resident feedback, public decision-making, data access and transparency, neighborhood forums, visualization and mapping, information crowd-sourcing, data utility, voting, and civic crowdfunding as well as for-profit businesses such as Airbnb (Patel & Sotsky, 2013). Patel and Sotsky’s civic

18

technology umbrella captures crucial aspects of civic technology activities such as transparency, crowd-sourcing, and mapping. Patel and Sotsky’s understanding of civic technology also includes for-profit businesses who use technology as a central aspect of the value proposition provided by businesses, but whose value proposition is not necessarily driven by community interests. Patel and Sotsky’s (2013) report provides an entry-point for investigating civic technology as a socio-technical innovation. Their definition is broad to the point of encompassing socio-technical innovations in both commerce and governance.

Including Airbnb in the definition of civic technology places for-profit business applications that provide a private value to consumers into the same category as volunteer built applications that provide information to the public about a community issue or problem. In doing so, they construct a definition of civic technology that encompasses all useful Information and Communication Technologies, regardless of the process by which they are built or the purpose they serve. Civic technology activity can include for-profit work, but this is generally in relation to programs and services under the jurisdiction of public entities. An application such as Airbnb is an innovative business model that reimagines what a hotel or lodging company does in the context of new technology that allows individuals to convert their homes or apartments into rentable spaces. Civic technology has a similar spirit to an application such as Airbnb because it reimagines what governance is in the context of new technology (David, McNutt, & Justice 2018).

2.3.3 Collaborative Processes

Hou (2018) points out that civic technology is a broad term that encompasses a wide range of societal sectors such as, business, governance, community informatics

19

and Human Computer Interface (HCI) civic engagement. Hou's presentation of various kinds of civic technology (i.e. vendor civic technologies, research-driven civic technologies, off-the-shelf civic technologies, and civic hacking and hackathons) underscores that civic technology involves multiple sectors. Hou’s presentation conveys a helpful starting point for unpacking the process through which public service and feedback platforms use the Internet and communication-based technologies as a central aspect of their value proposition.

2.3.3.1 Civic Hacking and Hackathons

Typically, the hackathon is a two to three-day contest in which participants attempt to solve a problem with the use of computer coding (Briscoe, 2014; Johnson &

Robinson, 2014). Briscoe writes about hackathons as a general phenomenon in the arenas of music, design, fashion and open data. Briscoe, citing others, defines a hackathon as a "problem-focused computer programming event" (Briscoe, 2014, 1). Hackathons exert considerable cultural significance in technological innovation. Many

IT organizations host such events on an annual basis, invite the public to participate, and offer monetary rewards to those who win. Hackathon event coordinators often market events by focusing events on a specific topic. Briscoe draws a helpful distinction between tech-centric hackathons and focus-centric hackathons. Intuitively, tech-centric events focus on improving a single application, platform, or the use of a specific language or framework. Focus-centric events, on the other hand, are issue, demographic, or organizationally oriented. Briscoe reports on the findings of a survey of 150 hackathon participants from the United States which finds that participants are overwhelmingly male and between the ages of 25 and 34. The top cited reasons for attending a hackathon are learning and networking. Finally, Briscoe argues that the

20

survey result finding that 70% of hackathon attendees were not attending their first hackathon suggests that there is a hackathon circuit in which attendees are interested in more than the individual results of a single hackathon (Briscoe, 2014).

Figure 4. Open Data Delaware, Data Jam 2017. (Quinn, 2018).

Johnson and Robinson (2014) attempt to evaluate hackathons as citizen engagement and government procurement. The authors argue that there is a need to track the outputs of hackathons over time to see if meaningful work is being done. If meaningful work is not being done, government sponsors risk the possibility of citizen fatigue. The authors describe the first civic hackathon, Apps for Democracy, held in Washington, D.C. in 2008. Others have followed up on the success of Apps for

Democracy and found that of the 47 applications developed during Apps for Democracy, only one was used after the event was conducted (Howard, 2011). The argument made by civic hackathon advocates is that the benefits of these events are

21

not found in any one event, but rather in the community and the knowledge gained in conducting these events (Headd, 2011). At the same time, individuals provided the D.C. city government with numerous hours of labor and only several of them received compensation, underscoring Johnson and Robinson’s (2014) questions about whether hackathon participants feel they are being exploited or view their participation as impactful. The issue here is that the purpose of civic hackathons is ambiguous, and that ambiguity has the potential to create tension between those who organize civic hackathons and those who participate in them. If civic hackathons are viewed as events designed to serve a civic purpose and people participate because they want to serve that purpose, then the potential for tension is low. If, however, civic hackathons are designed to serve a civic purpose, but people participate because they want to win money, then there is the potential for tension burnout. Participating in a civic hackathon is a significant time commitment.

Volunteering in a civic hackathon is at least a weekend-long commitment in which participants are attempting to construct solutions to complex problems and build a technology that is part of that solution. If participants in civic hackathons are doing so because they want to volunteer their time and effort for their community, then hackathons are truly serving a civic purpose. If, however, participants are there because they want to win money and prizes, then there is the potential for single event participation. This would be problematic because the knowledge and resources proponents of civic hackathons claim that the events contribute to the community is most likely minimal.

22

Gregg provides a broad overview of the rise of hacking as a mainstream societal phenomenon. She describes various hackers as "white" hat and "black" hat but does not distinguish between those who release government data without authorization and those that use authorized open government data. Gregg offers a penetrating analysis of the social events and rhetoric that emanate from hackathons. She argues that participation is sold as a political good to distract from cuts to government services. Moreover, she argues that political rhetoric venerating governmental performance masks the inevitable downgrade in service that assuredly accompanies such cuts (Gregg, 2015). Civic hack nights are an outgrowth of civic hackathons. They are a way to "regularize" the activity that takes place in hackathons in a sustainable manner. Hack nights do so by allowing more people over time to work on the same project. Also, hack nights present the opportunity for non-technical members of the community to participate through generating ideas for projects and serving as “project managers”

(Whitaker, 2015).1

2.3.3.2 Code For America

Within the United States and around the world, Code for America is a primary change agent in the civic technology movement. Code for America runs a community fellows program, supports brigades, holds convening activities including an annual summit, and builds technological products that have a civic purpose. Code for

America is a 501 (c) 3 nonprofit organization. The international division is called Code for All.

1 Personal communication with Dr. John Stephens, April 6, 2020

23

Jennifer Pahlka, the Executive Director of Code for America, was an O’Reilly

Media employee before founding Code for America.2 In 2012, Pahlka gave an influential TED talk in which she presented Code for America’s vision (Pahlka, 2012). In that talk, Pahlka describes the Code for America Fellows Program. At that time,

Pahlka described the Code for America Fellows Program as embedding individuals with technology skills in governments of all sizes in the United States. These fellows work with government employees to build applications and show what is possible with technology. In Boston, a Code for America fellow created an adopt-a-fire-hydrant application. It enabled people to volunteer to dig out fire hydrants after it snowed in Boston. The application went viral and was adopted by cities around the United States (Pahlka, 2012).

The Integrated Benefits Initiative started as a project between the City of San Francisco’s public benefits agency and Code for America (CfA) (Code for America, 2019). San Francisco paid Code for America $255,000 for a fifteen-month project. At the time, the California food-stamp participation rate was 50% of eligible residents. More than three-fourths of Americans who earn less than $30,000 annually use the Internet. However, GetCal Fresh, which was maintained by Hewlett-Packard, did not have text or e-mail support and the website did not work well on mobile devices.

Residents only needed to provide their name, address, and signature to the appropriate California agency in order to apply for benefits. The limited data points were necessary for a resident to apply enabled Code for America to redesign the website in

2 Code for America has a mailing list of 30,000 (Code for America, 2018).

24

order to make that low barrier to application clear to prospective applicants (Code for

America, 2019; Solomon, 2015). The GetCalFresh team took an iterative approach to update GetCalFresh services (Solomon, 2015). They made a map of ATMs that could be used without a surcharge fee. They updated the homepage. They used simple, direct language, let people know the application would not take a long time, and they added an apply button. They shortened the application and added a text service to let people know when an important phone call was on its way. They changed the document types accepted for upload and eliminated an unnecessary document that previously applicants had to sign in order to submit their applications (Solomon, 2015).

Figure 5. GetCalFresh. (GetCalFresh.Com, 2020)

The partners involved in the Integrated Benefits Initiative include Code for America, Center on Budget and Policy Priorities, and NAVA Public Benefit Corporation. Working collaboratively, these organizations are conducting pilot programs in Colorado, Alaska, Vermont, and Louisiana. The initiative was started in

25

August 2018. The goals vary across states. The goals of Michigan and Vermont are to hasten access to safety net benefits by eliminating barriers. Alaska’s goals are to improve access, visibility, and benefit delivery in remote communities. In Colorado and Louisiana, the goal is cost reduction (Code for America, 2019).

2.3.3.3 Code For America Brigades

The scope of Code for America Brigade activities encompasses all three components of civic technology – open data, technology, and innovative practice.

Code for America Brigades run hackathons, analyze open civic data, use technology as a resource, and build technology as a product with a community benefit. Brigades arose out of excess demand for participation in Code for America’s Fellowship

Program. In this case, excess demand refers to much greater interest in the fellowship program than there were fellowship slots. Code for America Brigades are the volunteer arm of Code for America (Schrock, 2016). Brigades exist within local communities and are organized by individuals within those communities with the support of Code for America. Code for America has a staff dedicated to working with Brigades on “priority projects” involving various civic issues (Code for America, 2020). Brigades have Memoranda of

Understanding with Code for America regarding how they will operate. To provide an example, Turkel, Suchanic, and Neil (2019) describe their participation in the 2018 Open Data Challenge facilitated by Open Data Delaware.

The Open Data Challenge statewide hackathon sought to address the challenge of accessibility in Delaware, particularly as it relates to transportation and natural resources. They describe building a paratransit portal for the State of Delaware, which they called MobiliDE.

26

Figure 6. MobiliDE Launch Page

Due to the high demand and the large scale of Delaware’s universal paratransit service (Institute for Public Administration, University of Delaware, 2013; Scott & Tuttle, 2007; Turkel, 2015; Tuttle & Falcon, 2003), MobiliDE sought to reduce costs by streamlining the ride reservation process and providing ride alternatives. From the

Delaware Open Data Portal, data sets were used on bus routes, bus stops, and bus schedules. Additional information sources provided an inventory of third-party service providers for inclusion by county. The completed prototype presented at the pitch night showcased a working website that could be used to create a user profile, reserve or cancel rides, view ride history, and when reserving rides, would showcase alternative routes. Figure seven provides a visual sample of the desktop version of the

27

prototype, which was also available in a mobile format. MobiliDE was awarded an

Open Data Challenge Ideation Award and a $1,000 grant. This is just one example of the activity of a Code for America Brigade. Code for America Brigades are connected to a national network of Brigades that all conduct work in local communities with the volunteer labor of local community members and partnerships with local governments and businesses.

2.4 Civic Technology As a Social Movement

Civic technology is one example of a growing number of movements aimed at increasing democratic participation in governance (McNutt, 2018; Sirianni, 2009; Smith, 2009). Other social movements such as participatory budgeting, Tech4Good, open data, and evidence-based practice are all social movements that demand greater involvement of citizens in the everyday work of government decision-making and service provision (Gilman, 2016; Noveck, 2015). These movements span the gamut of rational, re-educative, and pressure politics (Bennis et al., 1985). Political advocacy organizations, non-profits, and governmental institutions all work to produce societal change, but use different strategies to do so. Logically, political advocacy organizations would use pressure politics, non-profits would use re-educative strategy, and the government would use rational strategies to produce change. Civic technology represents the re-educative strategy of teaching new ways of engaging in civic participation to produce change. This section explains the specific type of re-educative strategy advanced by the civic technology movement in relationship to civic participation. This section argues that the re-educative strategy of civic participation advanced by the civic technology movement further blurs the boundaries among civic participation, governmental institutions, and social movements.

28

Tim O’Reilly’s influential article, Government as a Platform, conveys a vision of government organized around the principles of open source software development that relies heavily on crowd-sourcing (O’Reilly, 2010). O’Reilly argues that the role of government should be to facilitate participatory democracy through open government data. The vision of government as a facilitator necessitates a reprioritization of governmental tasks towards data collection, assembly, and dissemination as well as encouraging public participation in the role of designing and implementing government services. O’Reilly’s vision of civic participation in one in which civic participation is administrative in nature. He is not advocating a public forum concept of civic participation. O’Reilly’s vision does not see citizens sitting in rooms with governmental officials providing input on public plans. Instead, his vision of civic participation is citizens crunching governmental data to understand the specific details of governmental action (O’Reilly, 2010). Indeed, O’Reilly’s vision of civic participation is limited in certain ways. In formulating an innovative conception of civic participation, his vision leaves out people's technical skills. To participate in O’Reilly’s concept of civic engagement one must have computer skills, knowledge of data analytics, and/or computer programming.

Gordon and Walter (2016) extend O’Reilly’s innovative notion of civic participation by imagining ways in which all citizens, regardless of skill sets, can participate in this new mode of civic participation. Indeed, Gordon and Walter (2016) are skeptical of a vision of civic participation that reduces citizenship to empowering citizens to assist the government to become more efficient. In fact, they argue that inefficiencies need to be built into civic systems to empower citizen action. Such

29

inefficiencies can take the form of feedback, discussion, or open-source programming

(Gordon & Walter, 2016). Gordon and Walter (2016) distinguish technological efficiency from civic efficiency. Whereas technological efficiency is defined by cost- effectiveness, speed, and market distribution, John Dewey's (Dewey, 2012) notion of civic efficiency is defined by the ability to get things done with others, even if the participatory process is disruptive, messy, and unpredictable (Gordon & Walter, 2016). The varied perspectives offered by O’Reilly (2010) and Gordon and Walter

(2016) show the range of ideas present within the civic technology movement. Some are more focused on transitioning government to a more open, data-driven model in which citizens are empowered to participate in technical and analytic work. Others are more focused on thinking about how to make the technical and analytic work of government empower citizens on inclusive solutions for the whole community. There is agreement, however, that the ideal model of civic participation needs to go beyond deliberative democracy (Noveck, 2010, 2015, 2016). Noveck (2010) argues:

Whereas diverse viewpoints might make for a more lively conversation,

diverse skills are essential to collaboration…Deliberation measures the

quality of democracy on the basis of procedural uniformity and equality of

inputs. Collaboration shifts the focus on the effectiveness of decision-making

and outputs. Deliberation requires an agenda for orderly discussion.

Collaboration requires breaking down a problem into component parts that

can be parceled out and assigned to members of the public and officials

(Noveck, 2010, p. 62).

30

Noveck argues that deliberative democracy reduces citizenship to self-expression rather than direct input into government responses to problems and issues and the exact ways the government responds to problems. Deliberation can contribute to governing if systems are established to translate citizen input from public forums into decision-making and government services, but many times they are not (Peixoto & Sifry, 2017). Schrock (2018), McCann ( 2015), and Noveck (2010) promote the idea of putting the development of technology in the service of community interests. In

Schrock’s work, there is an understanding that optimistic technological determinism often accompanies the marketing of new technology products. In Noveck’s CITE work there is the crucial understanding that public projects which aim to involve the public often do so through tokenism (Arnstein, 1969). Moreover, even deliberative forums with the authority to enact solutions do not go far enough in establishing true collaboration among government and citizens. In Noveck’s construction, collaborative work does not start with a set of alternatives presented by public officials. It starts by asking the public what problems they face in everyday life and proceeding from their feedback. All of these writers, from O’Reilly to Noveck to McCann promote a conception of in which civic participation is an everyday part of the government. As such, all of these concepts blur the line between civic participation and governmental institutions.

In these conceptions of civic participation, citizens are empowered to design governmental programs and engage in the administrative work of government. To be clear, this does not mean that most examples of civic technology replace government. Rather, for the most part, examples of civic technology replicate or run concurrently

31

alongside existing bureaucratic structures and governmental services (Justice et al.,

2018). These conceptions of government and civic participation are further complicated by social movements working to spread the innovative practices of civic technology.

Social movements are composed of campaigns and repertoires that cultivate public representations of worthiness, unity, numbers, and commitment (WUNC) (Tilly & Wood, 2013). Campaigns are persistent, structured efforts to make collective claims on formal authority, while repertoires are the activities of social movements (Tilly &

Wood, 2013). Displays of WUNC involve appealing to those outside the movement through appropriate dress and association with community members of status, displaying symbols associated with the movement, displaying voluminous support, and displaying sacrifice and steadfastness (Tilly & Wood, 2013). Hendler (2016) conceptualizes civic technology as a 21st century social movement. Hendler (2016) measures the movement in terms of scale and growth, grassroots activity, sustained engagement, shared vision, collective action, and shared identity. Scale refers to whether the movement has penetrated the mainstream; grassroots activity refers to whether there is community-driven activity outside of the most committed members; sustained engagement refers to whether work is sustained over time; shared vision refers to whether it is associated with a specific set of changes that is commonly understood by all members of the movement; collective action refers to whether civic technology is associated with people who take action rather than simply engage in discussion, and shared identity refers to whether people who take action, discuss, and identify with the movement. The indicators used to operationalize

32

these are interest, conversation, action, affiliation, and funding. Indicators are used interchangeably to capture the various measures. From 2013 to 2015, across all indicators, there was tremendous growth in civic technology as a movement. Based on analysis of news articles, civic technology is not propelled by mainstream news coverage, but rather is a grassroots movement (Hendler et al., 2016). The civic technology movement does get coverage from specialized news outlets but does not receive coverage from large media outlets on a regular basis (Hendler et al., 2016). There is sustained engagement and action, but a shared vision and identity are not yet present (Hendler et al., 2016). Civic technology creates interesting spaces relative to governmental institutions, social movements, and civic participation. The civic technology movement is creating civic participation opportunities that empower citizens to design public policies and to do administrative work (McGuinness & Slaughter, 2019). These civic participation opportunities generally do not replace governmental services or agencies but are much more likely to either work collaboratively with government or replicate governmental efforts (Justice et al., 2018). Generally, such civic participation opportunities are conducted publicly, while the channels through which they are developed are largely private. That is to say, Code for America and other non-profits working to spread civic technology is funded by private donors and are not fully accountable to the public.

2.5 Related Social Movements

As previously stated, the civic technology movement is one of a host of movements working to create opportunities for everyday citizen involvement in government (Noveck, 2015). These movements, such as Open Government,

33

Tech4Good, online protest movements, evidence-based practice, and participatory budgeting, share similarities and differences with the civic technology movement. The open government movement is more focused on transparency than the civic technology movement, while online protest movements use tactics that are more politically aggressive than the tactics used by the civic technology movement.

2.5.1 Open Government

The principles of open civic data were outlined above. Open civic data and open government data are part of related but distinct social movements. As previously articulated, open civic data is a component of civic technology. The civic technology movement is interested in inclusivity in the everyday work of government, specifically about building technologies. The open government data movement is primarily interested in government transparency (Open Data Institute, 2020).3 At the end of 2007, a group of individuals from around the world situated at the intersection of technology, politics, and civic engagement convened in Sebastopol,

California to write a set of principles for open government data (8 Principles of Open Government Data, 2007). Sponsorship for the event came from the Sunlight Foundation, Google, and Yahoo (Tauberer, 2007). The consensus of this meeting was that citizens have a right to government information and the inability of the government to produce transparent records is a signal of government failure. The group made this consensus actionable by establishing principles of open government data.

3 Personal communication with Dr. John Stephens, April 6, 2020

34

The public presentation of government data is an all-at-once type of transparency to make government data anonymous (Heald, 2006). Heald points out that transparency can also function as a type of surveillance. The open government data movement attempts to dampen this criticism by ensuring that government data released publicly does not contain information that would reveal an individual’s identity. Open government data asks the government to proactively produce public data that describes the implementation of government programs, services, and use of government resources. This provides the opportunity for business (Gurin, 2014) to capitalize on government information, enables residents to monitor their communities, and enables government agencies to share information either internally with other agencies or externally with other governments. As such, it represents both forms of transparency, vertical and horizontal, identified by Heald. Heald distinguishes vertical and horizontal transparency. Vertical transparency references the ability of superiors to view the behavior of subordinates and subordinates to view the behavior of superiors. Horizontal transparency refers to the ability of all subunits of an organization to observe the behavior of other subunits of an organization. Ideally, open government data allows for all of these types of transparency to occur simultaneously.

There are several ways in which the Freedom of Information Act (FOIA) and open government data differ. First, FOIA does not require agencies to conduct research, analyze data, or create records; while, in certain instances, open government data standards do require that of government agencies. Second, the information released through FOIA is not open by default. Rather information released through FOIA is only done so because an individual requested certain information. Third,

35

information released through FOIA does not have to be machine-readable. Fourth, information released through FOIA is only shared with the requester, rather than the public at large. Finally, FOIA grants citizens the right to request information from all United States federal agencies, which agencies are obligated to fulfill with government funding unless they can show that the request meets one of the nine exemptions (Scalia, 1982).4 In the United States, the principles of open government data were institutionalized during the Obama administration. Before the Obama administration,

Congress passed the Federal Funding Accountability and Transparency Act of 2006, which requires that all federal awards information be made available to the public on USASpending.gov (Pub. L., 109–282, 2006).

In December 2009, President Barack Obama signed the Open Government Directive, which created Data.gov and directed government agencies to publish information online using the principles of open data. Additionally, agencies were mandated to provide three data sets that met the principles of open data within forty- five days of the issuance of the executive order. Finally, agencies were mandated to establish a plan for implementing the requirements of the Open Government Directive and assign responsibility for that implementation to personnel within the agency (The

White House, 2009). In 2013, President Barack Obama signed an executive order titled, Making

Open and Machine Readable the New Default for Government Information. The order directed OMB to issue an Open Data Policy consistent with the directive of the Open Data Initiative. Further, it directed individual agencies to conduct a full analysis of

4 Those exemptions largely deal with matters of personal privacy and national security or interest.

36

privacy, confidentiality, and security risks before releasing information. The order created the Cross-Agency Priority (CAP) Goal, which tasked OMB with monitoring the implementation of the Open Data Policy across federal agencies (“Executive Order -- Making Open and Machine Readable the New Default for Government

Information,” 2013). Subsequently, in 2014, Congress passed The Digital Accountability and Transparency Act of 2014 (The DATA Act), which mandated the standardization of government financial data (Pub. L. 113-101, 2014). States emulated the actions of the Obama administration by adopting open data policies and open data portals. Two prominent open data evaluation frameworks used to communicate the nature of data are the Five Star Scheme for Linked Open Data and the Eight Open

Government Data Principles (Zuiderwijk & Janssen, 2013). A factor that contributes to the lack of consensus on evaluation criteria is the disparate rationales for open data initiatives: transparency, civic engagement, and economic value (Attard et al., 2015).

Furthermore, while international collaboration organizations exist, such as OpenData4Development, agency collaborations on open data are rare. The full chain of government response to a societal problem often involves multiple agencies, but open data initiatives rarely mirror these arrangements. While the reentry services of a particular government could involve a department of corrections, a department of health and human services, and a department of transportation, open government data programs generally are siloed agency by agency.

There is no consensus on which factors to use in the evaluation of open government data. There is, however, considerable attention paid to evaluating the nature of data (meaning the document types in which data are presented) and the

37

functionality of open data portals. Open data portals are publicly available, electronic databases that catalog government’s use of public resources and the performance of government programs (Attard et al., 2015). Many state and local governments have such programs. An example is the Delaware Open Data Portal shown in Figure seven.

Figure 7. Delaware Open Data Portal

The public can access Delaware’s open data portal to view data sets on a wide variety of topics and programs including state spending, education, and transportation.

2.5.2 Tech4Good

While the open government data movement uses technology as a means to argue for greater government transparency, the Tech4Good movement uses the

38

wisdom of crowds to develop technological products (Podder et al., 2018).

Tech4Good uses robotics and artificial intelligence primarily to assist citizens with accessibility needs. The Tech4Good movement uses a model similar to Code for America’s Brigade model. Present in countries around the world, Tech4Good is an international network of community groups driven by NTEN and NetSquared with community meetings occurring locally in communities around the globe.

2.5.3 Data4Good

Tech4Good runs programs that parallel those of Code for America. Code for America runs a community fellowship program that embeds technologists within governmental organizations. In addition, Code for America supports a volunteer network known as the Code for America Brigade Network. The Data4Good movement, spearheaded by organizations such as Data Kind, facilitates pro-bono work conducted by data scientists who work for large corporations on behalf of non-profit and humanitarian organizations to conduct data analysis and build data tools

(Data4Good, 2020). Besides, Data Kind hosts short, intense “datadives” where volunteers dig through datasets as well as longer term projects where those involved do not have to leave their day jobs to contribute (Data4Good, 2020).

2.5.4 Evidence Based Practice The evidence-based practice strategy attempts to rationalize local government processes through program assessment, budget development, implementation oversight, outcome monitoring, and targeted evaluation. Program assessment involves knowing what programs a government is operating, how effective those programs are in achieving outcomes, and how existing services compare to alternatives. Budget

39

development includes performance data, evaluation or audit findings, and national research. Examples of programs include Results First, Data-Driven Justice, and the Correctional Program Check List (Dube, 2018). Congress institutionalized aspects of the evidence-based practice strategy by passing the Foundations for Evidence-Based Policy Act of 2018. Originally titled the OPEN Government Data Act, the law extends policies passed by Congress in open data legislation mentioned above. This includes the formation of White House maintained Chief Data Officer Council as well as the Federal Data Service

(Chappellet-Lanier, 2019).

2.5.5 Online Protest Movemments

Online protest movements make sophisticated use of civic technologies by enabling activists to broadcast political messages in real time (Carter, 2017). While there is skepticism that the ease to which civic technologies is wholly beneficial to online protest movements (Tufekci, 2017), social media platforms and other tools create new tactics ( McNutt & Justice, 2016) for protest movements. Online protest movements differ from the civic technology activity addressed in this dissertation is that they use aggressive tactics in order to reach their goals.

Zuckerman defines civic media as "the use of participatory media technologies for civic participation, political engagement, or social change" (Zuckerman, 2016, 54). He argues that current social media campaigns such as #iftheygunnedmedown operate with a different theory of change than those used in the 1950s and 1960s. The theory of change used in the 1950s and 1960s sought justice within the existing legal system, whereas #iftheygunnedmedown seeks to change media portrayals of young African American men. While certain scholars argue that such activity constitutes slacktivism,

40

Zuckerman argues it should be considered civic media (Zuckerman, 2016). Indeed,

Zuckerman argues that campaigns such as #iftheygunnedmedown and #blacklivesmatter represent an expression of Michael Schudson's notion of monitorial citizenship (Schudson, 1998). Such a notion of citizenship envisions citizens scanning media outlets for events of concern and using their voices on social media to influence those events. In contrast, proponents of efforts to increase civic participation such as voting, attending meetings, and deliberation, argue that the key is to lower barriers for participating in such activities. Zuckerman argues that the problem with lowering the barriers to participation is that this assumes institutions are effective. The reason a citizen would choose to tweet their participation over other modes of participation is that they have given up on traditional modes of participation (Zuckerman, 2016).

2.5.6 Participatory budgeting

Gilman views participatory democracy as a form of popular rule that facilitates collaboration among government officials and the general public (Gilman, 2016).

Gilman argues that there is an overlap among technology related government reform movements and government reform movements intended to increase citizen influence over governmental decision-making processes. Gilman argues that participatory budgeting offers one such example. Participatory budgeting gives citizens direct control over certain aspects of governmental budgeting, such as local capital expenditures, i.e., park benches or street lights. The argument for such a process is that citizens have unique expertise over such issues in their local neighborhoods (Gilman, 2016).

41

2.6 Conclusion

Civic technology, specifically Code for America, represents an evolutionary step forward in the identities of non-profit organizations. Smith and Lipski argue that the transition to non-profits serving as government-sponsored service providers complicated the identity of non-profit institutions (Smith & Lipski, 1995). After becoming government-sponsored service providers non-profit organizations were both mission-driven community organizations and professional providers of publicly funded welfare services. Smith and Lipski (1995) argue that serving as government- sponsored service providers complicated the identity of non-profit institutions because it bonded non-profit organizations together as political associations nationally and within states. Code for America represents an organization shaped by this transformation, but also one that represents a current transformation of non-profit organizations. Code for America is a government contractor, a non-profit organization, a civic engagement organization, and a social movement.

2.7 The Consequences Of Diffusion The Ryan and Gross (1950) study of the adoption of hybrid seed corn investigated the antecedents of technological conservatism. This study raises the question of why one farming community in Iowa adopted hybrid seed corn before another community adopted it. The authors find that farms with more secular social interactions were generally faster to adopt the innovation of hybrid seed corn than those with more traditional interactions. Interestingly, neither the financial ability of farmers to purchase seeds nor lack of detailed knowledge about the new seeds was observed to be a significant barrier to adoption. The Ryan and Gross study provides evidence that technological conservatism was tied to social conservatism.

42

This dissertation is interested in whether the findings of the Ryan and Gross study apply in the case of a different innovation and in an updated context. That is, on a regional scale, this dissertation is interested in asking whether the demographic composition of place impacts the adoption of a sociotechnical innovation. Ryan and

Gross investigated the question of whether community differences in sociodemographic composition influenced innovation adoption in specific American communities during the 1950s. This dissertation investigates this question in urban American communities during the 2010s. The updated context is an important difference because governmental policy and market forces shape contemporary urban American communities. Governmental policy and market forces shape decisions about where

Americans choose to live. The most famous articulation of this idea is Charles Tiebout’s “Pure Theory of Local Expenditures” (1956). In that piece Tiebout conceptualized local services as a marketplace in which individuals choose where to reside based on the quality of public services. Tiebout’s thesis takes on new importance in an updated context in which economies are dominated by technology (Porter, 1990), the decentralized administrative state is failing to distribute the production and benefits of economic productivity equitably (Barnekov et al., 1981), and Americans are sorting themselves geographically by values and political beliefs (Bishop & Cushing, 2008).

This dissertation contends that the reshuffling of American communities is correlated with the emergence of new forms of civic association. The outcomes of federal policies that expose urban America to the forces of the global marketplace not only have economic implications, they also affect patterns of civic life (Chambers,

43

2010). The overarching reason for this is that as new technologies are more integrated into economic life, individuals who work in those technological sectors begin to creatively adapt those technologies to areas outside of the narrow tasks which they are paid to conduct (Rogers, 2003).

2.7.1 The Diffusion Context

From the 1950s to the 1970s there was an even distribution of partisanship, education, and income across American cities. That changed starting in the 1980s.

Education levels increased in cities across the South and Pacific Northwest and declined in the Northeast and Midwest. While some cities lost population in the 1970s, in the 1990s urban population increased, but this change was no longer based on an expectation of economic well-being, but rather on lifestyle preferences. Cities offered the opportunity to network, keep loose affiliations, and generate new ideas (Bishop & Cushing, 2008). There have been major shifts in the location of American manufacturing centers to the American South, away from major urban centers, and, more recently, abroad (Bluestone & Harrison, 1982). These shifts have been accompanied by a reversal of earlier population migration flows. During the 1930s and 1940s, there was a shift from South to North and from the countryside to cities. During the 1960s and 1970s, there was a decreasing need for labor in manufacturing and blue-collar occupations. Employment in white-collar occupations increased as a result of new service activities (Noyelle, 1983). These changes in the geography of production and the use of labor have occurred largely because major productivity gains have been achieved in the manufacturing sector. This is largely a result of changes in transportation, the introduction of new technologies, and the growth of a whole range

44

of new services and service-like activities carried out within manufacturing corporations. These activities aimed at enhancing the precision and efficiency of industry and the ability to manage, plan and develop manufacturing resources (Castells, 1996; Noyelle, 1983; Thompson & Thompson, 1987).

The transformation from an industrial to a service-based economy, particularly services to corporations, accompanied a major shift in American urban policy. A “new privatism” emerged with the election of Ronald Reagan as President of the United States (Barnekov et al., 1981). What distinguished this form of privatism is that while

Warner’s privatism (1968) focused on the city as a wealth-creating mechanism for individuals, now the city is a mechanism for creating national wealth. Federal dollars to urban areas for a host of social programs were rolled back in the 1980s under the

Reagan era retrenchment as a way to foster economic growth and competition between regions and municipalities. The Reagan and Bush administrations slashed federal programs for local

governments. Reagan eliminated general revenue sharing, a $1.8 billion cut to the budgets of larger cities. He cut funding for public service jobs and job training by 69% and the CDBG [Community Development Block Grant] program by 54% and reduced the social services block grant and funds for

urban mass transit by 37% and 25%, respectively. The UDAG [Urban Development Action Grant] program was cut by 41%. Overall, federal

assistance to local governments fell 60%, from $43 billion to $17 billion. The

only program that survived cuts was federal aid for highways, which primarily benefited suburbs, not cities (Dreier et al., 2014).

45

The retrenchment of the federal government during a time of great economic transition drastically shaped urban migration patterns. Berry and Glaeser (2005) present evidence of an agglomeration effect regarding education in metropolitan areas in the United States in the second half of the twentieth century. Metropolitan areas with higher starting proportions of individuals with advanced education degrees saw nearly exponential growth rates in their educated populations. Furthermore, Berry and Glaeser (2005) find that this education agglomeration effect spills over to wages and housing prices. Metropolitan areas with exponentially increasing populations of individuals with advanced education degrees saw similar effects to their jurisdictions’ wages and housing prices (Berry & Glaeser, 2005). Between 1976 and 2004, only 33% of counties in the United States became more politically competitive in presidential elections. In 2004, over 60% of U.S. counties produced landslide elections. It is not simply that churchgoers are more likely to vote Republican than non-churchgoers; it is that conservative churchgoers live in certain places and liberal churchgoers live in other places; conservative union members live in certain places and liberal union members live in other places. While it is true that Democratic counties are more educated, earn more money, and are more racially diverse, it is also true that public opinion on political issues is sorted geographically so that individuals who live in a Republican county are more likely to take the Republican position on an issue regardless of demographics (Bishop &

Cushing, 2008).

46

THEORETICAL FRAMEWORK

The central research question of this dissertation is which communities are adopting civic technology and how do they differ from communities that are not adopting civic technology? In asking the question, the dissertation is specifically concerned with whether regions are adopting Code for America Brigades. Code for America Brigades were chosen as a proxy for civic technology activity is that Code for America is one of the primary agents working to diffuse civic technology in the United States. While other organizations such as DataKind, Tech4Good, and the Open Data Institute also work to spread components of the civic technology movement, Code for America is a highly visible organization. The work of the civic technology movement is important to study because it exemplifies how the “network society” is transforming civil society (Castells, 1996). The “network society” is Castells’ term for sociocultural phenomenon centered around information and communications technologies. Rather than describing contemporary society as organized around information, Castells (1996) argues that society is increasingly organized around technologies that enable individuals to share information with one another and/or quickly process and analyze data.

The civic technology movement is indicative of network society phenomena because its work is organized around the creation of information and communications technologies, data analysis, and the use of information and communications technology platforms. Code for America Brigades are a useful proxy for civic

47

technology because they are centered in local communities (Morris, 1981). Having groups centered in local communities is desirable because it allows for comparing the characteristic of communities that have adopted Code for America Brigades with those that have not adopted Code for America Brigades.

This dissertation generates hypotheses regarding the distribution of Code for America Brigades based on diffusion of innovation, social capital, and creative class bodies of literatures. Diffusion of innovation is a framework for studying how ideas, practices, and technologies spread through society. The diffusion of innovation framework contains subunits of study, including the attributes of innovations, who adopts innovations, the communication channels through which people learn about innovations, and the individuals and organizations who serve as change agents by advocating for the adoption of particular innovations (Rogers, 2003). Each one of these subunits could be used to study civic technology and Code for America. Studies could pose questions about the attributes of civic technology as an innovation (Turkel,

2019), about the motives of hackathon participants (Stepasiuk, 2014), or about the communication channels through which civic technologists communicate (Hendler et al., 2016). The core question, however, of the diffusion of innovation framework concerns the spread of innovations over time.

The variables of social capital and creative class represent known trends about polarization in the United States (Florida, 2012; Rupasingha et al., 2006). The work of

Rupasingha, Goetz, and Freshwater (2006) provides evidence that the institutions of traditional social capital measures – churches, bowling alleys, labor unions, voting, and filling out the census – are associated more with rural areas rather than urban areas. In an opposite trend, Florida’s work on the creative class shows that the urban

48

areas with higher concentrations of “creative” occupations are associated with higher levels of prosperity. The overarching contention of this dissertation is that new economic forms are giving rise to new forms of civic association. Based on the body of knowledge generated by diffusion of innovation scholarship, relatively prosperous places are generally early adopters of new innovations (Rogers, 2003). For the purposes of this study, traditional social capital is characteristic of industrial forms of association and economic distress. Creative class, on the other hand, represents prosperity. Traditional social capital is a proxy for economically distressed places. Creative Class is a proxy for more economically prosperous places. Civic technology is an example of networked social capital. Studies on social capital have found that traditional social capital is highly concentrated in rural communities (Rupasingha et al., 2006). As previously stated, this dissertation argues that measures of social capital which only include traditional forms of civic association fail to capture the full range of civic association activities. This dissertation hypothesizes that civic technology is a part of a new wave of civic participation emerging in urban areas referred to here as networked social capital. In this type of social capital, a central non-profit coordinates volunteer groups across the country.

Just as Code for America provides support for Brigades, other non-profits similarly provide support for networks of volunteer groups working in local communities on other issues from data analytics to infrastructure to climate change. This dissertation hypothesizes that these groups will be collocated with Code for America Brigades.

49

3.1 Propositions

This dissertation tests the findings of the diffusion of innovation literature in relation to voluntary associations. Rather than comparing secular communities and conservative communities (Ryan & Gross, 1950), this dissertation compares communities based on varied levels of traditional social capital and creative class.

Given that the unit of analysis of this study is county-level, DiMaggio and Powell’s institutional isomorphism concept (DiMaggio & Powell, 1983) is not fully applicable because institutions are not the subject of this research. Rather, regions are under study. However, that does not mean that pressure is not playing a role in the spread of Code for America Brigades at the regional level. Similarly, Granovetter’s (1973) argument that individuals benefit from loose or weak social ties with others is not fully applicable since individuals are not the subject of this study. However, that does not mean that loose affiliations among individuals do not play a role is the spread of Code for America Brigades at the regional level. These theories need to be adapted to the current research study. This is done below.

In Making Democracy Work, Putnam defines social capital as “the features of social organization such as trust, norms, and networks that can improve the efficiency of society by facilitating coordinated action” (Putnam, 1993, p. 167). Putnam (1993) finds that Italian regions with higher rates of civic participation correlated with higher functioning governments in terms of passed legislation and responsiveness. Based on this finding, Putnam concludes that associating with fellow residents produces greater stocks of social capital (Putnam, 1993). This includes pressure from others to conform to the behaviors of the group. Interpreting Putnam’s work, Norris (2013) argues that social capital does not belong to individuals, but rather exists between and among individuals. She argues, therefore, that social capital cannot be measured at the level

50

of individuals, but rather must be measured at the collective level (Norris, 2013). This dissertation follows DiMaggio and Powell (DiMaggio & Powell, 1983), Putnam, and Norris in hypothesizing that pressure to conform to group norms plays a significant role in the spread of Code for America Brigades. The details of the role pressure and networks play in the diffusion of Code for America Brigades is detailed below.

3.1.1 Traditional Social Capital

Putnam’s work on the decline of social capital (Putnam, 2000) is alarming because of its findings regarding the decline of the rich tradition of voluntary association in American civic life. In his most recent work, Our Kids, Putnam (2015) argues that whereas inequalities in gender and race were major barriers to opportunity in mid-20th century America, now the greatest threat to opportunity is inequality among communities. This dissertation tests ideas about the relationship between traditional forms of association and contemporary forms of association. One of the hypotheses of this dissertation is that places with relatively higher activity in the form of traditional social capital measures will have relatively lower amounts of civic technology activity. This would be the case because established patterns of association prevent experimentation with new forms of association (Morris, 1981). Places with relatively less traditional social capital will have relatively higher amounts of civic technology activity because loose patterns of association enable experimentation with new forms of association (Ryan & Gross, 1950).

51

3.1.2 Creative Class

This dissertation tests why certain communities are adopting the innovative practices referred to as civic technology while others have yet to do so. The dissertation hypothesizes that places with relatively higher amounts of creative class workers (Florida, 2012) will have relatively higher amounts of civic technology activity. The reasons for this, however, are not due to the innate characteristics of members of the creative class. The creative class is hypothesized to have an association with civic technology because as a group the creative class shares a set of characteristics conducive to mobilization (Olson, 1971; Salisbury, 1969; Salmon et al., 2010). First, they have the skill set for involvement (Brady et al., 1995; Schlozman et al., 2018; Verba & Nie, 1991). Second, they have a workplace structure conducive to volunteer efforts (Shirky, 2010). Third, they are recruited by non-profit organizations (Schlozman et al., 2018). Fourth, it is in their self-interest because participation in civic technology offers professional development and career advancement opportunities for individuals who participate (Salisbury, 1969; Olson, 1971).

3.2 Central Hypotheses Places with lower amounts of traditional social capital are expected to have a lower amount of civic technology due to their pre-existing social structures (Morris,

1981). Pre-existing social structures such as churches, bowling alleys, labor unions, and political parties engage individuals in civic activities that take time away from individuals participating in other activities. Affiliation with specific organizations leads individuals to devote effort and resources towards those organizations which deepens the connection those individuals have with those organizations. Individuals form bonds with organizations by participating in civic activities. They exert physical

52

effort, mental effort, and provide financial resources to those social institutions. They also reap the benefits of affiliation. These mutually beneficial arrangements detract from incentives to innovate in regard to social institutions. Similarly, places with higher amounts of the creative class are expected to have higher amounts of civic technology due to human capital resources. This refers back to pre-existing social structures (Morris, 1981), but of a different sort than those related to traditional social capital networks. Whereas the pre-existing social structures of traditional social capital networks are centered on civic activities, the pre-existing social structures of the creative class are centered on the workplace (Saxenian, 1996). The creative class is generally composed of young professionals (Florida, 2012) who are not as bonded to existing social institutions because they are just starting their careers, are less likely to have families, and are more likely to closely identify with their career choice (Florida, 2012). One hypothesis of this study is that civic technology activity is an outgrowth of the creative class work environment.

Figure 8. Hypothesis matrix, adapted from McNutt and Justice, 2016

53

The hypothesis matrix (Figure 9) illustrates the anticipated interaction of social capital and creative class. Traditional social capital networks are hypothesized to not overlap with civic technology networks. Given this assumption, places with high levels of both civic technology and traditional social capital will have high amounts of civic technology, while places with low traditional social capital and high creative class will have high civic technology. The reverse is true when traditional social capital is high and the creative class is low. As stated above, the assumption is based on an understanding of how social networks operate. Social networks develop through individuals participating in the associational activity. Individual participation in associational activity tends to lead to further activity within that association. In doing so, individuals form bonds with others who participate and often take on work that builds upon and strengthens and the existing social network.

3.3 Covariants

The primary hypothesis of this dissertation concerns the emergence of civic technology activity in the context of economic divergence. That being the case, the study is primarily focused on analyzing the distribution of civic technology in relationship to traditional social capital and creative class. That said, there are certainly other variables that could account for economic prosperity, namely population and education.

3.3.1 Distress Score

By including distress score, the study is controlling for the possibility that the overarching idea of this study – that new forms of civic association follow new forms of economic arrangements – is incorrect. That is to say, this study hypothesizes that

54

creative class is associated with Code for America Brigades. In this case, the creative class is a proxy variable for prosperity. The reason the creative class is a proxy for prosperity is that the literature on creative class is based on the finding that cities and regions with high levels of creative class workers have higher levels of income and wealth compared to those cities and regions with comparatively lower levels of creative class workers.

3.3.2 Networked Social Capital

In a recent report by the New American Foundation, a host of arrangements similar to Code for America Brigades were identified (McGuinness et al., 2018). Across the country, individuals in urban places are joining networks with a specific advocacy purpose, including resiliency to climate change, bikability, and criminal justice reform, among other issues. Many of these organizations encourage volunteerism and provide citizens with opportunities to engage in community events. In short, they represent a similar type of action as civic technology presented in the literature review. This dissertation hypothesizes that there is the collocation of these types of activities. That is, places with significant levels of civic technology activity will also have the presence of other networked social capital arrangements

(McGuinness et al., 2018).

3.3.3 Broadband Technology in contemporary society is ubiquitous. Since 2008, more people around the world are connected to broadband rather than dial-up Internet, and there are more things connected to the Internet than people (Townsend, 2013). As of 2014, 84% of U.S. households owned computers and 70% of households had access to a

55

broadband Internet connection (Rainie & Cohn, 2014). As of 2015, 68% of Americans owned smart phones. This is up from 35% in 2011 (Anderson, 2015). All of these suggest the ubiquity of Internet technology in contemporary society, but the digital divide still persists. Given that the digital divide is present among individuals and communities, this study hypothesizes that places with higher levels of broadband Internet access will have higher levels of civic technology activity.

3.3.4 Population

Previous studies have found the population size to be associated with early innovation adoption (Damanpour & Schneider, 2009; Kearns, 1992). Given these earlier findings, counties with relatively larger populations are expected to be positively associated with civic technology in comparison to places with relatively smaller populations.

3.3.5 Education

Similarly, Berry and Glaeser (2005) challenged Florida’s finding that the engine of economic growth in contemporary urban cities is creativity. In contrast, Berry and Glaeser (2005) argue that human capital is the engine of economic growth. In order to provide evidence for their claim, they run correlations between the concentration of PhDs in a given jurisdiction and its economic growth. The purpose of this dissertation is not to settle arguments over the cause of economic growth. Further, since there is scholarly evidence for both claims, this dissertation hypothesizes that the concentration of PhDs will be positively correlated with the emergence of civic technology.

56

The specific hypotheses for this study are as follows:

H1 Urban counties with higher levels of traditional social capital will have lower

levels of civic technology

H2 Urban counties with higher levels of creative class employment will have

higher levels of civic technology

H3 Urban counties with higher levels of advanced education will have higher

levels of civic technology

H4 Urban counties with larger populations will have higher levels of civic

technology

H5 Urban counties with higher levels of networked non-profit arrangements will

have higher levels of civic technology

H6 Urban counties with higher levels of distress scores will have lower levels of

civic technology

H7 Urban counties with higher levels of broadband speed will have higher levels

of civic technology

The assumptions of how social networks operate extend to the hypotheses regarding the covariates. Previous research has found traditional social capital to be associated with rural areas, while the creative class has been found to be associated with prosperous regions. Given these findings, the hypothesis that social capital will not be correlated with civic technology extends to distressed economic areas. On the other hand, given previous research, the hypothesis that creative class will be

57

associated with civic technology extends to advanced education, population, networked non-profit organizations, and broadband Internet.

58

METHODOLOGY

This dissertation investigates the diffusion of civic technology activity in United States counties. Specifically, it studies the diffusion of Code for America Brigades in urban U.S. counties. It tests the explanatory power of independent variables – social capital, creative class, broadband Internet, networked social capital, number of PhDs, population size, and distressed community index – on predicting the presence of Code for America Brigades. The dissertation uses logistic regression in order to conduct the analysis.

4.1 Subjects Counties differ in a host of ways. They vary in size, economy, education, and environment. Counties such as Los Angeles County, California, Cook County, Illinois, Harris County, Texas, and Maricopa County, Arizona are very large in size. Other counties such as Arthur County, Nebraska, Kennedy County, Texas, King County,

Texas, Kalawao County, Hawaii, and Loving County, Texas are extremely small. Some counties are economically dependent on farming, others on mining, others on manufacturing, some on government contracting and military installations, and some on tourism. Some counties are growing, while others are declining in population.

The population for this study is 1,165 urban counties and county equivalents in the United States. Rural counties are not included because previous work by McNutt and Justice (2016) demonstrated a lack of hackathons present in these counties. The

59

counties included in this study are those counties that form the building blocks of the

Office of Management and Budget’s 174 combined statistical areas (CBSA). The Office of Management and Budget delineates a CBSA as, "a geographic entity associated with at least one core of 10,000 or more population, plus adjacent territory that has a high degree of social and economic integration with the core as measured by commuting ties” (“2010 Standards for Delineating Metropolitan and Micropolitan Statistical Areas,” 37249). First, a CBSA must have an urban cluster of 10,000 people or an urbanized area of 50,000 people or more. In other words, it must have either a micropolitan statistical area or a metropolitan statistical area anchoring the CBSA. Second, a CBSA must have a central county and an outlying county. A central county has 50% of the population living in an urban area of 50,000 or more and has an urban area of 10,000 or more located within its boundary of which 5,000 reside in the county. The CBSA is formed by employment interchanges with an outlying county. There must be 25% of workers commuting from the outlying county to the central county, or vice versa. The 2010 delineation also states that CBSAs can be combined if there is an employment interchange of 15%. CBSAs are named for the two or three largest principle cities within their boundaries.

60

Figure 9. Census Data Hierarchy. (Walker, 2017)

The following details the procedure used for constructing a data set that includes these counties. An initial data set including all counties in the United States was imported into R. The original dataset for U.S. counties was obtained from the Missouri Census Data Center. This source was used because the Missouri Census Data Center has a number of cross walk data lists that allow users to request multiple geographic identifiers for particular geographic entities. This was important for the current study because it provided a tool for understanding the overlap between counties and CBSAs.

Data and analysis were conducted on all subjects in the population. Since the study was conducted on all subjects in the population, a sampling distribution was not drawn. Since no sampling distribution was drawn, the study does not have a sampling

61

bias. It should be noted, however, that the data sources do rely on sampling strategies to construct individual county-level data.

4.2 Design Structure

This is an explanatory study using a cross-sectional design. The purpose of this study was to test the overall relationship between the presence of civic technology and various community attributes. The study analyzed secondary and administrative data. Logistic regression was used to analyze the data.

4.2.1 Variables and Data Sources The purpose of this study was to test whether specific variables are predictors of the presence of civic technology activity. The dependent variable for this study was the presence of a Code for America Brigade in a given Combined Statistical Area. Using data available on Code for America’s website, a list of counties in which Code for America reports a Brigade existing was created. From there, based on McNutt and

Justice (2016), this study was interested in at least four categories of independent variables – social capital, creative class, broadband Internet, and networked social capital.

4.2.2 Dependent Variable Code For America Brigades Code for America Brigades represent an accurate proxy for civic technology activity because they use open data, organize hackathons, and build civic technologies.

Moreno (2018) documents Code for Tulsa’s work building a civic technology called Courtbot (Moreno, 2018). Courtbot is a text message alert system to remind individuals about court dates. The app was first developed through scraping data from a local community legal aid organization (Moreno, 2018). After the application was

62

first developed in Tulsa, Code for America Brigades in Vermont, Alaska, and elsewhere developed similar applications (Mutrux, 2018). This example illustrates the practices of open civic data, technology, and innovative practices. Since the first Code for America Brigade was started in 2009, the presence of

Brigades has multiplied with roughly seventy groups now in existence across the United States and around the world, including Code for Japan. There are currently eight-five brigades worldwide (The Code for America Brigade Network, 2020). These groups represent a variety of communities, generally in large cities; however, there are also Brigades located in smaller, more rural communities. Brigades vary in their formality, size, and level of activity. Some have sophisticated organizational structures, numerous members, and events. Others are fledgling groups with minimal community presence.

Figure 10. Code for America Brigades as of August, 2018

63

4.2.3 Independent Variables

The independent variables used in the study are intended to the test the hypotheses outlined in chapter three. These variables include traditional social capital, creative class, networked social capital, population, broadband Internet, PhDs, and distress community index. Descriptions of independent variable sources and variable construction are included below.

4.2.4 Traditional Social Capital Rupasingha, Goetz, and Freshwater (2006) cite studies that use surveys to ask individuals about their levels of trust of others in their communities. Glaeser (2000) argues that survey measurements of social capital are not valid because people cannot accurately report their level of trust in others. Given the complexity of the concept of social capital, this study operationalizes social capital based on Rupasingha, Goetz, and Freshwater’s (2006) operationalization and data. The main variable Rupasingha, Goetz, and Freshwater generate is a social capital index derived from a principal component analysis of the following variables.

64

The analysis used four factors. The first factor included religious organization, civic and social associations, business associations, political organizations, professional organizations, labor organization, bowling center, fitness and recreational sports centers, golf courses and country clubs, sports teams, and clubs. The second factor was voter turnout, the first factor was the Census response rate, and the fourth factor was the number of non-profit associations. Each factor was measured for in a given county divided by population per 10,000.5

social capitla 2014

-3.182 - -0.867

-0.865 - -0.442

-0.441 - 0.004

0.005 - 0.747

0.751 - 21.831

Figure 11. Social Capital by County (Rupasingha et al., 2006)

5 Rupasingha, Goetz, and Freshwater (2006) aggregates data collected at the county-level from various sources, including County Business Patterns, US Census, Population and Housing Unit Estimates, Dave Leip's Atlas of U.S. Presidential Elections, US Census 2010, and National Center for Charitable Statistics. The associational measure is an aggregate of the following organizations: civic organizations bowling centers, golf clubs, fitness centers, sports organizations, religious organizations, political organizations, labor organizations, business organizations, and professional organizations.

65

Table 1. Social Capital Original Variables Variable Description religious2014 Number of establishments in Religious organizations (NAICS 813110) civic2014 Number of establishments in Civic and social associations (NAICS 813410) business2014 Number of establishments in Business associations (NAICS 813910) political2014 Number of establishments in Political organizations (NAICS 813940) professional2014 Number of establishments in Professional organizations (NAICS 813920) labor2014 Number of establishments in Labor organization (NAICS 813930) bowling2014 Number of establishments in Bowling center (NAICS 713950) recreational2014 Number of establishments in Fitness and Recreational Sports Centers (NAICS 713940) golf2014 Number of establishments in Golf Courses and Country Clubs (NAICS 713910) sports2014 Number of establishments in Sports Teams and Clubs (NAICS 711211) pop2014 Population assn2014 The aggregate for all of above variables divided by population per 1,000 (1st factor) pvote2012 Voter turnout (2nd factor) respn2010 Census response rate (3rd factor) nccs2014 Number of non-profit organizations without including those with an international approach (4th factor) sk2014 Social capital index created using principal component analysis using the above four factors (nccs09 is divided by population per 10,000). The four factors are standardized to have a mean of zero and a standard deviation of one, and the first principal component is considered as the index of social capital.

66

4.2.5 Creative Class

To measure the presence of creative class in U.S. counties, data from the

United States Department of Agriculture, Economic Research Service were used. This data set recasts Florida’s original work by excluding occupations with a high creativity level if these occupations are proportional to the residential population they serve

(Mcgranahan & Wojan, 2007). Occupations such as health care practitioners and the secondary school teacher are not included because these occupations generally serve residential populations. College professor, on the other hand, tends to serve non- residential populations and is therefore included in the data set. A full listing of occupations Mcgranahan and Wojan (2007) use in their construction of their creative class index is included as Appendix B. While Mcgranahan and Wojan (2007) recast

Florida’s Creative Class (2012) measure and challenge the validity of his measure, they also agree with Florida’s (2012) creativity hypothesis over Gleaser’s (2005) human capital hypothesis. It is important to note that this measure is based on pooled American

Community Survey (ACS) data from 2007 to 2011. This is the case because the long form Census is sent to one in six American households while the ACS is sent to one in forty. The limited sample used by the ACS allows for conducting the survey on an annual basis, but limits reliability for smaller counties. The USDA data confronts this challenge by pooling the data into a five-year estimate.

67

Figure 12. Creative Class by County, USDA

68

Table 2. Creative Class Original Variables

Variable Full Variable Name TotEmpEst Total Employment Estimate TotEmpSE Total Employment Standard Error CCEst Creative Class Estimate CCSE Creative Class Standard Error CCShare Creative Class Share CCShareSE Creative Class Share Standard Error BohEST Bohemian Estimate BohSE Bohemian Standard Error BohShare Bohemian Share BohShareSE Bohemian Share Standard Error TotEmpRME Total Employment Root Means Error CCRME Creative Class Root Means Error BohRME Bohemian Estimate Root Means Error CCShareRME Creative Class Shared Roots Means Error BohShareRME Bohemian Estimate Shared Roots Means Error

4.2.6 Covariates

The data sources and variable construction for the covariates is discussed below. Those covariates include education, population, distressed community index, networked social capital organizations, and broadband Internet.

4.2.6.1 Education PhDs awarded was collected from the National Science Foundation’s (NSF) Webcaspar website (ncsesdata.nsf.gov/webcaspar/). Ph.D. data was downloaded from the NSF Survey of Earned Doctorates/Doctorate Records File. The survey is an annual census of institutions that spent at least $150,000 in separately accounted for research and development in each respective fiscal year.

4.2.6.2 Population The data on population came from the United States Census data set on urban and rural counties and county-equivalents. Based on the raw population counts of

69

county and county-equivalents, an ordinal variable was established using means- centering. A similar procedure was performed on most of the covariant measures.

4.2.6.3 Distressed Community Index

The data used in the study comes from the Economic Innovation Group (From

Great Recession to Great Reshuffling: Charting a Decade of Change Across American Communities, 2018). Data was obtained through an e-mail contact and is not fully included in order to comply with a data agreement.

4.2.6.4 Networked Social Capital Organizations The data used in the study were collected by the New America Foundation. This variable measures the strength of networked forms of social capital in particular counties.

Table 3. Networked Non-profit Organizations Initiative Description #GoOpen Works to assist educators adopt open source software in their daily classroom activities.

100 Resilient Cities (100RC) Resilience initiative

Alliance for Innovation Accelerate adoption of best practices by local governments

Benchmark Cities Works with police departments to set reasonable goals.

Big Jump Project Works with cities to increase bike ridership by redesigning physical infrastructure

Built for Zero Develops real-time data on homelessness

C40 Uses data to improve cities responses to climate change

Cities of Service Focused on service

70

City Leadership Initiative

Civic Analytics Network Run by Harvard University

Data-Driven Justice Police reform initiative

Educational Partnerships for Innovation in University-community partnership that Communities Networks (EPIC-N) seeks innovation at the local level with input from individual community members

FUSE Corps Recruit entrepreneurial individuals to serve in city government for a fixed amount of time

Innovation Teams (i-teams) “in-house innovation consults”

Living Cities Works to improve lives of low income people

MetroLab Network A University-City partnership designed to prioritize innovation, especially cutting- edge technology.

National Neighborhood Indicators Run by the Urban Institute, the organization Partnership (NNIP) seeks to inform policymaking with data and analysis

Next Century Cities Makes the case to greater Internet access

Police Data Initiative (PDI) Encourages cities to open data related to policing

StriveTogether Focuses on increasing educational opportunities through training and strategic assistance

TechHire Teaches technology skills and assists individuals in finding employment

Urban Sustainability Directors Network Peer-to-peer network of government (USDN) employees that works to encourage local governments adopt best practices to result in healthier environment, economic prosperity, and increased social equity

71

What Works Cities (WWC) Funded by the Bloomberg Foundation, What Works Cities assists with data analysis and evidence-based practice.

4.2.6.5 Broadband

The broadband data come from the Federal Communications Commission (FCC Releases 2018 Broadband Deployment Report, 2018). The agency released updated data on household broadband Internet speeds in 2018 (FCC Releases 2018

Broadband Deployment Report, 2018).

Table 4. Broadband Original Variables Variable Name Explanation State, County or County Equivalent - County, State - Population Evaluated - % of Pop. with Fixed 25 Mbps/3 Mbps 25 megabits per second download / 3 megabits per second upload % of Pop. with Mobile 5 Mbps /1 Mbps 5 megabits per second download / 1 megabits per second upload % of Pop. with Fixed & Mobile - Population Density - Per Capita Income ($2016) -

4.3 Data Analysis Analysis of the data was accomplished with a multistage procedure aimed at addressing the hypotheses. First, data was cleaned and coded and examined for irregularities. Next, a zero-ordered correlation matrix of all variables was conducted. Following that, a logistic regression equation was fitted to the data. Regression diagnostics were conducted to check for linearity, homoscedasticity, independence,

72

normality, and appropriate remedial strategies were employed. Missing data were managed with appropriate strategies. A logistic regression was conducted because logistic regression allows for conducting linear modeling with a binary dependent variable. Linear regression conducted with a continuous dependent variable is assumed to take on a normal distribution. Logistic regression models the probability of an event occurring and therefore does not use a continuous variable, but rather a binary variable of either zer or one (Gelman & Hill, 2007). The assumed distribution of a logistic regression, then, runs from zero to one with zero meaning an event did not take place and one meaning that it did occur (Peng & So, 2002). While events either take place or out, logistic regression analysis assigns a probability of a particular event occurring based on statistical analysis of the independent variables relative to the dependent variables. The current model used is direct, meaning that all of the varibles were included in one model (Stoltzfus, 2011). Further research will build on the results of this dissertation by using sequential modeling (Stoltzfus, 2011). The analysis calculates p- values for each independent variable, chi-squared for the model, and reports degrees of freedom. The independent variable calculations are reported as beta values and as odds-ratios. Odds-ratios allow for a clearer interpretation of the beta values relative to whether an individual independent variable increases or decreases the likelihood of the dependent variable occurring. When using logistic regression chi-squared is used to measure whether the model, in this case, the direct model, improves upon the performance of the null model.

73

퐿푁(푌̂I / 1 - 푌̂) = Bo + 퐵̂1i+ 퐵̂2i + 퐵̂3i + 퐵̂4i + 퐵̂5i + 퐵̂6i + 퐵̂7i + Ei

푌̂ = Probability of a Code for America Brigade 퐵̂1I = County Population 퐵̂2I = Social Capital Index 퐵̂3I = Creative Class Index 퐵̂4I = PhDs 퐵̂5I = Broadband 퐵̂6I = Networks 퐵̂7I = Distress Score

4.4 Limitations

The limitations of this study are that proxy variables were used to measure the dependent variable and the independent variables of interest. Civic technology is an evolving field and to measure its growth or decline, decisions must be made about what it is and how to measure it. The chosen course for this study is that civic technology is civic participation, but civic technology manifests itself in other ways as well, i.e., information and communication applications, open data policies, and open data portals. Given the definition of civic technology used in this dissertation, the dissertation argues that affiliation with recognized civic technology groups is a robust measure. Of course, the ideal measure would combine all aspects of civic technology into a measurement construct. Future work will build on the approach taken in this study.

Similarly, traditional social capital, creative class, and networked social capital are to be more fully developed. All three of these measures attempt to capture elusive concepts – that of relationships among individuals and of an evolving economy. This is particularly true when it comes to Networked Social Capital. While many of the

74

organizations included in the New America Foundation data are non-profit organizations coordinating groups across the country, this is not true of all groups included in the variable. Some of the groups included in the variable are government organizations.

75

RESULTS

The research question of this dissertation is who is adopting civic technology and why? The dissertation operationalized this research question by asking if a Code for America Brigade was present in urban counties in the United States. This question was asked in relation to other characteristics of those counties, including population, social capital, creative class, education, networked non-profits, and economic distress. The study hypothesized that urban counties with higher levels of population, creative class, education, and networked non-profits, would be more likely to have Code for America Brigades. The study also hypothesized that urban counties with higher traditional social capital and economic distress would be less likely to have Code for

America Brigades.

76

Table 5. Descriptive Statistics & Bivariate Correlations M SD 1 2 3 4 5 6 7 1 Brigade NA NA ------2 Population 201534 697905.6 .19** ------3 Social Capital -.44 .51 -.11** -.06* - - - - - 4 Creative Class .22 .04 .23** .5** .17** - - - - 5 PhDs 166.29 179.93 -.02 .22** .25** .22** - - - 6 Broadband .85 .19 .25** .17** .03 .28** -.01 - - 7 Networks .16 .11 .25** .47** -.19** .24** .06 .02 - 8 Distress Score 418.02 310.48 -.01 .31** -.28** .11** .05 -.02 .03 *p < .05. **p < .01.

Table 5 presents the Zero ordered correlations for the dependent variables, independent variables and covariants. The presence of a Code for America Brigade is positively correlated with networked social capital arrangements, creative class, population, and broadband Internet. The relationships are modest but statistically significant. Alternatively, the presence of a Code for America Brigade is negatively correlated with the traditional social capital arrangement, advanced education and economic distress score. The relationship with social capital is modest and the relationship with economic destress and advanced education are quite small.This suggests that areas with greater levels of creative class are more associated with Code for America Brigades while areas with greater levels of traditional social capital are less associated with Code for America Brigades.

The correlation matrix does not display evidence of multicollinearity. The strongest relationship between independent variables is between population and creative class, but it does not reach a threshold that warrants concern of multicollinearity. There is a similar dynamic between population and networked social capital, but again it does not reach a threshold that warrants concern of multicollinearity. Interestingly, PhDs have a similar association to social capital as

77

they do with creative class. This suggests that the human capital hypothesis of Berry and Gleaser (2005) is completely distinct from Florida’s creative class hypothesis. Given that there is no indication of multicollinearity and that the correlations reported in Table 5 lend evidence to the dissertations hyptotheses, it makes sene to move forward with the analysis. A logistic regression model was fitted to the data. The results are presented in Table 6.

Table 6. Summary of Logistic Regression Analysis for Variables Predicting CfA

Brigades

Predictors B SE B Odds Ratio

(Intercept) -8.308e+00*** 9.470e-01 .00

Population 4.759e-08 2.541e-07 1.00 Social Capital -9.338e-01*** 1.959e-01 .39

Creative Class 2.701e+01*** 3.224e+00 537053603781.47

PhDs 1.553e-03** 5.257e-04 1.00 Broadband 6.610e-01 4.555e-01 1.94

Networks 1.162e+01*** 1.097e+00 157772.27

Distress Score -2.547e-02*** 6.851e-03 .97

푥2 580.94***

df 7

*p < .05. **p < .01. ***p < .001.

Table 6 shows beta values, standard error, odds ratio, and p-value significance for each independent variable. The table also shows the chi-square and degrees of freedom measure for the entire model.

78

Social science researchers use several different methods to evaluate the performance of logistic regression models (Peng & So, 2002). The model in this study is evaluated using chi-squared (Peng & So, 2002; Stoltzfus, 2011). Chi-squared is a statistical test used to compare the cross-tabulations of categorical variables (Fox,

2003; Meier et al., 2012). When evaluating cross tabulation,s chi-squared evaluates whether the difference between observed and expected values are statistically significant. It does this based on overall proportions. For logistic regression, Chi- squared is used to evaluate the difference between the null model and the fitted model.

In this analysis, chi-squared was used to evaluate whether the fitted model out performed the null model in its error rate (Gelman & Hill, 2007). The chi-squared statistic test presented in the table shows the deviance value as well as the p-value.

The deviance value is 580.9 and the p-value is less than .001. This means that the fitted model out performs the null model for error rate. The fitted model made fewer incorrect predictions regarding whether an individual county did or did not have a

Brigade present compared to the null model. The combination of independent variables used in the direct model (Stoltzfus, 2011) improves upon a model with no independent variables. Except for population and broadband Internet, all beta values of independent variables are significant at the .01 level. This means that before examining the odds ratios, the beta values of the logistic regression analysis does not find population and broadband Internet access to be statistically significant. The practical implication of this statistical finding is that the statistical analysis does not find evidence that variation of population and levels of broadband Internet access significantly affect the presence or absence of Code for America Brigades. Conversely, the beta values of the

79

logistic regression analysis find social capital, creative class, PhDs, networks, and distress score to be statistically significant. The practical implication of this statistical finding is that the variation of social capital, creative class, PhDs, networks, and distress score significantly affects the presence or absence of Code for America

Brigades. The odds ratio provides a likelihood estimate for the variables. A likelihood estimate provides a numeric estimation of the effect that each independent variable has on the probability of the dependent variable. The likelihood estimate provides an estimate of whether a one-unit increase of the independent variable increases or decreases the probability of a positive occurrence of the dependent variable. Creative class has an overwhelming impact on the likelihood of the presence of a Brigade, as does the presence of networked social capital organizations.

80

Figure 13. Receiver operating characteristic (ROC) Curve

Figure 13 is a Receiver Operating Characteristic (ROC) Curve. When interpreting classification tables, it is important to note the ratio of negative to positive observations. The reason this is important is that if there are very few positive observations then the model will overperform relative to making negative predictions rather then positive predictions. When observing the presence of Code of America

Brigades, there are 657 negative observations (no Code for America Brigade) and 508 positive observations (presence of a Code for America Brigade). That means that 56% of the observations are negative and 44% of the observations are positive. The

81 negative to positive observation ratio is slightly skewed towards negative observations.

The area under the curve is an important measure for evaluating the accuracy of the model as is the shape of the curve. The Y axis shows positive outcomes and the

X axis shows all other outcomes. The gray, diagonal line is the baseline prediction. In that baseline prediction, the model cannot predict the outcome more accurately than random chance. The further away the blue line is from the diagonal, the better the model is performing.

82

DISCUSSION

The research questions posed by this dissertation concern who is adopting civic technology and why. The dissertation operationalizes this question by analyzing the presence of Code for America Brigades in relation to socioeconomic characteristics of urban counties in the United States. The main hypothesis is that new economic forms are giving rise to new forms of civic association, exemplified by Code for America

Brigades. This dissertation operationalizes this hypothesis through the independent variables of traditional social capital and creative class. Many scholars have argued that American society thrives on voluntary associations among its people. By choosing to spend time with one another in churches, bowling alleys, or in the case this dissertation is concerned with, online forums and public spaces utilizing technology to solve community problems, people generate resources that have the potential to confer benefits in the form of building skills, getting jobs, and solving problems.

83

Given the questions addressed by this research project, the following hypotheses were generated:

H1 Urban counties with higher levels of social capital will have lower levels of

civic technology

H2 Urban counties with higher levels of creative class employment will have

higher levels of civic technology

H3 Urban counties with higher levels of advanced education will have higher

levels of civic technology

H4 Urban counties with larger populations will have higher levels of civic

technology

H5 Urban counties with higher levels of networked non-profit arrangements will

have higher levels of civic technology

H6 Urban counties with higher levels of distress scores will have lower levels of

civic technology

H7 Urban counties with higher levels of broadband speed will have higher levels

of civic technology

84

6.1 Findings

Traditional social capital was statistically significant but not substantively significant. This makes sense in relation to the hypothesis. The theoretical reasons for this were established in the conceptual chapter of the dissertation. Social capital did not predict civic technology activity. If anything, the presence of higher levels of social capital slightly decreased the probability of civic technology activity. This result supports the finding that traditional social capital networks are not significantly connected to civic technology networks, particularly Code for America Brigades.

Creative class was statistically significant and substantively significant. This also makes sense in relation to the hypotheses. The theoretical reasons were established in chapter three of the dissertation. Creative class did predict civic technology activity. The presence of higher levels of creative class drastically increased the probability of civic technology activity. This result lends evidence to the finding that creative class networks are extending to civic technology networks, particularly Code for America Brigades.

As stated in chapter five, the traditional social capital index is a proxy for a group of activities associated with civil society; it measures a grouping of activities that arose in particular historical circumstances thought to result in social capital.

Similarly, creative class is not a measure of social capital. Rather, it is a measure of occupations that are hypothesized to have common incentive structures for career advancement. The hypothesis of this dissertation is that new economic forms lead to new forms of civic association. A more specific formulation of this hypothesis is that the emergence of the creative class as an economic force presents the opportunity for these individuals to be mobilized for social goals incentivized through professional development and career advancement (Salisbury, 1969).

85

Past research on the diffusion of innovation has found population and economic health to be predictive of early adoption (Damanpour & Schneider, 2009; Kearney et al., 2000; Moon, 2000). A large body of literature argues that economic health is premised on technological factors (Porter, 1990). Neither broadband Internet access nor population were statistically significant. The reason that broadband Internet may not predict civic technology activity in a significant degree is that broadband Internet access is no longer an advanced factor (Porter, 1990). While a digital divide persists among pockets of urban communities, the level at which the analysis was conducted may mask those inequities. The reason that population may not predict civic technology activity in a significant degree could be methodological. It could be that because the study population was urban counties, that the smallest population in the range under study was large enough to exhibit agglomeration effects. Advanced education was statistically significant but not substantively significant. This is interesting. The results support a finding that lends evidence to the creative class hypothesis but not the human capital hypothesis. There is overlap between the creative class variable and the PhD variable. While there is not significant multicollinearity, it could be that the creative class variable is absorbing the impact of the PhD variable.

Distressed community index was statistically significant, but the odds ratio suggests that the variable is not substantively significant for predicting the presence of

Code for America Brigades. The reason the question of the relationship between civic technology and economic distress was raised in the first place is due to the economic polarization the United States has experience since the 1970s. The industrial Midwest saw the collapse of the steel industry while Silicon Valley saw the meteoric rise of the

86

technology industry. Given that creative class is associated with prosperity, it was hypothesized that economic distress would negatively predict civic technology activity. This is true, but only to a minimal extent. This could be due to economic inequality throughout the United States. This is to say that economic polarization does not only occur among regions, but is also occurs within regions. There could be certain regions, though certainly at all, with a large creative class index that could have also have significant income inequality which results in economic distress. Networked social capital was statistically significant and substantively significant. This is in line with the hypothesis of the dissertation. The hypothesis was that places with relatively more networked social capital arrangements would be more likely to have Code for America Brigades than places were relatively less networked social capital arrangements. The results of the logistic regression analysis provide clear evidence to support hypotheses one, two, five, and seven. Similarly, the results of the logistic regression analysis do not provide clear evidence to support hypotheses three, four, and six. Of the seven hypotheses generated in chapter three, the hypotheses related to social capital and distress score were anticipated to decrease the likelihood of Code for America Brigade adoption. Of these two independent variables, social capital was the only variable that conveyed significant substantial evidence of decreasing the likelihood of Code for America Brigade adoption. The remaining hypotheses (creative class, population, advanced education, networked social capital, and broadband

Internet) were anticipated to increase the likelihood of Code for America Brigade adoption. Of these creative class and networked social capital conveyed significant

87

substantial evidence of increasing the likelihood of Code for America Brigade adoption.

6.2 Conclusions

The diffusion of innovation framework assumes that over time all innovations will be fully adopted by society. This, however, is only an assumption. In practice, societal dynamics shape the speed and extent to which communities adopt specific innovations. This dissertation investigated the associational patterns of urban

American communities to provide insights regarding the differences among communities that are adopting Code for America Brigades and those that are not. The dissertation first outlined civic technology as an innovation. The civic technology movement is an attempt to intentionally build technologies that effectively serve the individuals and communities who use those technologies. Often, the civic technology movement attempts to realize this goal by involving individuals and communities in the process of designing and building technologies. As a volunteer practice, the civic technology movement is organized around open civic data, technology, and innovative practices. Code for America Brigades exemplify the civic technology movement because they make use of all three components of civic technology. In generating hypotheses about the differences among urban American communities the dissertation used insights from the diffusion of innovation literature as well as the social capital and creative class literatures. The dissertation relied heavily on the concept of networks and the principle of self-interest in order to generate expectations as to why certain communities were adopting Code for America Brigades and others were not.

88

Networks are channels of communication through which information and ideas spread. Individuals are connected to networks through their jobs, families, and voluntary associations. Networks are held together by social capital. Affiliation through employment, familial bonds, and voluntary association carry certain obligations that maintain the network’s structure. Attending holiday events, showing up for work on time, and taking free time to attend a lecture or training demonstrate affiliation and commitment to particular networks to the other members of that network. Networks also offer benefits to individual members. While showing up on

Christmas Day means that an individual might have an awkward conversation or two, it also means that they are not alone on an important holiday. Assisting fellow employees with projects at times means extra work, but builds good will that can facilitate coordinated future actions. Attending a local Code for America Brigade event may mean that one has to postpone a well-earned night of binging a favorite show, but it provides access to individuals in close proximity with technology and data analysis skills willing to in part their knowledge. Network affiliation generally precludes affiliation with other networks because network obligations necessitate further involvement in those networks and do not leave time for other commitments. Networks are not identical. They differ in terms of norms and obligations. Civic technology, particularly Code for America, is one of a number of networked social capital arrangements that stand in contrast to traditional social capital arrangements. Whereas traditional social capital arrangements are based on commitments to older value sets (Bishop & Cushing, 2008) and are generally based on reciprocity among community members, networked social capital arrangements are

89

based on a reaction to contemporary societal problems, and generally integrate technology as an aspect of that response (Schuler, 1996), and offer benefits to those who contribute (Olson, 1971). In other words, civic technology is a social movement to make the everyday work of government data-driven, participatory, and technologically accessible (Noveck, 2015). Communities, then, with substantial amounts of individuals who can access these benefits are more likely to adopt Code for America Brigades than communities who cannot access such benefits. Benefits include professional development and career advancement. Communities with large numbers of individuals in the creative class were hypothesized to be those communities that can access these benefits. The dissertation operationalized that idea with secondary data sources on Code for America Brigades, traditional social capital, creative class, and networked social capital. The results of the logistic regression analysis found evidence that communities with relatively higher numbers of individuals in the creative class are more likely to adopt Code for America Brigades while those with relatively higher levels of traditional social capital are slightly less likely to adopt Code for America Brigades. This finding is consistent with theories of collective action which argue that individuals do not participate in social movements solely to realize collective benefits

(Olson, 1971). This is not to say that individuals who participate in the civic technology movement do not believe in the ethical and/or professional standards of the movement. Rather, it is to say that the realization of collective benefits is not enough to sustain engagement (Olson, 1971). In order for Code for America Brigades to fully diffuse, Code for America needs to offer substantial individual benefits for that outside of the creative class.

90

Code for America needs to focus on being a platform for civic participation. O’Reilly

(2010) argued that the government should facilitate civic participation. The formation of Code for America in part answered this call by providing a forum for technologists to experiment with building effective solutions to community problems utilizing technology. The findings of this dissertation make clear that up to this point it is generally communities that could be considered a constituency of Code for America that has been mobilized to this point. Code for America has the opportunity to make inclusive platform governance a central aspect of its mission.

91

REFERENCES

2010 Standards for Delineating Metropolitan and Micropolitan Statistical Areas.

(2010). Federal Register. https://www.gpo.gov/fdsys/pkg/FR-2010-06-

28/pdf/2010-15605.pdf

Arnstein, S. (1969). A ladder of citizen participation. Journal of the American Institute

of Planners, 35(4), 216–224.

Attard, J., Orlandi, F., Scerri, S., & Auer, S. (2015). A systematic review of open

government data initiatives. 32(4), 399–418.

Barnekov, T., Rich, D., & Warren, R. (1981). The New Privatism, Federalism, and

The Future of Urban Governance: National Urban Policy in the 1980s. Journal

of Urban Affairs, 3(4), 1–14.

Bass Warner, S. (1968). The Private City. Trustees of the University of Pennslyvania

Press.

Bennis, W. G., Benne, K. D., & Chin, R. (1985). The planning of change. Holt,

Rinehart, and Winston.

Berry, C. R., & Glaeser, E. L. (2005). The divergence of human capital levels across

cities. Papers in Regional Science, 84(3), 407–444.

Bishop, B., & Cushing, R. G. (2008). The big sort: Why the clustering of like-minded

America is tearing us apart. Houghton Mifflin.

92

Bluestone, B., & Harrison, B. (1982). The Deindustrialization of America. Basic

Books.

Borgstrom, A., Druker, D., & Sparrow, J. (2005). The Technology Opportunities

Program (TOP): Networking Our Nation-A Decade of Lessons Learned.

Community Development: Journal of the Community Development Society;

36(1), 103–119.

Bowyer, B., & Kahne, J. (2016). Revisiting the Measurement of Political Participation

for the Digital Age. In Civic Media: Technology, Design, and Practice. The

MIT Press.

Brady, H. E., Verba, S., & Schlozman, K. L. (1995). Beyond SES: A Resource Model

of Political Participation. American Political Science Review, 89(2), 271–294.

https://doi.org/10.2307/2082425

Briscoe, G. (2014). Digital Innovation: The Hackathon Phenomenon. Creativeworks

London, Queen Mary Research Online, 24.

http://qmro.qmul.ac.uk/xmlui/handle/123456789/11418

Carter, D. B. (2017). The Structure of Environmental Movements in Small U.S. States

and the Implications for Environmental Policy. PhD Dissertation. University

of Delaware.

Castells, M. (1996). The rise of the network society. Blackwell Publishers.

Chambers, C. A. (2010). Seedtime of reform: American social service and social

action, 1918-1933. University of Minnesota Press.

93

Chappellet-Lanier, T. (2019, January 15). The OPEN Government Data Act is now

law. FedScoop. https://www.fedscoop.com/open-government-data-act-law/

Civic Tech Field Guide. (n.d.). Civic Tech Field Guide. Retrieved March 26, 2020,

from https://civictech.guide/

Code for America (2020) About the Brigades.

https://brigade.codeforamerica.org/about

Code for America. (2018, March 28). Who we are. Code for America.

https://www.codeforamerica.org/who

Code for America. (2019, March 6). Integrated Benefits Initiative. Code for America.

Connor, D. M. (1988). A new ladder of citizen participation. National Civic Review,

77(3), 249–257. https://doi.org/10.1002/ncr.4100770309

Damanpour, F., & Schneider, M. (2009). Characteristics of Innovation and Innovation

Adoption in Public Organizations: Assessing the Role of Managers. Journal of

Public Administration Research and Theory, 19(3), 495–522.

https://doi.org/10.1093/jopart/mun021

Data Ethics Canvas. (2019). Open Data Institute. https://theodi.org/wp-

content/uploads/2019/07/ODI-Data-Ethics-Canvas-2019-05.pdf

Data4Good. (2020, March). Data4good. https://data4good.com.au/

David, N., McNutt, J. G., & Justice, J. B. (2018). Smart Cities, Transparency, Civic

Technology and Reinventing Government. In Smart Technologies for Smart

94

Governments (pp. 19–34). Springer, Cham. https://doi.org/10.1007/978-3-319-

58577-2_2

Davies, S., Wiley-Schwartz, A., Pinkett, R., & Servon, L. (2003). Community

Technology Centers as Catalysts for Community Change. Ford Foundation.

Dewey, J. (2012). The public and its problems: An essay in political inquiry.

Pennsylvania State University Press.

DiMaggio, P. J., & Powell, W. W. (1983). The Iron Cage Revisited: Institutional

Isomorphism and Collective Rationality in Organizational Fields. American

Sociological Review, 48(2), 147–160. https://doi.org/10.2307/2095101

Dreier, P., Mollenkopf, J., & Swanstrom, T. (2014). Place matters: Metropolitics for

the twenty-first century (Third). University Press of Kansas.

Dube, S. (2018). How Counties Can Use Evidence-Based Policymaking to Achieve

Better Outcomes. Pew Research Center. https://pew.org/2KS9YAJ

FCC Releases 2018 Broadband Deployment Report. (2018, February 2). Federal

Communications Commission. https://www.fcc.gov/document/fcc-releases-

2018-broadband-deployment-report

Federal Funding Accountability and Transparency Act Subaward Reporting System,

Pub. L. No. 109–282 (2006). https://www.fsrs.gov/

Florida, R. L. (2012). The rise of the creative class, revisited. Basic Books.

Fox, W. (2003). Social statistics. Wadsworth Publishing Co.

From Great Recession to Great Reshuffling: Charting a Decade of Change Across

American Communities. (2018). Economic Innovation Group.

95

Fung, A. (2006). Varieties of participation in complex governance. Public

Administration Review, 66(Special Issue), 66–75.

Gelman, A., & Hill, J. (2007). Data analysis using regression and

multilevel/hierarchical models. Cambridge University Press.

GetCalFresh.com. (2020, February). https://www.getcalfresh.org/en

Gilman, H. (2016). Participatory Budgeting and Civic Tech. Georgetown University

Press. http://press.georgetown.edu/book/georgetown/participatory-budgeting-

and-civic-tech

Gilman, H. R. (2016). Democracy reinvented: Participatory budgeting and civic

innovation in America. Brookings Institution Press.

Gordon, E., & Walter, S. (2016). Meaningful Inefficiencies: Resisting the Logic of

Technological Efficiency in the Design of Civic Sytems. In E. Gordon & P.

Mihailidis (Eds.), Civic Media: Technology, Design, Practice (1 edition). The

MIT Press.

Graeff, E. (2018a). Evaluating Civic Technology Design for Citizen Empowerment.

PhD Thesis, Massachusetts Institute of Technology.

Graeff, E. (2018b, September 20). The Need for Empowerment-based Design in Civic

Technology. The Internet, Policy & Politics Conference, Oxford University.

Granovetter, M. S. (1973). The Strength of Weak Ties. American Journal of

Sociology, 78(6), 1360–1380.

Gurin, J. (2014). Open Data Now: The Secret to Hot Startups, Smart Investing, Savvy

Marketing, and Fast Innovation. McGraw-Hill.

96

Heald, D. (2006). Varieties of Transparency. Oxford University Press.

Hendler, J., Sosnik, O., Prudner, G., Chernicoff, D., McCarthy, J., Ree, O., McGhee,

J., Bloomfield, E., & Ribeiro, M. (2016). Engines of Change: What Civic Tech

Can Learn From Social Movements. Omidyar Network.

https://www.omidyar.com/spotlight/what-can-civic-tech-learn-social-

movements

Hou, Y. (2018). Understanding the Design and Implementation of Civic Technologies

in Resource-Limited Public Organizations. University of Michigan.

Johnson, P., & Robinson, P. (2014). Civic Hackathons: Innovation, Procurement, or

Civic Engagement? Review of Policy Research, 31(4), 349–357.

https://doi.org/10.1111/ropr.12074

Justice, J., McNutt, J., Melitski, J., Ahn, M., David, N., Siddiqui, S., & Ronquillo, J.

(2018). The Civic Technology Movement. In Technology, Activism, and Social

Justice in a Digital Age. Oxford University Press.

Kearns, K. P. (1992). Innovations in local government: A sociocognitive network

approach. Knowledge and Policy, 5(2), 45–67.

https://doi.org/10.1007/BF02692805

Lathrop, Daniel., & Ruma, Laurel. (2010). Open government: Collaboration,

transparency, and participation in practice. O’Reilly.

Maloney, W., Smith, G., & Stoker, G. (2000). Social capital and urban governance:

Adding a more contextualized ‘top-down’ perspective. Political Studies, 48,

802–820.

97

Markusen, A. (1991). The military-industrial divide. Environment and Planning D:

Society and Space, 9, 391–416.

Markusen, A. (1996). Sticky Places in Slippery Space: A Typology of Industrial

Districts. Economic Geography, 72(3), 293–313.

https://doi.org/10.2307/144402

McCann, L. (2015). Experimental Modes of Civic Engagement in Civic Tech. Smart

Chicago Collaborative. http://www.smartchicagocollaborative.org/wp-

content/uploads/2015/10/experimental-modes.pdf

Mcgranahan, D., & Wojan, T. (2007). Recasting the Creative Class to Examine

Growth Processes in Rural and Urban Counties. Regional Studies, 41(2), 197–

McGuinness, T., Passen, H., & Ross, D. (2018, December). Networks and American

Renewal: An Interactive Map and Collection of Articles on Connectivity

Across American Communities. New America.

https://www.newamerica.org/national-network/reports/networks-and-

american-renewal/

McGuinness, T., & Slaughter, A.-M. (2019, Spring). The New Practice of Public

Problem Solving. The Stanford Social Innovation Review, 17(2), 26–33.

McNutt, J. G. (2018). Technology, activism, and social justice in a digital age. Oxford

University Press.

McNutt, J., & Justice, J. (2016). Predicting civic hackathons in local communities:

Perspectives from social capital and creative class theory. University of

Delaware, Working Paper Series.

98

McNutt, J., Justice, J., Melitski, J., Ahn, M., Siddiqui, S., Carter, D., & Kline, A.

(2016). The diffusion of civic technology and open government in the United

States. Information Polity, 21, 153–170.

Meier, K. J., Brudney, J. L., & Bohte, J. (2012). Applied statistics for public and

nonprofit administration. Wadsworth Publishing Co.

Moreno, C. (2018, August). Courtbot software aims to help break the cycle of high

incarceration of women in Tulsa, Oklahoma. Unpublished manuscript.

Morris, A. (1981). Black Southern Student Sit-in Movement: An Analysis of Internal

Organization. American Sociological Review, 46(6), 744–767. JSTOR.

https://doi.org/10.2307/2095077

Mutrux, M. (2018, October 19). Redeploying Courtbot. Code for America Brigade

Congress, Charlotte, NC.

Norris, P. (2013). Democratic phoenix: Reinventing political activism. Cambridge

University Press.

Noveck, S. B. (2015). Smart Citizens, Smarter State: The Technologies of Expertise

and the Future of Governing. Harvard University Press.

Noveck, B. S. (2011). The single point of failure. In Innovating government (pp. 77-

99). TMC Asser Press.

Noveck S. B. (2016). Re-Imagining Government through Civic Media: Three

Pathways to Institutional Innovation. In E. Gordon & P. Mihailidis (Eds.),

Civic Media: Technology, Design, Practice. The MIT Press.

99

Noyelle, T. (1983). The Rise of Advanced Services: Some Implications for Economic

Development in U.S. Cities. Journal of the American Planning Association,

49(3), 280-290.

Olson, Mancur. (1971). The logic of collective action: Public goods and the theory of

groups. Harvard University Press.

OpenPlans. (2012). Field Scan of Civic Technology. Living Cities.

O’Reilly, T. (2010). Government as a platform. In Open government: Collaboration,

transparency, and participation in practice.

https://www.mitpressjournals.org/doi/pdf/10.1162/INOV_a_00056

Pahlka, J. (2012, February). Coding a better government.

https://www.ted.com/talks/jennifer_pahlka_coding_a_better_government

Patel, M., & Sotsky, J. (2013). The Emergence of Civic Tech: Investments in a

Growing Field. Knight Foundation.

https://knightfoundation.org/articles/knight-foundations-civic-tech-report-why-

it-matters

Peixoto, T., & Sifry, M. (Eds.). (2017). Civic tech in the global south. International

Bank for Reconstruction and Development/The World Bank.

Peng, C.-Y. J., & So, T.-S. H. (2002). Logistic Regression Analysis and Reporting: A

Primer. Understanding Statistics, 1(1), 31–70.

Podder, S., Kumar Singh, S., Tanguturi, P., & Roy, P. (2018). Technology for Good:

Scaling up Social Transformation in the Fourth Industrial Revolution.

100

Accenture. https://www.accenture.com/us-en/insights/future-

workforce/artificial-intelligence-blockchain-techforgood

Porter, M. (1990). The Competitive Advantage of Nations. The Free Press.

Putnam, R. (1993). Making Democracy Work. Princeton University Press.

Putnam, R. D. (2000). Bowling alone: The collapse and revival of American

community. Simon & Schuster.

Putnam, R. D. (2015). Our kids: The American Dream in crisis. Simon & Schuster

Quinn, H. (2018, March 21). There’s an open data hackathon April 7 and it’s called

“Data Jam” this year. Technical.Ly Delaware.

https://technical.ly/delaware/2018/03/21/theres-open-data-hackathon-april-7-

called-data-jam-year/

Rogers, E. M. (2003). Diffusion of innovations. [Fourth Edition] Free Press.

Rupasingha, A., Goetz, S. J., & Freshwater, D. (2006). The production of social

capital in US counties. Journal of Socio-Economics, 35, 83–101.

https://doi.org/doi:10.1016/j.socec.2005.11.001

Ryan, B., & Gross, N. (1950). Acceptance and Diffusion of Hybrid Corn Seed in Two

Iowa Communities. Research Bulletin (Iowa Agriculture and Home Economics

Experiment Station), 29(372, number 1).

Salisbury, R. H. (1969). An Exchange Theory of Interest Groups. Midwest Journal of

Political Science, 13(1), 1–32. JSTOR. https://doi.org/10.2307/2110212

Salmon, C. T., Fernandez, L., & Post, L. A. (2010). Mobilizing public will across

borders: Roles and functions of communication processes and technologies.

101

Journal of Borderlands Studies, 25(3–4), 159–170.

https://doi.org/10.1080/08865655.2010.9695778

Saxenian, A. (1996). Regional advantage: Culture and competition in Silicon Valley

and Route 128. Harvard University Press.

Schlozman, K. L., Brady, H. E., & Verba, S. (2018). Unequal and unrepresented:

Political inequality and the people’s voice in the new Gilded Age.

Schrock, A. (2016). Case Study: Code for America—Scaling Civic Engagement

through Open Data and Software Design. In Civic Media: Technology, Design,

& Practice. MIT Press.

Schrock, A. (2018). Civic Tech. Rouge Academic Press.

Schudson, M. (1998). The good citizen: A history of American civic life. Martin

Kessler Books.

Schuler, D. (1996). New community networks: Wired for change. ACM

Press/Addison-Wesley Publishing Co.

Sellers, J. (2002). The Nation-State and Urban Governance: Toward Multilevel

Analysis. Urban Affairs Review, 37(5), 31.

Shirky, C. (2010). Cognitive surplus: Creativity and generosity in a connected age.

Penguin Press.

Sirianni, C. (2009). Investing in Democracy: Engaging Citizens in Collaborative

Governance. Brookings Institution Press.

Smith, E. (2018, January 20). The Tale of the Free-Net, the Cheap Way Dial-Up Users

Got Online—Motherboard. Motherboard.

102

https://motherboard.vice.com/en_us/article/qvwayd/cleveland-free-net-the-

cheap-way-dial-up-users-got-online

Smith, G. (2009). Democratic innovations: Designing institutions for citizen

participation. Cambridge University Press.

Smith, S., & Lipski, M. (1995). Nonprofits for Hire: The Welfare State in the Age of

Contracting. Harvard University Press.

Solomon, J. (2015). A User-Centered Approach to Food Stamps.

https://www.youtube.com/watch?v=lqTFi2U2Ebc&t=228s

Stempack, M. (2016, April 27). Towards a taxonomy of civic technology. Microsoft on

the Issues. https://blogs.microsoft.com/on-the-issues/2016/04/27/towards-

taxonomy-civic-technology/

Stepasiuk, T. (2014). Civic hacking: A motivational framework. New Visions for

Public Affairs, 6, 21–30.

Stephens, J. (2017). Civic Technology: Open Data and Citizen Volunteers as a

Resource for North Carolina Local Governments. The University of North

Carolina at Chapel Hill.

Stephens, J. (2018, June). What is civic tech for?: Barn Raising V Empowerment.

Stoltzfus, J. (2011). Logistic Regression: A Brief Primer. Academic Emergency

Medicine, 11, 1099–1104.

Tauberer, J. (2007). 8 Principles of Open Government Data. https://opengovdata.org/

The Code for America Brigade Network. (2020, April).

https://brigade.codeforamerica.org/

103

The Data Spectrum – The ODI. (n.d.). Retrieved April 7, 2020, from

https://theodi.org/about-the-odi/the-data-spectrum/

The White House. (2009). Open Government Directive.

https://obamawhitehouse.archives.gov/node/7011

Thompson, W., & Thompson, P. (1987). National Industries and Local Occupational

Strengths: The Cross-Hairs of Targeting. Urban Studies, 24(547), 1987.

Tiebout, C. M. (1956). A Pure Theory of Local Expenditures. Journal of Political

Economy, 64(5), 416–424.

Tilly, C., & Wood, L. J. (2013). Social movements, 1768-2012. Paradigm Publishers.

Tocqueville, A. de, Mayer, J. P., & Lerner, M. (1966). Democracy in America. Harper

& Row.

Townsend, A. M. (2013). Smart cities: Big data, civic hackers, and the quest for a new

utopia. WW Norton & Company.

Tufekci, Z. (2017). Twitter and Tear Gas: The Power and Fragility of Networked

Protest. Yale University Press.

Turkel, E. (2019, Spring). The Attributes of Civic Technology. Paper presented at the

American Society of Public Administration, Washington, D.C.

Turkel, E., Suchanic, E., & Neil, R. (2019). Civic Hackathons as Deliberative

Democracy: Reflections from Participation in the 2018 Delaware Open Data

Challenge. New Visions for Public Affairs, 11.

Verba, S., & Nie, N. H. (1991). Participation in America: Political democracy and

social equality. The University of Chicago Press.

104

Verhulst, S., Young, A., & Srinivasan, P. (n.d.). An Introduction to Data

Collaboratives. GovLab, New York University.

Walker, K. (2017, April 4). Use Census geographic data in R with tigris.

https://www.youtube.com/watch?v=lZuVxVONK9g&__s=hpmyiy9wyzwapfz

ug5q9

Warner, M. R. (2014). S.994 - 113th Congress (2013-2014): DATA Act (2013/2014).

https://www.congress.gov/bill/113th-congress/senate-bill/994

Whitaker, C. (2015). The @CivicWhitaker anthology: Three years of organizing,

writing, and documenting in Chicago civic tech. Smart Chicago Collaborative.

Zuckerman, E. (2016). Effective Civics. In E. Gordon & P. Mihailidis (Eds.), Civic

Media: Technology, Design, Practice. The MIT Press.

Zuiderwijk, A., & Janssen, M. (2013). Open data policies, their implementation and

impact: A frameworkfor comparison. Government Information Quarterly,

31(1), 17-29.

Zukin, C., Keeter, S., Andolina, M., Jenkins, K., & Carpini, M. X. D. (2006). A New

Engagement?: Political Participation, Civic Life, and the Changing American

Citizen. Oxford University Press.

105

Appendix

CODE BOOK

Variable Data Source Type Means-Centering Code for America Code for America, Bivariate N/A Brigades New America Foundation Networked Social Code for America, Continuous No Capital New America Foundation Metropolitan U.S. Census Bivariate N/A Social Capital Penn State Continuous No Creative Class USDA Continuous No Broadband Internet FCC Continuous No Education (PhDs) NSF Webcaspar Continuous No Distressed Economic Innovation Continuous No Community Index Group

106

CREATIVE CLASS CODES

Creative class as reformulated by ERS Occupation title Standard Occupation Code (SOC) Management occupations Top executives 11-1000 Advertising, marketing, promotions, public relations, 11-2000 and sales managers Financial managers 11-3030 Operations specialties managers, except financial 11-3010, 11-3020, 11-3040 managers through 11-3070 Other management occupations, except farmers and 11-9020 through 11-9190 farm managers Business and financial operations occupations Accountants and auditors 13-2011 Computer and mathematical occupations Computer specialists 15-1000 Mathematical science occupations 15-2000 Architecture and engineering occupations Architects, surveyors, and cartographers 17-1000 Engineers 17-2000 Drafters, engineering, and mapping technicians 17-3000 Life, physical, and social science occupations

Life and physical scientists 19-1000 and 19-2000

Social scientists and related workers 19-3000

Legal occupations

107

Creative class as reformulated by ERS Occupation title Standard Occupation Code (SOC) Lawyers 23-1011

Education, training, and library occupations

Postsecondary teachers 25-1000

Librarians, curators, and archivists 25-4000

Arts, design, entertainment, sports, and media occupations

Art and design workers* 27-1000*

Entertainers and performers, sports, and related 27-2000* workers* Media and communications workers 27-3000 and 27-4000

Sales and related occupations

Sales representatives, services, wholesale and 41-3000 and 41-4000 manufacturing Other sales and related occupations, including 41-1000 and 41-9000 supervisors *These two categories comprise the arts occupation subset.

108