Abstract
Background:
In the United States, students’ experiences are shaped by racioethnic, socioeconomic, and linguistic segregation. School choice, and especially charter schools, generally exacerbate existing levels of school segregation. Counter to this trend, hundreds of intentionally diverse charter schools (IDCS), with a mission to promote school diversity, have opened across the country.
Purpose:
This study aims to clarify how IDCS operationalize their mission to serve diverse populations. To do so, we examine contextual factors that influence diversification efforts, the strategies that IDCS use to pursue diverse enrollments, and whether sample IDCS are more diverse than comparison schools.
Research Design:
This study uses a convergent mixed-methods design, in which qualitative and quantitative data were independently collected and analyzed, then merged for final analysis. The qualitative phase included thematic analyses of 101 interviews and 40 focus groups with IDCS staff and leaders to identify key contextual factors and practices that shape diversification efforts. For quantitative analyses, we used student and school-level administrative data from sample and comparison schools in New York, Colorado, and California. We explored researcher-created diversity outcomes for sample and comparison schools, including a racioethnic diversity index representing the probability that any two students chosen at random will be of a different race/ethnicity, and locally created diversity goals of having 40%–50% of enrolled students qualify for free/reduced price lunch (FRL) and no racial majority group among enrolled students.
Conclusions/Recommendations:
We found that local context, including race-neutral state policies and local housing patterns, created barriers to recruiting and enrolling diverse students. To overcome these barriers, IDCS staff developed data-driven recruitment and enrollment practices that were differentiated by the target group. Practices focused on increasing awareness of the school, building trust, and investing parents in the schools’ diversity mission. Enrollment data show that sample IDCS are more diverse than a set of comparison schools, with mixed results when analyzing diversity outcomes with locally defined goals. We conclude that, despite contextual barriers, choice schools have considerable agency in fostering school diversity.
Though it has been more than 60 years since the courts ruled that state-sanctioned segregated schools were unconstitutional (e.g., Brown v. Board of Education, 1954; Méndez v. Westminster School District, 1947), racial, socioeconomic, and linguistic segregation continue to shape the experiences of American public school students (Marcotte & Dalane, 2019; Orfield et al., 2016; Owens et al., 2016), perpetuating inequitable access to educational opportunities (Berends & Penaloza, 2010; Spatig-Amerikaner, 2012; Weathers & Sosina, 2022). As districts have been released from federally mandated desegregation plans and, in many cases, experienced resegregation (Fiel & Zhang, 2019; Reardon et al., 2012), the role of local actors in countering segregation has become increasingly important.
Over the same period that American schools experienced resegregation, enrollment in choice schools increased (Wang et al., 2019). The rise of these schools, including magnet schools, intradistrict choice options, and charter schools, means that students and families increasingly have a range of choices in the education marketplace (Cookson, et al, 2018). One particular subset of choice schools, charter schools, have been critiqued for exacerbating segregation because charter schools are more likely to be racially and socioeconomically segregated than traditional public schools and can cause small increases to traditional public school segregation (Alcaino & Jennings, 2020; Frankenberg et al., 2010; Frankenberg & Siegel-Hawley, 2013 Monarrez et al., 2022; Wells et al., 2019).
Counter to this trend, hundreds of intentionally diverse charter schools (IDCS) have opened across the country. While specific school designs vary, IDCS are generally defined as schools with a commitment to diversity at the core of their school mission (Potter & Quick, 2018). Because these schools are relatively new and understudied, limited research has examined the degree to which IDCS successfully recruit and enroll diverse students, the practices they use to do so, and how context influences these processes. (Jabbar & Wilson, 2018; Villavicencio, 2016) Accordingly, this mixed-methods study, drawn from a larger project focusing on IDCS practices and outcomes (Wohlstetter et al., 2021), explores the overarching research question, How do context and school practices shape diversity at IDCS?, through the following questions:
What contextual factors influence diversification efforts at IDCS?
How do IDCS pursue diverse enrollments? How do IDCS communicate their diversity mission during recruitment and enrollment? How do IDCS recruitment and enrollment practices support diversity?
Are IDCS more diverse than comparison schools and are they meeting locally determined diversity goals?
We start by providing a review of the literature on school segregation and choice schools. We then describe our methods and data before presenting our findings. To preview our findings, we found that local housing context and state policy environment influenced IDCS’ use of recruitment and enrollment practices to increase student diversity. IDCS often operated within segregated housing contexts or neighborhoods experiencing gentrification, which meant that recruitment and enrollment practices became critical tools to support student diversity. State policies were often seen as a factor constraining IDCS diversity, and strategic navigation of these policies was required to meet diversity goals. IDCS used a wide range of data-driven recruitment and enrollment practices and employed intentional, differentiated messaging about their diversity mission that aimed to increase families’ investment in diversity. IDCS ultimately had greater student diversity than comparison schools, while success in meeting locally determined goals varied by operator. Although no causal claims can be made about specific practices or outcomes, these findings suggest that, when responsive to local context, schools can successfully increase student diversity. We conclude with a discussion of implications for policy, research, and practice.
Literature Review
In recent decades, there has been a rise in racial and socioeconomic school segregation, with pervasive segregation of low-income students of color both across and within school districts (Orfield et al., 2016; Owens et al., 2016; Reardon et al., 2012; Richards & Stroub, 2020). As such, reducing segregation and establishing integrated schools have become a focus among many education reformers and advocates. Segregation is an ethical and moral issue because it perpetuates and reflects a national legacy of racial and ethnic discrimination (Rosiek & Kinslow, 2015). Empirically, school segregation is a powerful predictor of disparities in academic achievement and resource provision (Berends & Penaloza, 2010; Palardy et al., 2015; Spatig-Amerikaner, 2012; Weathers & Sosina, 2022).
In contrast, school integration has been linked to improved academic, social-emotional, and economic outcomes for students of all racial and socioeconomic backgrounds, with the largest gains among students from historically marginalized groups (Benner & Crosnoe, 2011; Bifulco et al., 2009; Hanushek et al., 2009; Johnson, 2011; Juvonen et al., 2018; Liebowitz, 2018; Siegel-Hawley, 2020; Wells, 2009). For example, Johnson (2011) found long-term benefits for Black students who attended desegregated schools—including increased educational and occupational attainment, improved adult health, and reduced probability of incarceration—but found no impact on White students. However, Benner and Crosnoe (2011) found that White students experienced academic benefits from attending integrated schools with balanced representation of different racioethnic groups. Others also found evidence that students of a variety of backgrounds can benefit from diverse schooling environments. For example, Saatcioglu (2010) found that, over three decades, attending diverse schools decreased the risk of dropping out of high school for White, Black, and Hispanic students. Wells and colleagues (2009) found that decades after attending desegregated schools, Black, White, and Hispanic graduates valued their experiences—despite challenges and instances of discrimination—and felt they had learned important lessons about how to live in a multicultural society that they could not have acquired any other way. This finding echoes research showing that diverse schools may yield benefits for society at large, not just for individual students, including higher levels of social cohesion (Mickelson & Nkomo, 2012; Mikulyuk & Braddock, 2018). As the United States becomes more diverse—and more politically polarized, particularly around issues of race, immigration status, and poverty (Pew Research Center, 2019)—it is worth exploring a policy mechanism that evidence suggests may promote social cohesion.
However, diversity alone, either as a goal or as a metric, is often not enough to ensure equity for historically marginalized students. Alongside research on the benefits of school diversity, there is also evidence that historically marginalized students can face discrimination, resegregation through tracking, and internal opportunity gaps in integrated schools (Emerick, 2021; Lewis & Diamond, 2015; Noguera, 2001; Tyson, 2011; Wells & Crain, 1997). When schools fail to adapt policies, practices, and structures to promote equitable opportunities and outcomes for historically marginalized students, it can create schools that are diverse without being equitably integrated (Powell, 2002). While schools may espouse support for diversity and/or equity, there remain critical, structural ways in which schools may continue to reify existing inequities, often constricting opportunities for students of color (Lewis & Diamond, 2015).
In the United States, traditional public school enrollment is tied to students’ residence, so when neighborhoods are segregated, as is the case in much of the United States, traditional public schools are also segregated (Frankenberg, 2013). There is a long history of using public choice schools, such as magnet schools, to break the link between residential and school segregation (Cookson et al., 2018). Some have argued that charter schools are also well-positioned to combat school segregation by decoupling school enrollment from residence (Archbald, 2004; Frankenberg & Lee, 2003; Welner, 2013). However, in practice, charter schools are more likely to be racially and socioeconomically segregated than traditional public schools (Bifulco & Ladd, 2007; Frankenberg, 2003; Miron et al., 2010; Whitehurst et al., 2016) and can increase segregation among traditional public schools (Alcaino & Jennings, 2020; Clotfelter et al., 2019; Monarrez et al., 2022; Wells et al., 2019).
Charter schools’ practices are one driver of segregation in the charter sector. Charter schools are more likely to be located in neighborhoods with high proportions of Black and Latino/Hispanic 1 residents, which drives higher rates of enrollment among those groups and contributes to segregation (Henig & MacDonald, 2002; Lubienski et al., 2009). Charter schools’ marketing, recruitment, and enrollment practices can also increase segregation by influencing which parents learn about a given charter school, shaping application and enrollment patterns (Jabbar, 2016; Jennings, 2010; Mommandi & Welner, 2018). Segregative marketing and recruitment practices tend to target particular student groups or deter students of particular races, classes, and abilities from applying to charter schools (Lubienski, 2007; Wells, et al., 2019), including students with disabilities and students classified as English learners (Estes, 2004; Lacireno-Paquet et al., 2002). Many charter schools have a mission to serve “at-risk” youth and focus on enrolling low-income students of color in urban areas, which further contributes to segregation (Eckes et al., 2011; Welner, 2013). While some segregated charter schools represent culturally affirming spaces for historically marginalized students (Fox & Buchanan, 2014; Wilson, 2016), other charter schools with a mission to serve “at-risk” or other historically marginalized students have implemented this focus in damaging ways (Golann, 2015). For example, Hernández (2022) found that a charter management organization that aimed to serve historically marginalized students used race-evasive, deficit-based messages about racially minoritized students to secure funding and support, in effect perpetuating racism in their community.
There is also evidence that families’ preferences are drivers of charter school segregation. Family enrollment decisions may be linked to racial bias, particularly among White families (Billingham & Hunt, 2016; Haderlein, 2021; Hailey, 2021; Johnson & Shapiro, 2003; Ladd & Turaeva, 2020). For example, Ladd and Turaeva’s (2020) study of North Carolina families who transferred to charter schools found that White families’ preferences for schools with high shares of White students led to increased racial segregation. However, families’ segregative choices can also be explained by factors such as residential segregation and programmatic offerings (Bifulco & Ladd, 2007; Garcia, 2008; Jacobs, 2013). For example, Jacobs (2013) found that Washington, DC, parents’ strong preference for nearby schools led to a reproduction of residential segregation within charter schools along racial, socioeconomic, and linguistic lines. Additionally, Garcia (2008) found that specialized college preparatory charter schools were more likely to be racially segregated than other types of charter schools or traditional public schools, illustrating how parental preference for programmatic offerings may drive segregation. Important work has also highlighted how Black families have used school choice to advocate for strengthened educational services and as an act of political resistance, often framed in response to the failures of the public school system to provide equitable opportunities for Black students (Cooper, 2005; Stulberg, 2008). However, informational asymmetries, burdensome application processes, and a lack of responsiveness can constrict the sense of agency that Black families have when navigating the school choice marketplace (Pattillo, 2015).
While families’ and schools’ practices can contribute to school segregation, there is evidence that public school choice programs are a promising strategy for establishing diverse schools. In a study of four intentionally diverse public schools, three of which were choice schools, Siegel-Hawley (2020) identified a number of practices that supported school diversity, including setting diversity goals, using lottery-based (as opposed to competitive) admissions processes, having a racially diverse faculty, engaging in extensive targeted outreach, and publicly communicating the value of diversity. Other studies focused on diverse charter schools have found that many use similar practices to promote diversity (Jabbar & Wilson, 2018; Villavicencio, 2016). Other charter school practices that support diversity include forging connections with community organizations, using multilingual messaging, and providing culturally responsive information to families (Garcia & Morales, 2016; Villavicencio, 2016). At the same time, some diverse districts and schools have used messaging and made programmatic changes that cater to White families, in effect commodifying families of color and/or linguistically diverse families rather than centering the needs and experiences of these families (Cucchiara, 2013; Palmer, 2010; Posey-Maddox, 2014; Turner, 2020; Woody, 2020). Here, as with school diversity programs more generally, recruitment practices to support diversity have been implemented in inequitable ways. These conflicting characterizations of schools’ practices suggest that more evidence is needed about how schools can equitably recruit and enroll diverse students.
In part as a response to growing segregation in the charter school sector (Wohlstetter et al., 2017), there has been a rapid rise in the number of IDCS—charter schools that have a mission to serve students from diverse backgrounds (Potter & Quick, 2018). IDCS represent an important context to learn about the affordances and challenges of intentional diversity in public schools of choice. Yet, how IDCS operate—including specific practices used by IDCS educators to attract a pool of diverse applicants—is relatively understudied. Jabbar and Wilson’s (2018) qualitative study of recruitment practices at 13 IDCS in Minnesota and Louisiana is an exception. They found that IDCS engaged in intentional recruitment practices, with specific attention toward recruiting lower income families. IDCS educators saw recruitment efforts as a critical mechanism to enact their diversity mission, though many schools struggled to meet student diversity goals. Within these recruitment efforts, staff members worked with the community, sought to open schools in locations where they could attract a diverse student body, and tapped into social networks. At the same time, state and local policies designed to prevent exclusionary enrollment practices created barriers to diversity (Jabbar & Wilson, 2018), paralleling other evidence that charter school leaders had to flexibly interpret state policies in order to meet diversity goals (Villavicencio, 2016). Our study builds on this work by focusing on different states, examining local context and providing evidence on diversity outcomes.
The present study comes from other recent work where we explored a range of IDCS outcomes and practices. We found that IDCS were more diverse and had lower chronic absenteeism and suspension rates than comparison schools. Performance outcomes were somewhat more mixed: IDCS had better English language arts outcomes in four out of five jurisdictions, with worse outcomes in New York City, and no differences in math scores. This work also highlighted approaches that were integral to the mission and operation of IDCS, including community building, restorative justice, and culturally responsive pedagogy (Wohlstetter et al., 2021). The current study extends these findings by exploring how sociopolitical context shaped diversity outcomes within these schools and how schools exercised their agency to recruit and enroll a diverse student population.
Theoretical Framework
Parents—a term that we use here to refer to all student caregivers—are key agents in determining school enrollment patterns. As such, understanding how school leaders’ actions influence parents’ navigation of school choice is critical to understanding school enrollment patterns. Dominant school choice discourses position parents as rational actors who gather information about potential schools, evaluate that information, and effectively choose the highest performing options for their children (Blast & Walberg, 2004; Chubb & Moe, 1990). Though there is evidence that parents of all racial and socioeconomic backgrounds consider perceived academic quality when making school choice decisions (Glazerman & Dotter, 2017; Kleitz et al., 2000), evidence is also growing that nonacademic factors can influence school choice decisions as much as, if not more than, academic considerations (Billingham & Hunt, 2016; Garcia, 2008; Harris & Larsen, 2014; Jacobs, 2013; Ladd & Turaeva, 2020). In particular, many marginalized students—including students of color, linguistic or ethnic minority students, and LGBTQ+ students—and their families consider supportive social environments that reduce the risk of discrimination to be a top priority in their school choice decisions (Cooper, 2005; Fox & Buchanan, 2014; Posey-Maddox et al., 2021; Wilson, 2016). Further, many families—especially Black families—have reported experiencing barriers that functionally limit their ability to access preferred, or otherwise higher quality, schools of choice (Ellison & Aloe, 2019; Pattillo, 2015). As evidence mounts that parents’ school choice decisions are complex, socially situated negotiations influenced by a variety of factors—including families’ preferences, information access, and school leaders’ practices (Cucchiara & Horvat, 2014; Jabbar & Lenhoff, 2019; Lareau & Goyette, 2014; Villavicencio, 2013)—there is a need to move beyond overly simplistic rational choice models that do not take various inputs and context into account.
To understand how IDCS practices relate to school diversity, we draw on and expand C. A. Bell’s (2009) application of a choice set framework. This framework draws on bounded rationality theory, which acknowledges that people’s decisions differ because their decision-making processes are influenced by their knowledge and experiences (Simon, 1990). In this view, when families choose a school, they do not evaluate every option available to them. Instead, families subjectively consider a variety of priorities and information sources to create and select from a smaller choice set. C. A. Bell (2009) found that parents primarily learned about schools through word of mouth, largely from friends or other parents in their community, which ultimately produced stratified choice sets that could exacerbate school segregation.
According to C. A. Bell (2009), “by considering the content of the choice set and the processes used to create it, we are better able to understand how parents’ agency and market structures interact” (p. 193). The present study extends this framework by examining interactions between schools’ agency, context, and market structures. Specifically, we explore how administrators in schools with a diversity mission attempt to influence parents’ choice set construction in pursuit of enrolling a diverse student population. In the context of our study, we consider an IDCS to be part of a parents’ choice set if they decide to enter that school’s enrollment lottery. We consider how schools’ context influences the recruitment and enrollment practices that IDCS use to influence parents' choice sets and, in turn, schools’ diversity. This is outlined in Figure 1, which maps the pathway from IDCS context, to practices, to choice sets. While we acknowledge that parents’ social context and IDCS operational context are often related—for example, residential segregation influences both—this study seeks to clarify schools’ experience of and influence on the school choice process; so, for the purposes of this study, we will consider schools’ context distinctly. In doing so, we hope to understand the nature of schools’ agency in a school choice system and whether this agency can be used effectively in pursuit of a diversity mission.
Theoretical model.
Defining Diversity
In this study, we define student diversity using theory and our sample schools’ conceptualizations of diversity. We ground our exploration into diversification efforts in a definition of diversity that is focused on school-level enrollment, operationalized through variation in enrollment patterns by student characteristics such as race/ethnicity, gender, socioeconomic status, language, and more. We recognize that this definition of diversity is limited in depth, as internal integration and equitable intergroup contact, not merely numeric representation, are the ultimate goal of diversity (Allport, 1954; Leong, 2013; Perry, 2004; Powell, 2005)—particularly given the previously discussed internal segregation present in many schools that have diverse student populations (Lucas & Berends, 2007; Wells & Crain, 1997). However, we do not have access to data that would suggest the degree to which sample schools are truly integrated. As such, any conclusions about integration efforts are beyond the scope of this study. Instead, in this study, any references to diversity only refer to enrollment, which is a prerequisite for more holistic versions of integration and so remains worthy of study.
As will be discussed further, in constructing quantitative measures of diversity, we adopt an emic perspective and focus on the aspects of diversity most salient to our sample schools, specifically, race/ethnicity 2 and socioeconomic status. Although interviewees discussed other dimensions of diversity, such as linguistic diversity, gender identity, and disability, IDCS did not set formal goals around these criteria. While these are valuable components of a diverse learning environment, we prioritized only those aspects of diversity that sample schools included in their goals. Thus, we analyzed schools’ diversity in terms of racioethnic diversity and socioeconomic diversity. Additionally, across sample IDCS, there was not a single agreed-on definition of student diversity, and many IDCS pursued multiple goals. In recognition of this, we drew on multiple measures of diversity.
Using sample schools’ multifaceted framing of student diversity as a starting point, we measured diversity in three ways. In the case of research on school segregation, dominant discourses and researcher approaches tend to rely on a Black–White binary (Nieto, 2004). As Nieto (2004) explained, “the Black-White binary is not only incorrect but . . . it helps to obscure the common struggles of African Americans, Latinos, and others for equal education” (p. 22). Our sample schools did not rely on a Black–White binary. Instead, schools’ diversity goals were context responsive, which meant that the way they defined diversification of their population was often a response to their geography, existing enrollments, and local population. So, for some schools, pursuing racioethnic diversity meant recruiting across all racioethnic categorizations, while other times, schools were focused specifically on shifting their student population toward or away from a certain group. Accordingly, the term “diverse” is not a euphemism for a particular student group and we recognize that IDCS can define and pursue diversity in multiple ways. As discussed in the Data and Methods section, we quantitatively mirrored this orientation by using a diversity index that gives equal weight to students of various racioethnic groups. As further measures of diversity, we considered schools’ racioethnic and socioeconomic data separately to examine whether schools met locally determined diversity goals focused on racioethnic and socioeconomic diversity. Although these categories represent interrelated social constructs, they have unique impacts on students’ lived experiences, and evidence indicates that efforts to establish diversity based on a single aspect of diversity does not necessarily produce other types of diversity (Linn & Welner, 2007; Reardon et al., 2006).
Data and Methods
This study, which is drawn from a larger study of IDCS (Wohlstetter et al., 2021), used a convergent mixed-methods design (Creswell & Plano Clark, 2018; Ponce & Pagan-Maldonado, 2015). We collected and analyzed qualitative and quantitative data independently, then integrated both types of data to inform our conclusions (Creswell, 2013). Including qualitative and quantitative data allowed us to develop a more robust understanding than would have been possible with either type of data alone, as well as to triangulate and extend findings that emerged from each data source (Creswell & Plano Clark, 2018; Fetters et al., 2013; Greene et al., 1989; Leavy, 2017).
Schools’ inclusion was based on three criteria. First, schools had to have been in operation for at least three years as of 2016–2017 to ensure that they were relatively well established at the time of data collection. Second, sample schools had to be members of the Diverse Charter Schools Coalition (DCSC). To be a member of the DCSC, schools must be a nonprofit, brick-and-mortar charter school that meets a set of diversity standards. These standards include (1) having diversity among students, staff, and board members; (2) having diversity missions, goals, and strategies that focus on equity; and (3) being inclusive regarding socioeconomic, racial, ethnic, language, gender, and disability groups. Schools apply to join the coalition, and, once accepted, have access to programming focused on promoting diversity, including professional development and networking opportunities. Though coalition members pay sliding-scale dues, the majority of the coalition’s programs are funded by philanthropy, providing an incentive for schools to participate (Diverse Charter Schools Coalition, n.d.). Because schools that join the DCSC dedicate time and resources to participate in its programming, they represent a sample that has shown high levels of institutional commitment to diversity; this makes them an opportune site to explore the practices of choice schools that are intentionally pursuing diverse enrollments. Finally, to be included in the sample, schools needed to have an explicit commitment to diversity in their mission statement. While our inclusion criteria ensured that all sample IDCS had an explicit commitment to diversity, having a mission to serve a diverse student population is not a guarantee of meeting this goal. By focusing on the practices and outcomes of schools that have a mission to serve a diverse student population, we can better understand how specialist mission statements are operationalized through policy and practice.
Here, we describe the sample of schools in our study, first describing the full quantitative sample and then describing a subset of schools from the quantitative sample that comprised the qualitative sample. Appendix A includes all schools that were part of the analytic sample, denoted by whether they were in the qualitative and/or quantitative analytic sample, and their pseudonym.
Quantitative Data
Our quantitative sample includes 47 IDCS campuses managed by 12 operators. Eight charter management organizations (CMOs) operated 33 campuses, and four independent operators oversaw 14 campuses. Thirty campuses were in California, 12 in Colorado, and five in New York City. School districts in New York and Colorado provided student-level records for students enrolled in IDCS. For sample schools in California, we used publicly available demographic data from the Common Core of Data.
The quantitative analytic sample also includes 6,126 comparison schools. The definition of comparison schools varied by jurisdiction. In Colorado, comparison schools were all schools attended by IDCS applicants not admitted to an IDCS sample school. In California, comparison schools were those that served similar grades within a five-mile radius of each sample school. In New York, comparison schools were those attended by a matched group of traditional public school students.
Qualitative Data
As is common in a mixed-methods research design, qualitative participants were recruited from within the quantitative sample (Creswell & Plano Clark, 2018). Qualitative IDCS were purposefully sampled from the quantitative sample to produce maximum variation in terms of geography, governance structure, years of operation, and size (Patton, 2002). The final qualitative analytic sample included 27 IDCS schools managed by 12 unique operators. 3 Of these, 21 school campuses were operated by eight CMOs and six schools were operated by four independent operators. Seventeen campuses were in California, three in Colorado, and seven in New York City. Qualitative data collection was conducted in two phases. During the first phase, we interviewed school and network leaders (n = 70). In the second phase, a team of researchers visited each school to conduct 101 separate interviews with school principals and leadership team members and 40 focus groups with teachers and teacher leaders. For CMOs, we visited at least three school campuses operated under the CMO. For independent operators, the number of sites visited depended on the number of schools they operated. Interviews and focus groups followed semi-structured protocols (Patton, 2002) developed by the research team. Interview questions were tailored to the participants’ role in the organization and focused on approaches, challenges, and strategies for creating and operationalizing intentionally diverse charter schools. Interviews and focus groups were roughly 30–60 minutes.
Survey Data
We supplemented our analyses with data from parent surveys administered in 22 of the 27 IDCS campuses in the qualitative sample (one operator did not administer the survey). Surveys were sent to all parents of students enrolled in fourth, eighth, and 11th grades at these campuses. The approximate parent response rate was 20%, resulting in 356 respondents. Table 1 presents descriptive statistics for parent survey respondents, as compared with descriptive statistics for the students attending schools in the survey sample. The percentage of parent respondents was low and not representative of the full sample. In particular, parents of Latino/Hispanic students and parents of students who qualified for free/reduced price lunch (FRL) are underrepresented and parents of White students are overrepresented. We interpreted parent survey responses with caution, using them as qualitative data to triangulate and support our findings.
Descriptive Statistics for Parent Survey Respondents and Full Student Sample.
Note. FRL = free/reduced price lunch. Parent sample descriptives report the demographics of the students whose parents responded to the survey. The student sample is the full quantitative sample. All differences were statistically significant (p<.000).
In the survey, we asked parents to report on a range of topics regarding the school’s approach to diversity and inclusion. For this study, we drew on open-ended comments and two close-ended questions. The first question asked which factor was most important in selecting a school: location, school mission, academic performance, safe environment, or other, with the option to explain an “other” response. The second question asked respondents to rate their agreement with the statement, “Integration benefits all students.” We used open-ended explanations and comments submitted through the survey as qualitative data.
Quantitative Analyses
We drew on quantitative data to answer Research Question 3: Are IDCS more diverse than comparison schools, and are they meeting locally determined diversity goals? As previously discussed, diversity can be measured in many ways. Accordingly, our analysis employed three quantitative measures of diversity. First, we employed a theoretically informed, researcher-created racioethnic diversity index. The diversity index is a modified Herfindahl index, which, when calculated with student race/ethnicity as the focal characteristic, can be interpreted as the probability that any two students chosen at random from a school will be of a different race/ethnicity. Thus, higher diversity index values indicate more racial diversity. We drew on data from the 2016–2017 school year, calculating a racial diversity index value for each sample and comparison school. Then, we averaged the racial diversity index values across sample and comparison schools by state and compared these values using t-tests.
Second, to employ an emic perspective, we used two diversity measures from qualitative sample schools’ diversity goals. We assessed whether the qualitative sample of IDCS—from which these goals were derived—met the socioeconomic diversity goal of having 40%–50% of enrolled students qualify for FRL. We also evaluated whether schools met the racioethnic diversity goal of having no racial majority group among enrolled students. Given this goal, sample schools were considered racioethnically diverse if no single racioethnic group represented more than 50% of enrolled students.
Qualitative Analyses
All interviews were recorded, transcribed, and coded with a mix of deductive and inductive approaches. We collaborated with research team members from the larger project from which this study was drawn to create an initial codebook a priori, informed by the interview protocol and our study goals, which was inductively expanded in response to emergent themes (Miles et al., 2014). This set of codes aimed to capture key study constructs, including recruitment practices, policy context, and perceptions of diversity, along with other topics not relevant to the present study, such as instructional approaches and discipline practices. A team of six researchers participated in the first round of deductive coding and engaged in multiple coding pilots to establish interrater reliability at or above 70% (Charmaz, 2006). If interrater reliability on a pilot transcript was below 70%, group members discussed discrepancies until disagreements were resolved and then revised the codebook as necessary, after which all researchers engaged in a subsequent round of pilot coding. Once interrater reliability was established, all transcripts were coded using the a priori codes. After the first round of coding was completed by the full research team, the first and second author applied inductively created codes to identify practices, policies, and outcomes specifically related to recruiting and enrolling diverse students. Drawing on coded excerpts, the same two researchers wrote analytic memos for each operator in the qualitative sample to clarify emergent patterns in the data and draw connections between themes (Saldaña, 2015).
Survey Analyses
To provide a deeper understanding of Research Question 1 (What contextual factors influence diversification efforts at IDCS?) and Research Question 2a (How do IDCS communicate their diversity mission during recruitment and enrollment?), parent survey data were converged with qualitative data (Creswell & Plano Clark, 2018). For each of the two survey questions included in our analyses, we calculated descriptive statistics, disaggregated by the race/ethnicity of the student whose parent or guardian was responding. For the Likert-type item that ranged from strongly disagree to strongly agree, we calculated the percentage of respondents who selected each option. Open-ended survey responses were coded using the qualitative phase codebook. We used survey responses to triangulate and provide examples or unique insights from parents, not as stand-alone representative findings.
Mixed-Methods Analyses
Convergent mixed-methods studies require multiple strategies to integrate qualitative and quantitative data (Creswell & Plano Clark, 2018). First, we employed iterative mixed-methods integration, in which qualitative findings were used to develop and refine quantitative approaches (Moseholm & Fetters, 2017). We coded interviews to establish how qualitative sample schools defined diversity in their goals and mission, then used this information to construct two outcomes of diversity to use in our quantitative analysis. Second, our interpretation relied on triangulation to explore how quantitative, survey, and qualitative data confirmed, contrasted, explained, or enhanced one another (Creswell & Plano Clark, 2018; Greene et al., 1989). We drew on this analysis to answer our overarching research question: How do context and school practices shape diversity at IDCS?
Findings
In this section, we present our findings, ordered by research question. We first identify contextual factors that influenced IDCS diversity efforts. Second, we describe how IDCS used messaging, recruitment, and enrollment practices to pursue diverse enrollments. Finally, using multiple measures of diversity, we provide evidence to suggest that IDCS are more racially and socioeconomically diverse than comparison schools and vary in whether or not they are meeting locally defined diversity goals.
What Contextual Factors Influence Diversification Efforts at IDCS?
Across interviews and focus groups, IDCS personnel described how complex contextual factors influenced their ability to establish diverse schools. The two most prevalent factors discussed by interviewees were local housing patterns, including segregation and gentrification, and state policy.
Housing Patterns: Segregation and Gentrification
Interviewees from all three states described local housing patterns—specifically segregation and gentrification—as barriers to sample schools’ diversification efforts, with some interviewees describing strategic responses to these barriers. Interviewees from nine operators across all three states identified housing segregation as impacting student diversity. According to one Colorado school leader, the school was “a more White and more affluent school starting out, which reflects [this neighborhood]. This is one of the wealthiest and more segregated neighborhoods in the city.” By opening in a segregated neighborhood, the school had low levels of diversity from its inception; thus, school staff had to work against local housing patterns by actively recruiting students from outside the neighborhood.
Housing and neighborhood segregation can cause other challenges to diversity. For example, IDCS educators described how segregated housing patterns led to longer commutes for certain students and families, which they perceived as deterring some families from enrolling. According to the regional executive director of a Southern California CMO, [Los Angeles], like many urban developments, is a highly segregated city . . . and when you’re trying to build a diverse school community, that becomes in some ways even more of an issue because you need parents in a segregated city to traverse across communities to get to you, and so where the school is physically located is very important.
A New York parent echoed these concerns in an open-ended survey response, calling for the school to support diversity through “better transportation opportunities” because “a lot of children of color live in underserved neighborhoods, and these underserved neighborhoods are not close to the school.” Here, we see that housing segregation intersects with transportation issues to create a barrier to student diversity.
Residential segregation also influenced how information about the schools was circulated through parent social networks. In particular, IDCS leaders whose schools were located in a neighborhood with wealthier and White families reported that those parents tended to share information about the school among their relatively homogeneous social circles. This pattern of information sharing led to Whiter and wealthier applicant pools. For example, a California IDCS administrator described how locating their school in a segregated area of the city impacted the ability of families of color to hear about the school: We had a lot of middle-class White families throw their names in the lottery. Especially because of our location, because now the families who are closer to the city, those mostly of color, it was harder to hear about us and, if they did, the commute was hard. . . . What we said is all of a sudden, “Oh my God, our school is more White than what we were aiming for.” That was just, wow. That was hard for us, because it wasn’t our main goal. I would say in the past five years, every year is trying to figure out “how do we change that?”
These quotes highlight how residential segregation can create a barrier to school diversification efforts, especially when coupled with segregated social networks and a lack of transportation options. We also see how residential segregation drove IDCS to use recruitment practices to counteract this barrier. Educators from some schools did attempt to strategically respond to segregation, largely through intentionally selecting school sites in areas where there was less residential segregation or where multiple segregated neighborhoods intersected. One school leader said that their schools’ location at the intersection of various segregated neighborhoods allowed them to “naturally” meet their diversity goals.
Interviewees from five IDCS operators in New York and Southern California also identified gentrification as a barrier to meeting diversity goals. In many cases, educators who discussed gentrification worked at IDCS that had carefully chosen their locations in response to existing patterns of residential segregation. However, gentrification and the resulting changes in the neighborhood demographics altered the recruitment and enrollment assumptions made by school founders when they were choosing a school site. Operators in New York and Southern California discussed this as a major barrier to meeting their diversity mission. An IDCS director of operations explained, Gentrification is killing us right now in Brooklyn . . . we strategically chose our campus locations for communities that would be able to source students and families from all types of backgrounds. However, over the 10 years that [the school] has existed, . . . the neighborhood has just significantly changed, and so, we do have challenges.
School staff in Southern California were also pessimistic about the influence of gentrification on their diversity mission. One school’s director of community engagement noted, “I think the reality of the housing market in Los Angeles, and the affordable housing crisis, [is] going to continue to make it hard to have intentionally diverse and integrated schools. And so, we are sort of scraping by.” Taken together, our evidence suggests that although some theorized that charter schools might increase diversity through decoupling schools from residential segregation (Archbald, 2004; Eckes & Trotter, 2007; Frankenberg & Lee, 2003; Welner, 2013), housing dynamics continue to place considerable pressure on schools’ ability to recruit and enroll diverse students, particularly given inadequate public transportation access in many communities.
State Policy: Lottery, Enrollment Targets, and Facilities
Across interviews, participants explicitly linked state policies to their ability to recruit and enroll a diverse student body. Although there were some exceptions, these state policies were typically described as constraining the ability of school staff to meet diversity goals. Lottery policies were the most discussed state policy. At the time of the study, all study jurisdictions permitted lottery preferences for students based on socioeconomic status, geographic location, or English learner status, although the specifics of these preferences varied by state. While all operators in our sample used some combination of lottery weights or preferences, interviewees from different IDCS characterized lottery policy in divergent terms. Some interviewees felt that policies permitting weighted lotteries supported their diversity mission; others viewed the legislative prohibition of race- or ethnicity-based lottery preferences as limiting their school’s ability to achieve diversity goals.
Interviewees from seven operators described how lottery weights based on socioeconomic status or a student’s address supported school diversity because they used these weights as proxies for race and ethnicity. For example, one interviewee explained, “Because of California legislation, we cannot select applicants based on ethnicity. We get around that by selecting applicants based on where they live, which is a proxy, basically . . . for your socioeconomic status and your ethnicity.” In line with prior research indicating that intentionally diverse schools interpret and implement race-neutral state policies in ways that support their diversity mission (Jabbar & Wilson, 2018; Villavicencio, 2016), this response illustrates how interviewees navigated barriers created by race-neutral lottery policy. In contrast, staff from four operators that used socioeconomic lottery weights as a proxy for race described this strategy as ineffective. For example, interviewees from a New York school expressed frustration with the attempted workaround, noting, [When using socioeconomic lottery weights], we can end up with a lot of poor White kids, and we can end up with a whole school of White kids that are the full range of economic status, and still have to follow all the laws, which isn’t our [diversity] goal.
A California participant noted that their school’s geographic lottery preference for neighborhood residents began to work against the school’s diversity goals: White and wealthy families began to move into the neighborhood to take advantage of the lottery preference, illustrating how gentrification and lottery policy intersected to hinder diversity.
Leaders at New York IDCS also found enrollment target policy to be in conflict with their diversity mission. In our sample, New York law was unique in that it required charter schools to set enrollment targets that reflected the demographics of traditional public schools in their community school district.
4
Because of the prevalence of gentrification and segregation in New York City, interviewees from sample schools reported that this policy and housing context interacted in ways that created barriers to diversification. One interviewee explained, One of the conundrums in New York is that the state charter law has language that says . . . your charter school should really come close to or mirror the demographics of the district in which it's located. So, you can't create a diverse charter school in . . . any of the community school districts that, themselves, are drawn along the lines of residential segregation and based on historical redlining.
This IDCS set a goal for its schools to demographically match the city, rather than their neighborhood district, which put their diversity goals in conflict with state policy. Here, we can see that, like lottery policy, other state policy decisions may have unintended consequences for school diversity.
Interviewees from California IDCS also reported that two state facilities policies—Proposition 39 and Senate Bill (SB) 740—impacted their ability to pursue a diversity mission. Proposition 39 requires that school districts provide facilities to charter schools that serve students residing in that district. However, the local district chooses the facility and is not required to offer it for more than one year at a time, which led participants to characterize the policy in conflicting terms. One IDCS staff member said, Prop 39 can be beautiful . . . right up until Prop 39 is extremely volatile, political, and only a one-year arrangement, and so [at one of our campuses], as an example, we’ve been in, I wanna say, eight facilities over eight years, which is just brutal.
Participants from this school went on to describe how school enrollment and diversity levels fluctuated with each new location. Overall, while participants saw the facilities provided by Proposition 39 as a crucial financial support, they were critical of its implementation and the impacts it had on their diversity efforts. Another participant mentioned that they turned to SB 740 funding after “inadequate” Proposition 39 facilities created a “challenging” relationship with their local district. SB 740 provides funding to charter schools to purchase facilities in neighborhoods with high shares of traditional public school students who qualify for FRL. One IDCS leader cited SB 740 as one of the only policies that helped them meet diversity goals, because it helped them permanently locate the school at a site that supported diverse enrollments. Here, both facilities policies influenced where schools were located, which, because of previously discussed housing patterns, influenced schools’ diversity.
How Do IDCS Pursue Diverse Enrollments?
As we have seen, IDCS work within complex contexts. Housing patterns and different state policies are powerful forces; ignoring or failing to strategically respond to these factors can have consequences for schools’ diversity. We found that, in response to these contextual factors, staff at IDCS developed sophisticated recruitment and enrollment practices designed to influence parents’ choice sets and, in turn, support diversity. Recruitment practices encouraged parents to enter a school’s lottery, while enrollment practices attempted to influence whether parents whose children had been offered a spot would enroll. Here, we first discuss how IDCS staff communicated their diversity mission as part of their recruitment and enrollment efforts. Then, we detail recruitment and enrollment practices that interviewees described as supporting diversity.
How Do IDCS Communicate Their Diversity Mission During Recruitment and Enrollment?
To influence the choice sets of different target families, IDCS incorporated messaging about their school’s diversity mission throughout recruitment and enrollment phases. Messaging was intentional and differentiated for two types of parents: (1) those who already supported the school’s diversity mission and (2) parents who were unsure about intentional diversity. Messaging to parents who already supported the school’s diversity mission aimed to raise awareness through mass recruitment practices, such as attending citywide events or making direct contact through mailers or social media. These outreach practices ensured that families who were already invested in diversity added the sample school to their choice set. Speaking about parents at a Colorado school that initially struggled to meet its diversity goals, an administrator explained that parents “wanted a diverse school. They didn’t necessarily get it that first year, but that’s what they wanted.”
In contrast, some interviewees reported that many families required convincing about the benefits of attending a diverse school. Many of these parents expressed concerns that an IDCS’ diverse environment—which, in most cases, was quite different from the educational environment those parents experienced—would disadvantage their children in some way. For example, a California participant explained that hesitant parents often said things such as, “My uncle didn’t go there, my aunt didn’t go there, my family hasn’t gone there, I don’t know what that is.” This type of unfamiliarity with diverse schools made messaging about diversity a critical strategy for IDCS leaders.
For hesitant parents, IDCS staff differentiated their messaging to address parents’ concerns and, in doing so, attempted to build those parents’ commitment to diversity. Interviewees felt that families’ concerns varied based on parental background. This led schools to develop multiple ways to communicate their diversity mission and its benefits to parents. Participants explained that White, middle-class or wealthy families were often deciding between their school and magnet or private schools with higher funding levels and a greater breadth of academic and extracurricular programming. This perception was paralleled by open-ended survey responses from parents of White-identifying IDCS students, indicating that they had been considering their IDCS alongside many other schooling options. IDCS staff reported that hesitant families from these historically advantaged groups also expressed concerns about lower academic standards or narrowed programmatic offerings at charter schools. When messaging the benefits of diversity to these hesitant families, IDCS staff tended to emphasize social and cultural advantages that students gained by attending their schools. This strategy is exemplified by the framing offered by one NYC administrator: What I often say is, “If you send your child to [our school], they will have friends from all walks of life, and they will learn to collaborate for the 21st century. You could always buy a tutor for your child or give your children academic experiences outside of school quite easily, but you can't buy diverse friends. When you send your children to [our school], you know that they're going to have a better understanding of the world and therefore be more able to operate within it moving forward.”
This messaging inverts the deficit-based characterizations that tend to follow urban public schools attended by high proportions of low-income students of color (Valencia, 2010). Instead, this framing asks advantaged families to focus on the benefits of diverse contexts—a message that has been found to be appealing to many White families (Underhill, 2019).
Open-ended survey responses from White and/or wealthy parents echoed IDCS staff’s framing of school diversity as an asset. A White parent from Colorado explained that her son is experiencing “a rigorous academic program that is substantively enhanced by the diversity of the student body and the life lessons that he is learning in and out of the classroom.” This parent’s comments parallel IDCS leaders’ perception that for historically advantaged families, diversity differentiates IDCS from other schooling options. While we cannot know if this parent already held these views or developed them as a result of IDCS’s messaging, the close alignment of her description with the school’s mission suggests that this is a salient framing for some families. By messaging the exclusive value proposition of diversity, IDCS attempt to increase the likelihood that White and/or wealthy families will include their school—not just private or magnet schools—in their choice sets.
Hesitant families of color and linguistically diverse families, on the other hand, were seen as tending to express concern that their child would face marginalization at IDCS. Recruitment messaging for historically marginalized families tended to focus on building trust and demonstrating cultural responsiveness. To illustrate the importance of trust when messaging a diversity mission, one administrator described a conversation with a Latino/Hispanic family: She said, “It’s like sending my kid into the belly of the beast.” Driving into [a White wealthy neighborhood] is just this, “I’m not welcome here.” I think we had to do a lot of work to say, “That may actually be true in the community as a whole, let’s not pretend that it’s not true. It is uncomfortable, it is segregated, it is different. But within our school, things will be different.” I think we delivered on that, and I think her story was then told to friends, neighbors, family, one student at a time.
Here, we see that IDCS messaging directly acknowledges the negative experiences that many parents of color have experienced in diverse environments. By recognizing these lived experiences and ensuring cultural responsiveness, sample schools attempted to build trust among parents, which in turn increased the likelihood that IDCS would become part of their choice set. This trust was reflected in the open-ended survey response of a Black parent whose child attended a New York IDCS. They said that their IDCS “feels like home for students and parents. . . . [It is a] very good model of school integration.” Ultimately, we see that IDCS understood that families’ experiences with and attitudes toward diversity would vary based on their positionality and used explicit, differentiated messaging to attempt to address these varied concerns.
Among parents who responded to the survey, we see that IDCS’ diversity mission was influential. Almost all of the small subset of parent survey respondents (98.4%) indicated that they believed integration benefited all students (Table 2), with nearly half (48.43%) saying that the school’s mission was the most important factor in selecting the school for their students (see Table 3). That respondent parents, including parents of color, reported that they saw integration as beneficial and chose the school based on its mission suggests that IDCS are effective in messaging their diversity mission, at least among respondent families. Although this is not a representative sample, these data imply that for some families, an IDCS’ diversity mission is an influential factor in their school choice process; this makes messaging about the diversity mission a promising strategy that warrants further study with a larger, more representative sample of parents.
Parent Survey Responses to the Statement, “Integrating students from different backgrounds benefits all students.”.
Note. This table reports the share of respondents in each category. ~ Indicates a column where the percentage is rounded because of small sample size.
Parent Survey Responses to the Question, “Of the factors listed, which was most important in selecting this school for your student?”.
Note. This table reports the share of respondents who selected each factor. Columns may not add to 100 because of rounding.
How Do IDCS Recruitment and Enrollment Practices Support Diversity?
Recruitment Practices
Having developed nuanced ways to communicate their diversity mission, IDCS from all jurisdictions then used a variety of recruitment strategies during which staff shared these messages, in an attempt to recruit a diverse student population into the lottery pool. Recruitment practices included making direct contact with families, establishing feeder patterns, and using data to inform recruitment strategies.
Direct Contact
Many interviewees described how school staff made direct contact with prospective families through in-person events. Among the most common direct contact strategies was attending citywide or regional school fairs. At one independent New York IDCS, personnel explained that most students heard about their school “through our fairs. We attend the district fairs throughout September and October and so we send . . . family, staff and students to attract them. We have our table set up, and that’s how they tend to hear about us.” Other schools hosted community events at the school. In describing their recruitment practices, one California school administrator explained, Every year we do a community fair where it’s free for everyone. We invite the community to come here. It’s games and it’s food and it’s interacting with the staff and seeing the classrooms, and that’s been our most effective way to increase our visibility within diverse communities.
Here, large events that draw crowds of families, either at the school or elsewhere in the community, were a critical way for IDCS to gain exposure and provide information to families, effectively widening their choice sets.
Staff from other sample schools made direct contact with families through strategies such as advertising or canvassing. A New York school that tracked how enrolled families learned about the school found that most did so through social media, mailers, and canvassing. This mix of contact strategies was common: Half of sample IDCS operators reported using face-to-face canvassing, while more than 80% (10 of 12) reported contacting families through mail, phone calls, or online. In most cases, interviewees reported that direct contact recruitment materials were differentiated based on the target demographic. There was also a focus on having recruitment materials available in multiple languages and featuring images of students from diverse backgrounds.
Staff at all but one IDCS operator mentioned using family members of current students as recruiters in their direct contact practices. Familial involvement in recruiting could include having family members visit feeder schools to talk about the school, attend recruitment fairs, or conduct home visits with potential families. Staff from a Colorado school explained that these parent volunteers “drive a significant portion of that trust and relationship building” during recruitment activities. Many schools were intentional about asking diverse families to recruit among their own community. For example, staff at a New York school noted that when they were focusing on recruitment in “a Russian district, we may send more families that are Russian-speaking to help connect with those students.” One New York school took the practice further by involving current students in designing and implementing recruitment strategies. These practices parallel findings indicating that parental word of mouth and social networks are highly influential in school choice decisions (C. A. Bell, 2009). Parents and students are uniquely positioned to influence their peers and increase prospective families’ confidence in the school, in turn ensuring that a school makes it into those families’ choice sets.
A small subset of CMO-based IDCS had more extensive direct contact strategies that employed marketing software or online directories to manage these processes. One CMO received the email addresses of every local family from the traditional public school district. Armed with this information, CMO staff used marketing software to contact families in English and Spanish, the most prevalent home languages in the district. The software tracked family engagement with recruitment messages, and these data were used to plan follow-up recruitment actions. An interviewee from a New York CMO said that their school used software to get “very, very specific towards which people you’re targeting” online in ways that helped them meet diversity goals.
Ultimately, the extensive and targeted nature of these direct contact programs suggests that IDCS saw seeking out, contacting, and building trust with diverse families as a way to interrupt segregative contextual forces and help them reach their diversity goals.
Feeder Organizations
Staff from 11 of 12 IDCS operators discussed recruiting by establishing feeder organizations. Community organizations that IDCS sought to establish as feeders included schools, day care centers, community centers, camps, social services agencies, houses of worship, and nonprofits. Interviewees described how IDCS staff, teachers, or families would visit and build relationships with community members at the organizations from which they hoped to recruit students. According to an interviewee from a Colorado middle school, “We visit every fifth-grade classroom within about a five-mile radius with our teachers. They go in and they get to know teachers and their peers and try to build respect.” Similarly, staff from elementary schools described establishing feeder patterns with diverse pre-K operators.
Some IDCS developed their feeder organization strategy in response to local context. For example, staff from one IDCS facing gentrification and declining enrollment among students of color explained that it only engaged with organizations that helped counteract those forces: The only places we go are to community centers, day care centers, public housing, community rooms. We turn down any private preschools that want us to join and any forums that will primarily serve affluent families. Those families already know about us.
Here, in addition to using recruitment practices to respond to gentrification, we also see that IDCS are aware of the ways that existing word-of-mouth networks among affluent parents, who “already know about us,” can undermine diversity efforts. In effect, this IDCS is using feeder organizations to overcome segregative contextual pressures.
Building relationships with feeder organizations indirectly influences parents’ choice sets. In building a relationship with a feeder organization, schools are increasing awareness of their school among the staff, including teachers, leaders, and counselors. These staff members are often trusted members of diverse parents’ social networks. Further, feeder organization staff are often involved in helping students and parents navigate school transitions and applications. By building a positive reputation among staff at trusted feeder organizations, IDCS may be recommended as part of that process. Parental word of mouth within feeder schools and community organizations may also influence prospective parents’ choice sets. In this way, focusing on feeder organizations offers an indirect, yet effective, way to influence parents’ choice sets.
Data-Driven Decision-Making
IDCS did not use recruitment practices indiscriminately: The use of data to inform recruitment practices was itself a distinct tool to support diversification. All but one sample IDCS discussed using data to develop recruitment practices that were responsive to each school’s particular context. Most IDCS administrators reported using data to analyze current school demographics in order to set recruitment targets that would support the attainment of diversity goals. One California administrator explained that their recruitment strategy involved looking at what your numbers look like in your previous classes and then doing that targeted outreach where you might say . . . there’s a community here that’s underrepresented, so we are going to do the door hangers or we’re gonna go and see if we can get into some churches or community groups and see if we can invite people in because they might just not know about the school.
Here, the school administrator details how their school’s diversity data informed their strategy for influencing parents’ choice sets through targeted recruitment.
IDCS with multiple campuses also reported using data to develop tailored recruitment strategies based on each campus’s demographics and location. A Colorado CMO that set goals for the proportion of students who qualified for FRL explained that while some of their campuses were meeting diversity goals, others were not: We have some schools where their [low-income] population is closing in on 90% . . . we intentionally recruit students for those schools to make sure that we can keep this element of integration alive. That does feel weird sometimes because . . . what we’re really doing is targeting middle-class White families.
Here, we see that developing data-driven recruitment strategies pushes this IDCS to conduct targeted recruitment that supports their vision for diversity, even while doing so creates discomfort.
Unsurprisingly, data-driven strategies were more elaborate at CMOs with dedicated recruitment staff than at independent operators. One CMO shared that they evaluated the effectiveness of recruitment strategies, for example, by tracking which social media images generated the most engagement from target groups. Two New York CMOs explained that their schools monitored the demographic composition of lottery entrants so they could target additional recruitment activities toward groups that are underrepresented in the lottery pool. Active monitoring of lottery data allowed these schools to make real-time adjustments in recruitment strategies to support diversity. Though CMOs tended to have more robust data systems, both CMOs and independent IDCS used data to gauge the degree to which their recruitment activities were overcoming contextual barriers to diversity, and to adjust their recruitment and enrollment practices to influence parents’ choice sets.
Enrollment Practices
After spending considerable time and energy cultivating a diverse lottery pool via recruitment practices described earlier, sample IDCS shifted their focus to a set of practices that would increase the likelihood that a diverse set of students would actually enroll in their schools, despite contextual barriers to diversity. Specifically, sample schools continued to pursue diversity through weighted lottery schemes and outreach to admitted families.
Lottery Weights
All sample schools discussed using lottery weights to overcome contextual barriers to diversity. Though schools’ recruitment practices aimed to cultivate a diverse lottery pool, lottery weights were characterized as a critical tool for interrupting imbalanced lottery application patterns by assigning higher odds of winning the lottery to certain families. This was the case for a New York school, where a teacher felt, in their words, that “we can live our equity mission through this lottery.” Lottery weights were also frequently mentioned as a strategy to overcome segregative housing context, such as in the previously discussed zip-code-based lottery weights used by a California IDCS. Paralleling this strategy, a Colorado IDCS used socioeconomic lottery weights to overcome economic segregation in the neighborhood around their school. According to one administrator, Students who receive free and reduced lunch get three tickets in the lottery for every one ticket of a non-FRL student. That was a really tough one, we had to sacrifice our [federal grant] funding for a year, as that was not approved. . . . We worked through that and did a lot of lobbying at the local, national level to get this through. It’s been wonderful. We are in a neighborhood where I think the local elementary schools are about 5 to 10% FRL, we now will be 50% [FRL] this year. When we started, it was 27%.
Here, an IDCS used its lottery to prioritize its diversity mission over financial considerations and, in doing so, reported having greater socioeconomic diversity than neighboring schools.
Some schools, however, initially underestimated the role of lottery weights in supporting diversity, perhaps because of the prohibition of race or ethnicity based preferences. These schools came to realize that changing their lottery weights in response to local context could support a more diverse student population, even in the absence of racial weights. This was particularly common among schools in New York and Southern California, which also reported high levels of gentrification. For example, the executive director of a Los Angeles-area CMO explained, Being here in [a wealthy neighborhood] we've had declining diversity since year one. We started out at 35% free and reduced lunch, we're down to 20%. We've stabilized around 20 to 25% because we have a free and reduced lunch [lottery] preference. . . . In this past year, we had roughly 800 applicants, and I think 68 of them were free and reduced lunch. Now, because of our preference, they all got in at the top of the list.
A participant from a New York school described a similar experience of declining school-level diversity and how a geographically weighted lottery helped change that trend: All of a sudden, we were seeing that the children living in public housing right [near the school] were not getting into our lottery anymore. And I think one major mistake that I'm responsible for is not noticing that and paying attention to that early on. Once we did notice and start paying attention to what was happening, then we put a [geographic lottery preference in place] to protect that class diversity.
Here, we see that, despite policy barriers, weighted lotteries were used as a stopgap against the segregative influence of housing context, particularly in affluent or gentrifying neighborhoods.
Enrollment Outreach
Two parental decision points ultimately determine a charter school’s diversity: the decision to enter a lottery, and, if granted a lottery spot, the decision to enroll. A subset of IDCS also put considerable effort into influencing post-lottery enrollment decisions through direct contact and using parent volunteers. An administrator from a New York City school explained the specific actions they took to ensure that linguistically diverse families enrolled: When they receive an offer, we follow up with them with someone that speaks Spanish with them and share what’s going on in our school and share with them how great it is, leveraging our non-White parents to talk to other non-White parents and say this is actually a safe place, this is a culturally safe place to send your child, etc., and making them comfortable.
Similarly, an IDCS administrator in a gentrifying New York neighborhood reported that they gave the “white glove treatment” to low-income families who received a spot in the lottery. As these examples illustrate, enrollment outreach efforts tended to mirror techniques used during initial recruitment.
Looking across qualitative data, we see that intentional recruitment and enrollment practices are necessary to overcome contextual barriers to student diversity. Recruitment practices aimed to influence parental choice sets and induce parents to enter the lottery, while enrollment practices encouraged target parents to enroll their children. Across recruitment and enrollment practices, sample schools’ differentiated messaging about their diversity mission aimed to build trust and cement the schools’ position in parents’ choice sets.
Are IDCS More Diverse Than Comparison Schools and Are They Meeting Locally Determined Diversity Goals?
Next, we quantitatively explore the outcomes of sample IDCS’ recruitment and enrollment practices, using multiple measures of racioethnic and socioeconomic diversity. As previously discussed, we employ a researcher-created diversity index and two school-created measures: (a) if 40%–50% of students qualify for FRL, and (b) having no racial majority (i.e., no single racioethnic group represents more than 50% of enrolled students).
IDCS and Comparison Schools’ Diversity Outcomes
Using a modified Herfindahl index, which represents the probability that any two students chosen at random from a school will be of a different race or ethnicity, we found sample IDCS to be more diverse than comparison schools (see Table 4). Specifically, the average sample IDCS diversity index was 0.60 versus 0.47 among comparison schools. We also examined the racioethnic breakdown of students, noting that the greatest difference between IDCS and comparison schools was the share of White students, which was 13 percentage points higher among IDCS. The shares of Black students were lower in IDCS than comparison schools, while differences in enrollment share of Asian/Pacific Islander and Latino/Hispanic students were not significant. Given that students of color are increasingly attending racially isolated schools, particularly in the charter sector (Orfield et al., 2012), and that sample IDCS were in highly segregated districts, enrolling higher shares of White students and lower shares of students of color results in greater overall diversity at these IDCS. We also found that half of IDCS students qualified for FRL, compared with 69% of comparison school students. Though this is a large share, we interpret this value as consistent with socioeconomic diversity goals.
Diversity Index for Sample and Comparison Schools, 2016–2017.
Note. The diversity index is a modified Herfindahl index, which can be interpreted as the probability that any two students chosen at random from a school will be of a different race/ethnicity. Higher diversity index values indicate more racial diversity. The student race/ethnicity and free/reduced price lunch panels reports mean percentage of respective variables for sample and comparison schools in each jurisdiction for 2016–2017. To determine if differences between sample IDCS and comparison schools were statistically significant, t tests were performed. IDCS = intentionally diverse charter schools; Comp. = comparison school.
Comparison schools are schools attended by a matched group of traditional public school students.
Comparison schools are all schools attended by IDCS lottery applicants not admitted to an IDCS, including traditional public schools and charter schools.
Comparison schools are schools within a 5-mile radius of sample schools; includes students in Grades K–12. Enrollment share by subgroup equals more than 100 because of rounding error.
✢Includes American Indian and Alaska Native; two or more races.
p < .1. ** p < .05. *** p < .01.
However, these trends mask considerable variation when disaggregated by state. While the IDCS diversity index values in each state were similar to each other (and, in all three cases, higher than comparison schools), the data on subgroup enrollment shares reveal that the means to achieving greater diversity differed by state. In relation to comparison schools, New York IDCS enrolled slightly greater shares of students of other races (5% vs. 2%), nearly 3 times the share of White students, and about half the share of Latino/Hispanic students. New York IDCS also enrolled lower shares of students who qualify for FRL (48% vs. 74%), resulting in more socioeconomic diversity than comparison schools. Colorado IDCS enrolled higher shares of Asian/Pacific Islander students than comparison schools (5% vs. 2%), while differences for other subgroups were not statistically significant. However, small enrollment changes appear to have collectively created a higher degree of overall diversity: Colorado IDCS were significantly more diverse than comparison schools (.56 vs. .44). It is noteworthy that there was little difference in the share of students who qualified for FRL among Colorado IDCS (70%) and comparison schools (74%). We interpret this as an indication that Colorado IDCS were not successful in achieving socioeconomic diversity.
Among California IDCS, the share of Latino/Hispanic students enrolled in sample IDCS was 9 points lower than comparison schools (47% vs. 58%). Similar to New York IDCS, California IDCS enrolled much higher shares of White students (30% vs. 16%) and slightly higher shares of students of other races/ethnicities (9% vs. 5%) than comparison schools. Further paralleling New York IDCS, California sample schools enrolled much lower shares of students who qualified for FRL (43%) than comparison schools (63%), which we interpret as attaining socioeconomic diversity. These divergent patterns highlight that there is no single strategy for diversification—instead, IDCS in all three states attained similar overall levels of racioethnic diversity (as measured by the diversity index) with different demographic compositions that reflected the districts in which they were located.
IDCS Success in Meeting Locally Determined Diversity Goals
Many operators had diversity goals focused on the proportion of students who qualified for FRL, the most common being a goal of having between 40% and 50% of enrolled students qualify for FRL. As seen in Figure 2, only five out of 12 IDCS operators met this goal, with three operators enrolling greater than 50% FRL-eligible students, and four operators enrolling less than 40% FRL-eligible students. Four of the five operators who met this goal were California-based charters.
Average percent eligible for free/reduced price lunch by operator in 2016–2017.
Avoiding a racial majority among enrolled students was another commonly discussed diversity goal among IDCS. As seen in Figure 3, nine of 12 IDCS operators met this goal in 2016–2017. Among the three operators who did not meet this goal, Latino/Hispanic students comprised more than 50% of enrolled students. As with the diversity index, we also saw considerable variation at the local level in the distribution of students by race/ethnicity. Enrollment tended to be dominated by one or two demographic groups, with much lower enrollment of other groups. However, operators differed in whether White, Black, or Latino/Hispanic students represented the largest share (there were no sample IDCS operators where Asian students represented the greatest enrollment share). An exception to this pattern was Neighborhood Charter, where we saw percentages of White, Latino/Hispanic, and Asian/Pacific Islander students clustered around 30%. This single outlier highlights the difficulty in recruiting and enrolling a truly racially balanced population.
Racioethnic enrollment by operator in 2016–2017.
Discussion
Looking across all data sources, we can begin to answer our overarching research question: How do context and school practices shape diversity at IDCS? First, our participants’ reports of the influence of residential segregation, gentrification, and race-neutral policies are in line with prior research examining the causes of increased segregation in charter schools (e.g., Jacobs, 2013; Lubienski et al., 2009) and the recruitment challenges faced by intentionally diverse schools (e.g., Jabbar & Wilson, 2018; Villavicencio, 2016). To overcome these influences, educators in IDCS developed data-driven recruitment and enrollment practices that attempted to communicate the value of their diversity mission and influence the choice sets of diverse parents. A subset of IDCS parents identified the schools’ diversity mission as a critical factor in their enrollment decision, suggesting that schools’ messaging about the benefits of diversity can be influential. Quantitative analyses showed greater racioethnic and, and to a lesser degree, socioeconomic diversity among IDCS than comparison schools. These same data highlighted that there is no “one-size-fits-all” way to achieve greater diversity, with IDCS in different states recruiting different populations to establish multiracial/multiethnic schools. This variation implies that IDCS were successful because they were attentive to local context when recruiting for diversity. Taken together, these findings suggest that because schools’ recruitment and enrollment practices were responsive to local context and driven by a diversity mission, they were effective in promoting school diversity. We conclude that, despite contextual barriers, schools have considerable agency in fostering a diverse student population.
Despite our robust findings, this study has a number of limitations. First, our claims are not causal, although our rich data sources allowed for extensive triangulation. Areas for future research include identifying the causal effect of state policy on diversification efforts, perhaps exploiting variation in timing or geography. A key policy to examine is the use of lottery weights and the impact of using specific weights on diversity outcomes. Further, while we highlight the diversification efforts of IDCS, we did not explore the consequences of IDCS on surrounding school diversity, which may be a critical area for future research if IDCS continue to grow in popularity. Similarly, we did not examine internal integration. Future studies could look at tracking practices and opportunities for equitable intergroup contact at IDCS.
As an additional limitation, our sample was not representative of the United States or IDCS as a whole. Schools were drawn from regions with robust choice markets, high numbers of IDCS, and diverse populations. Our findings, then, may not be generalizable to regions with fewer school choice options or less racioethnic or socioeconomic diversity among the general population. Further, our parent sample was small, and White parents were overrepresented, limiting the claims we can make based on survey findings. Given the important role of parents in the choice system, researchers can expand existing literature on parents’ engagement with school choice by exploring how a representative sample of parents experience recruitment and enrollment processes of intentionally diverse schools. In light of work that has found that some schools center the needs and experiences of White communities over those of racially and linguistically diverse communities (Palmer, 2010; Turner, 2020), it is important to critically analyze how differentiated recruitment efforts are developed, disseminated, and experienced. While the practices identified in this study all point to the important goal of diversification, intentionally crafting a student population can lead to a reliance on stereotyping and/or commodifying students’ racial identities. Our survey suggested that some parents are satisfied with IDCS’ diversity mission, but more data are needed that explore how students and parents experience these practices and their schooling environments, particularly those from historically marginalized groups.
Finally, our qualitative sample only included charter schools. As school choice options continue to grow, and increasing numbers of traditional public school districts move to a portfolio management model (Bulkley & Henig, 2015), researchers should explore the degree to which the same contextual factors and school practices influence diversity within other school types. Finally, there is a tradition of rich theoretical work examining school integration, school segregation, and families’ navigation of school choice, including those using critical race theory (e.g., D. Bell, 2004) or social reproduction theory (e.g., Bourdieu & Passeron, 1990; Lareau & Weininger, 2003). Future work could employ theoretical constructs such as interest convergence or social capital to deepen our theoretical understanding of intentionally diverse public schools of choice.
Conclusion
Though our sample schools made progress in combating school segregation, considerable time and effort was required to overcome contextual barriers to diversity. Accordingly, our findings suggest that a broad range of policies could help support school diversity, including those that expand pupil transportation, support integrated housing, address parent information asymmetries, and reduce entrance lottery restrictions. Further, it appears that policies, some of which were designed to prevent discrimination or support schools’ missions, had unintended consequences for school diversity. As such, policy makers should carefully evaluate policies regarding school facilities, enrollment, or funding for their impact on diversification efforts.
Overall, though some school leaders feel they are not able to influence student diversity (Giersch, 2019), our findings illustrate that schools have significant agency and can establish diverse schools if they are responsive to local context and devote considerable efforts toward making diversity a priority. Although our findings are based on a sample of charter schools, because school segregation remains an issue across all school types, our study suggests promising practices for any school system working to establish diverse schools. At the same time, it is unreasonable, if not impossible, to expect individual schools to singlehandedly counteract the systemic forces driving school segregation. School segregation is a national issue impacting neighborhood schools and choice schools alike, and, as such, it requires a national response (Linn & Welner, 2007; Marcotte & Dalane, 2019; Orfield et al., 2016; Owens et al., 2016). While we found that schools have agency in combating school segregation, it is critical that these efforts be considered alongside federal, state, and local policies to support diversification and equalization of resources across schools and districts.
Footnotes
Appendix
Quantitative and Qualitative Analytic Samples.
| Operator a | Schools in Quantitative Sample | Schools in Qualitative Sample | Operator Type |
|---|---|---|---|
| Borough Charter | 2 | 3 | CMO |
| Grove Charter | 1 | 1 | Independent |
| Global Charter | 3 | 3 | CMO |
| Community Charter | 2 | 2 | CMO |
| Common Ground Charter | 1 | 2 | Independent |
| Cultural Charter | 1 | 1 | CMO |
| Workshop Charter | 12 | 3 | CMO |
| Neighborhood Charter | 1 | 3 | CMO |
| Journey Charter | 1 | 1 | Independent |
| Exploration Charter | 9 | 3 | CMO |
| Canyon Charter | 2 | 2 | Independent |
| Inquiry Charter | 12 | 3 | CMO |
Note. CMO = charter management organization. aAll names are pseudonyms.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors received financial support from the Walton Family Foundation for this research, but the views expressed here are solely that of the authors.