Is Responsiveness to Student Voice Related to Academic Outcomes? Strengthening the Rationale for Student Voice in School Reform

Purpose: Drawing on relevant literature, we conceptualize three pathways through which responsiveness to student voice may promote academic goals. Then, we analyze panel data of students in the Chicago Public Schools to examine this relationship. We focus on the “responsiveness” of teachers and administrators to student voice, because prior work has highlighted that opportunities for student voice are often symbolic. Our central goal is to examine whether responsiveness to student voice is related to academic outcomes. Research Methods/Approach: We draw upon student records collected in Chicago Public Schools from the 2017–18 and 2018–19 school years, as well as students’ responses to questions administered in the district’s 2019 5Essentials Survey. We estimate models of the effect of responsiveness to student voice on students’ ninth-grade academic outcomes, controlling for students’ eighth-grade attendance and grade point average. Findings: In schools that students regard as responsive to their expressed critiques, students have better grades and attendance and reduced rates of chronic absenteeism. This is the first large-scale study employing panel data to examine the relationships between responsiveness to student voice and academic performance. Implications: Proponents of student voice have long emphasized its benefits in terms of democratic values and respect for the full humanity of young people. Scholars argue these benefits may be particularly important for students from marginalized communities. This study indicates that there are academic benefits as well. However, future studies are warranted to deepen our understanding of the mechanisms through which responsiveness to student voice yields valued benefits.

Scholars argue these benefits may be particularly important for students from marginalized communities. This study indicates that there are academic benefits as well. However, future studies are warranted to deepen our understanding of the mechanisms through which responsiveness to student voice yields valued benefits.
In June 2020, as protests around the nation broke out in response to the murders of George Floyd, Breonna Taylor, and Tony McDade, students at Burroughs High School 1 on Chicago's South Side were frustrated at the lack of discussions taking place in their classrooms. While a national reckoning with anti-Black racism and police violence was being held at dining room tables and covered in the news, they wanted their teachers to open space in school to process and make sense of these injustices. An ad hoc group of students, members of their school's student voice committee (SVC), worked with a trusted teacher to craft a letter to all school staff. It was signed by more than 80 students and outlined their concerns. In the letter, students described their disappointment with the silence of the mostly White staff. They noted that whereas students took to the streets to protest and make their voices heard, their teachers-whom they expected to be just as outraged-largely failed to address the events despite their centrality to the lives of the young people they engaged with daily. The school administration met with students to discuss the letter.
Rather than seeking to minimize student concerns, the administration invited students to partner to develop a plan forward. The assistant principal worked with students to develop a teach-in where staff could learn about these issues and JOSEPH KAHNE is the Dutton Presidential Professor of Education Policy and Politics at the University of California, Riverside. His research and school reform work focus on ways to identify and support equitable and impactful forms of civic and democratic education. BENJAMIN BOWYER is a project scientist with the Civic Education Research Group at the University of California, Riverside School of Education. His research interests center around the effects of social context on the formation of political attitudes and behavior. JESSICA MARSHALL is a PhD student in learning sciences at Northwestern University, where her research has focused on the political and sociocultural dimensions of learning in civics education. Previously she served as the director of Social Science and Civic Engagement for the Chicago Public Schools. ERICA HODGIN is the codirector of the Civic Engagement Research Group at the University of California, Riverside. Her current research focuses on the distribution, quality, and influence of youth civic learning and digital civic learning opportunities.

Is Responsiveness to Student Voice Related to Academic Outcomes?
how to address such relevant and critical topics in the classroom. The assistant principal and students would go on to colead a workshop for school leaders at a district-wide leadership conference on how to work collaboratively with students to design professional learning for adults. This is but one example of the potential power and value of student voice. Through writing a letter to their administration and then codesigning professional learning for adults in their school and, later, throughout the district, students had opportunities to develop leadership skills and an understanding of how to bring about change. The school community and the district benefited as a result of students' efforts to promote authentic discussion about a relevant societal issue. Research on student voice also suggests that the students at Burroughs High School may well have developed agency, a sense of belonging, and competence as civic actors through their engagement in this work (see Mitra 2004).
The initiative at Burroughs High School grows out of and reflects a 10-year commitment by the Chicago Public Schools (CPS) to promote student voice and civic engagement. One way CPS has promoted student voice is through SVCs, which aim to bolster youth leadership, amplify student voice, and promote a positive school climate. SVCs have grown significantly in Chicago-from 5 schools in 2012 to 161 schools in 2020. The SVCs are part of a multifaceted district-wide commitment to build a culture of student voice and responsiveness to it. Broadly speaking, schools in Chicago are encouraged to include students in ongoing decision-making and development of school improvement plans that cover topics ranging from school culture to curriculum.
The CPS's systemic student voice efforts are relatively unique. Indeed, although reformers have long highlighted the value of "student voice," especially for those from marginalized communities (Benner et al. 2019;Toshalis and Nakkula 2012), within the United States such efforts have most commonly been initiated by independent student voice groups or by individual schools, often in after-school settings. Student voice initiatives have rarely been part of systemic state-or district-level policy or reform efforts (Mitra et al. 2014).
Often, the rationale for student voice efforts emphasizes democratic aims. We believe this is a worthy priority, but it has been insufficient to spur systemic action. The focus of our article reflects the belief that further conceptualization and evidence connecting student voice to academic outcomes might spur greater interest in systemic reform. Indeed, although student voice efforts have been shown to promote democratic commitments, desired forms of youth development such as a sense of agency, and a more supportive school culture, the field lacks large-scale studies that rigorously examine links between student voice and academic performance.
This study responds to that need. Specifically, with data collected in Chicago's district-wide surveys, we examine the relationship between varying levels of responsiveness to student voice and academic outcomes.
What Is Responsiveness to Student Voice? Scholars conceptualize "student voice" as the variety of ways in which youth share or have input into their education, their school, and the decisions that have an impact on their lives (Benner et al. 2019;Mitra et al. 2014). At their best, efforts to support student voice move beyond listening to include responding and, ideally, partnering with students while making space for youth-led efforts and decisions (Chicago Public Schools 2020; Toshalis and Nakkula 2012). As Cook-Sather (2006) finds in her review of the literature on student voice, these efforts are premised on the belief "that young people have unique perspectives on learning, teaching, and schooling; that their insights warrant not only the attention but also the responses of adults; and that they should be afforded opportunities to actively shape their education" (359).
With these priorities in mind, in this article, we focus not only on whether students had opportunities to express their voice but also, when they did, on whether students felt teachers and administrators were responsive. Our emphasis on responsiveness reflects a view of students as critical members of the school community who should have agency in decisions and activities affecting their academic experience and the school overall. Responsiveness does not require that adults respond positively to all of the critiques students raise, however; rather, responsiveness requires that students' ideas are taken seriously by school staff.
We contend that responsiveness to student voice is a particularly important focus for policy and reform in settings with a majority of students of color, such as Chicago, because youth of color often experience social and political marginalization (Cohen 2010;García Bedolla 2005;Junn 1999) as well as fewer opportunities for meaningful voice in school (McFarland and Starmanns 2009). Indeed, such marginalization often undermines students' motivation, their sense of belonging, and their academic achievement (Cohen and Garcia 2008;Gray et al. 2018).
Furthermore, a central challenge regarding the implementation of student voice initiatives is that they often result in "tokenism" or "co-optation" rather than meaningful engagement with student critique (Lundy 2007). This kind of shallow engagement is particularly common when it comes to schools with high percentages of low-income students, students of color, and other marginalized groups, such as students with disabilities (Benner et al. 2019;McFarland and Starmanns 2009). Tokenistic experiences are unlikely to alter students' senses of agency, belonging, or competence and may, in fact, further alienate youth and have negative consequences if they are not given due weight (Alderson 2000). Thus, responsiveness to student voice, as we conceptualize it here, takes seriously the critiques forwarded by young people, particularly marginalized youth.

Democratic Rationales for Student Voice Efforts
Proponents of student voice often take as a starting point the premise that listening and responding to youth when they raise concerns is fundamentally important in a democratic society. Advocates argue that schools should be run in ways that expect, respect, and reflect contributions from all members, particularly those who are not often represented. This view is consistent with Article 12 of the United Nations Convention on the Rights of the Child (UNCRC; United Nations 1989), which states that children should have "the right to express [their] views freely in all matters affecting the child" and that "the views of the child should be given due weight." Accordingly, following the vision articulated in Article 12, listening to students-and responding (even if not always agreeing)is essential if students' humanity is to be fully recognized.
Consistent with such priorities, John Dewey viewed listening and responding to student voice as a way for schools to enact and model democratic values. As succinctly stated in Democracy and Educational Administration, "All those who are affected by social institutions must have a share in producing and managing them" (Dewey 1937(Dewey /2008. This stance reflected Dewey's view of democracy as a way of life characterized by collaboration, communication, and respect "in all social relationships" (225). "Voice" as a priority for institutional practice in schools, Dewey argued, helps ensure that decisions reflect students' goals and priorities while enabling integration of student input regarding effective strategies for reaching these goals. In the process, responsiveness to voice also models the democratic spirit of inclusion (consistent with a focus on belonging, discussed below) while providing students with the agency and skills to pursue democratic engagement in other contexts.
Student voice and inclusion in school governance is also considered a core element of effective civic education (Gould et al. 2011;Levine and Kawashima-Ginsberg 2017). Empirical studies show that when students have opportunities to experience and practice voice in the classroom and in relation to school governance, it fosters democratic commitments and capacities (Mager and Nowak 2012;McFarland and Thomas 2006;Quintelier and Hooghe 2013;Torney-Purta 2002). Conversely, there is evidence that negative student experiences and marginalization in the daily life and operations of schools can have a negative impact on young people's long-term political development (Bruch and Soss 2018). Indeed, amid calls for the civic education field to engage more fully on issues of race and equity (Clay and Rubin 2019;Cohen et al. 2018;Mirra and Garcia 2017), responsiveness to student voice may well be a valued approach, as it provides an important opportunity for marginalized youth who are not often represented at decision-making tables.
The Limited Persuasive Power of the Democratic Rationale Unfortunately, the democratic rationale for promoting youth voice has not provided sufficient impetus for educational institutions to implement significant student voice policy or reform. A common explanation for the failure of educational institutions within the United States to promote student voice as a systemic goal is that efforts to promote voice conflict with long-established norms in schools that are characterized by sizable power distinctions between students, teachers, and school leaders (Mitra 2009). There is also evidence that prioritizing student voice contrasts sharply with broader cultural norms within the United States. As Kilbourne (1998) documents, one reason that the United States did not ratify Article 12 of the UNCRC was because of concerns that it might undermine adult authority. As a result, although the UNCRC spurred a focus on youth voice in other countries, this did not occur in the United States (Mitra et al. 2014).
Although this cultural explanation for the lack of systemic policy and reform focused on student voice clearly has merit, we are not sure that it captures the full story. In particular, we posit that an additional factor constraining the student voice movement is a lack of evidence that student voice affects valued academic outcomes to which school leaders are often held accountable. Indeed, although scholars have made the case that responsiveness to student voice should spur achievement, evidence is limited.

The Academic Rationale for Student Voice
In addition to the democratic rationale and a related "rights" rationale for student voice, scholars and reformers in the United States and elsewhere have advanced what Cook-Sather (2006) has called an "improvement" rationale. They have argued that student voice should be viewed as a significant dimension of a healthy school and classroom culture and that attending to student voice can support academic outcomes and broader school improvement (Benner et al. 2019;Cook-Sather 2002;Fullan 2015;Mitra 2009). Proponents also argue that it would be a mistake to imagine that high-quality learning environments can be pursued without attention to student voice. This belief grows from research that shows that when students engage as cocreators and authorities of their learning environments, it yields generative and productive opportunities for learning (Cook-Sather and Hayward 2021;Espinoza et al. 2020;Gutiérrez et al. 1999). Those championing academic and school improvement rationales for student voice do not generally question the democratic rationale. Rather, they position democratic and academic aims as intersecting priorities: youth are viewed as a resource for classroom and whole-school improvement at the same time that youth voice is viewed as a way to support development of democratic capacities and commitments ( Joselowsky 2007). From the standpoint of promoting reform designed to advance "student voice," however, this connection to academics is significant. It ties student voice to the central priorities of educators and reformers. Such a focus highlights the importance of the following question: Will responsiveness to student voice promote academic outcomes?

Pathways through Which Voice May Promote Academic Outcomes
In his widely read text, The New Meaning of Educational Change, Michael Fullan (2015) asks, "What would happen if we treated the student as someone whose opinion mattered in how learning occurred and for what purpose?" (174). He details the importance of hearing from students to ensure that they find learning relevant and responsive. And Fullan notes that both cognitive scientists and sociologists find substantive learning gains when there are strong relationships between educators and students and between students and the course content. Consistent with this perspective, and with the broader literature on student voice, we specify three distinct pathways through which responsiveness to student voice may yield benefits for students' academic outcomes (see fig. 1). First, there is an organizational pathway, through which responsiveness to student voice might influence school-level change. Scholars argue that student voice can support school improvement efforts and promote more impactful reform by influencing the way a school is organized and run (Levin 2000;Osberg et al. 2006;Rudduck et al. 1996). Second, there is a classroom-level change pathway, through which responsiveness to student voice might lead to improvement in curriculum and instruction ( Johnson and Nicolls 1995;Rudduck and Flutter 2000). This strategy is reflected in the example from Burroughs High School, in which students provided teachers with insight into ways that they could effectively engage with students on the topic of anti-Black racism and police violence.
The third pathway emphasizes the development of student agency, belonging, and competence. Scholars argue that when processes are enacted that engage and respond to student voice-both about the way the school is organized and run (school-level change) and about curriculum and instruction (classroom-level change)-young people gain enhanced agency, a sense of belonging to the school community, and leadership capacities (Mitra 2004; see also Fullan 2015; Toshalis and Nakkula 2012). Evidence supporting this claim comes from Vieno and colleagues (2005), who used hierarchical linear modeling and found that when students experience opportunities for voice and agency-participation in making rules and organizing events-students' "sense of community" increased at the student, classroom, and school levels (327). Relatedly, Mitra's (2004) study of Whitman High School, in Northern California, provides a detailed account of two groups of students who each focused on specific reform and improvement priorities. She found that as students worked collectively to bring about changes, these opportunities promoted participants' sense of agency, belonging, and competence-important dimensions of youth development.

Does Student Voice Affect Academic Outcomes?
Unfortunately, few studies examine how opportunities for voice influence academic outcomes. Case studies of school voice initiatives (Mitra 2004), including a sizable literature tied to Youth Participatory Action Research (see Anyon et al. 2018;Rodríguez and Brown 2009), make valuable contributions by examining programmatic design and ways that student voice efforts have fostered outcomes such as agency, belonging, and competence related to leadership skills. To make the case that opportunities for student voice are likely to affect academic outcomes, such findings are paired with studies showing that agency, belonging, and competence lead to improved academic results, especially for marginalized populations (Benner et al. 2019;Gray et al. 2018;Toshalis and Nakkula 2012). Gray and colleagues (2018), for example, highlight studies demonstrating a relationship between a young person's sense of "belonging" and academic outcomes, and they emphasize that student voice can foster this sense of belonging. The authors recommend, for example, that schools and educators consider whether "Black students' voices and perspectives [are] reflected in [your] school's academic and student affairs policies" (105). Similarly, scholars point out that attending to student voice can strengthen students' relationships to educators and engagement with course content, and they note that other studies have shown that such relationships make academic success far more likely (Benner et al. 2019;Fullan 2015). These studies, however, do not directly assess the connection between student voice efforts and academic outcomes.
Indeed, there is no direct evidence that efforts to promote student voice promote academic outcomes for an entire school population (Benner et al. 2019;Mager and Nowak 2012). In their review of qualitative and quantitative studies of student voice efforts, Mager and Nowak (2012) found just 15 studies that examined academic outcomes. Almost all of the evidence regarding relationships between student voice and academic outcomes came from single-school, qualitative case studies.
The study that follows aims to respond to this gap. Specifically, by drawing on panel data, this study examines the relationship, over time, between teacher and administrator responsiveness to student voice and academic outcomes.

Context
The schools within CPS provide a valuable testing ground on which to assess the relationship between responsiveness to student voice and academic outcomes. As described above, the school district has placed a considerable emphasis on promoting student voice through a range of initiatives. Of particular importance, the Student Voice 360 programs emphasize academic aims, and both principals and district leaders routinely highlight this academic rationale when explaining why this reform effort is so important. In fact, in 2020, the district launched a working group tasked with developing a strategic plan to support schools in integrating student voice in school improvement and decision-making efforts. This plan is integrated into a district-wide initiative to improve the city's high schools. Although a comprehensive assessment of CPS's student voice efforts is beyond the scope of this article, the district's investment in student voice infrastructureincluding the implementation of district-wide student surveys-provides an opportunity to examine the relationship between student voice and academic outcomes.
Chicago's demographics also make it an important context in which to consider these issues. District-wide, 77% of students were eligible for free or reducedprice lunch in 2018-19, compared with 49% in Illinois as a whole and 52% nationally. 2 Thirty-seven percent of CPS students identified as African American and 47% as Hispanic/Latinx, compared with 21% and 15%, respectively, statewide and 17% and 16% nationally. If, as we argued above, responsiveness to student voice is particularly important for the academic success of students from marginalized groups, then it is important to examine data from a setting like Chicago.

Method
We draw upon student records and survey data collected in CPS schools. These data include administrative and academic records for CPS students and schools from the 2017-18 and 2018-19 school years, as well as students' responses to questions administered in the district's 2019 5 Essentials Survey, which is given annually each spring and covers a range of areas from school climate to classroom experiences in a variety of subjects. Our primary data analyses look at the cohort of students enrolled in eighth grade in 2017-18 and in ninth grade in 2018-19 (see table 1). In this study, we limit our attention to students attending schools run by the CPS district and exclude students who attend charter schools within the district. 3 We focus on this cohort of students for two principal reasons. First, ninth grade is a critical year for the ultimate academic achievement of students. Studies in Chicago and elsewhere demonstrate that students who have good attendance records and earn good grades during freshman year are much more likely to graduate high school and to attend college than students who struggle in terms of their attendance or grades (Allensworth and Easton 2007;Jackson 2018;Neild et al. 2008). As such, interventions to promote high school students' academic success are expected to have a sizable impact during the ninth-grade year. Second, because it is a year in which students typically begin attending a new school, ninth grade provides an opportunity to test the effect of school climate on students' academic performance. That is, the transition to high school, in some ways, represents an exogenous change that allows us to assess the impact of a school's environment on ninth-grade students' attendance and grades. Analytical Approach Our analysis estimates the effect of responsiveness to student voice on students' academic achievement. These panel data are crucial, as they allow us to estimate models of the effect of responsiveness to student voice on students' ninth-grade (2018-19) academic outcomes that control for students ' eighth-grade (2017-18) attendance and academic performance, as well as student-level demographic and socioeconomic variables and school-level control variables. The eighth-grade academic outcomes, in particular, are strong predictors of ninth-grade outcomes.
Their inclusion in the model helps us guard against the possibility that there is some selection mechanism by which those students who are most likely to be academically successful in high school are more likely to attend more responsive high schools. We estimate two models of responsiveness to student voice on academic achievement that, as summarized in table 1, differ primarily in the level at which students' perceptions of their schools' responsiveness are operationalized. The first approach relies on student-level answers to two survey questions concerning responsiveness, whereas the second employs the school-level averages of these responses. 4 The advantage of the first approach is that there is likely to be substantial variation between students within the same high school in their experiences with their teachers and school administrators. For example, students who happened to be enrolled in one English class might have a very different opinion of their teacher's responsiveness to their concerns and suggestions than otherwise similar students in the same school who had a different English teacher. However, there is some ambiguity about the direction of the causal relationships between these student-level perceptions of responsiveness and their academic outcomes. For instance, a student who received a worse-than-expected grade might rate the teacher as being less responsive than a classmate who received a better-thanexpected grade. Moreover, it is also possible that some teachers might be most responsive to their more academically successful students. In either case, it is a student's academic performance that is influencing their perception of responsiveness, rather than the reverse. Consequently, there may be a bias whereby analyses that use student-level perceptions of responsiveness as predictors of academic outcomes overestimate the effect of responsiveness on these outcomes.
The second analytic approach avoids this challenge to the estimation of the effects of responsiveness by aggregating students' answers to the responsiveness questions at the school level. Specifically, we take the average of the responses of the ninth-grade panel in each school and employ that as the measure of responsiveness in the equation predicting the student-level academic outcomes. The strength of this design is that it more clearly establishes that it is the characteristics of the school that explain the variation in student outcomes. This helps to eliminate the possibility that any observed relationship between responsiveness and academic outcomes is spurious because of some unmeasured tendency of certain students to rate their school as being responsive (e.g., students who like going to school may both be more likely to rate their teachers as being responsive and be academically successful). As noted above, a disadvantage of this approach relative to the first is that we do not capture the variation in responsiveness that students may experience within the same school. Consequently, we may underestimate the effect that variation in responsiveness has on students' academic performance. If the first approach provides an upper bound of our estimate of the effect on responsiveness on academic outcomes, then the second approach provides a lower bound. In addition, because the panel includes data from only 86 high schools, the statistical power of this model is substantially reduced compared with the first model.

Academic Outcomes
We employ two measures of attendance and one of academic achievement as our dependent variables. The first of these outcome variables is the proportion of school days a student was absent. Student absence rates are strong determinants of academic achievement (Goodman 2014), and ninth-grade attendance, in particular, is predictive of high school graduation and college enrollment (Allensworth and Easton 2007;Jackson 2018). We operationalize this variable as the proportion of the days that a student was absent out of the total number of days in the academic year during which the student was enrolled in the school. 5 As seen in table 2, the mean proportion of absences for ninth graders in 2018-19 was 0.071. Because there is particular concern with students who are chronically absent from school (Hamlin 2021), we also employ a second, dichotomous measure of attendance. Following the US Department of Education's (2019) definition, we classify a student as being chronically absent if they were absent for 15 or more days during the school year. Twenty-seven percent of the ninth-grade sample meets this criterion, a somewhat higher percentage than the national average for high school students (21%) reported by the US Department of Education (2019).
The third measure to assess students' academic performance is their grade point average (GPA). Ninth-grade GPA is a strong predictor of future academic achievement, including graduation from high school and college enrollment (Allensworth and Easton 2007;Easton et al. 2017). We calculate GPA for ninth graders on an unweighted 0-to-4-point scale based on all of a student's semester grades in all of their courses during the academic year. The mean GPA for ninth graders in 2018-19 is 2.81. For eighth-grade students, we use their marks on their end-of-year standardized assessments in core subjects to calculate a measure equivalent to GPA (i.e., we convert the letter grades onto a standard 0-to-4-point scale and calculate an average for each student).

Responsiveness to Student Voice
The key independent variable in this study is a school's responsiveness to student voice. Crucially, we conceptualize this as involving more than the opportunity for students to express their opinions or complaints. We also include students' perceptions that their input is heard and heeded. We operationalize this by asking students about the responsiveness of their teachers and school leaders, the two most relevant groups of actors to whom they might address concerns. Specifically, we employ two items from the student survey that ask students whether they agree with the following statements about their school: "If students express concerns about a school policy, school leaders are responsive," and "If students express concerns to their teachers about their class, teachers are responsive." Respondents answered both questions along 4-point scales from "strongly disagree" to "strongly agree" (recoded 0-1 to assist with the substantive interpretation of the multivariate models). Responses to these items are fairly strongly correlated (r p 0.54; 67% of respondents gave the same answer to both questions). Consequently, we combine the two items into a single index (0-1) of responsiveness (a p 0.70) by taking their mean. As discussed above, we operationalize responsiveness to student voice at either the student level (M p 0.607,

Other Student Characteristics
In addition to their academic records, we employ several student-level control variables. These include race/ethnicity, gender, free or reduced-price lunch status, English-language learner status, and special education classification. In addition, we include a dummy variable that indicates whether the student was enrolled in any honors-level courses in the core subjects during ninth grade as a further control for academic differences across students.

School Characteristics
We also include a set of school-level control variables. A dummy variable indicates whether the high school is a neighborhood school (54% of ninth-grade students) or another type of CPS-run school, including selective-enrollment high schools. Additional school-level variables include the proportion of students at the high school who identify as Hispanic/Latinx, the proportion of students eligible for free or reduced-price lunch, and the number of students enrolled.

Estimation Approach
To estimate the effect of a school's responsiveness to student voice on students' academic outcome, we employ a multilevel analysis that accounts for the fact that the data set is composed of students who are nested within schools. In addition to the student-level and school-level independent variables, these models contain a term representing the random intercept unique to each school. For the two continuous dependent variables (the proportion of school days absent and GPA), our models employ linear mixed-effects models. Because the third dependent variable (chronic absenteeism) is dichotomous, we fit a generalized linear mixed model using a logit link function. 6

Results
The majority of ninth-grade students in CPS thought their school leaders and teachers were responsive to their concerns. Fifty-nine percent agreed that school leaders responded when students had concerns about school policies, and an additional 12% strongly agreed. An even larger percentage said that their teachers would respond to students' concerns about their classes (61% agreed and 17% strongly agreed). Table 3 presents the results of the models of the three academic outcomes for the panel of ninth graders. The first column displays the results for the model of students' absences that includes the student-level measure of responsiveness. Consistent with expectations, students who reported that their teachers and school leaders were more responsive to their concerns tended to have lower rates of absences from school. This coefficient is statistically significant ( p ! .001). Similarly, as reported in the second column, we find that responsiveness measured at the school level appears to diminish students' absences, though this effect is of borderline statistical significance ( p ! .10).
To gauge the substantive magnitude of these estimated effects, table 4 shows the expected absence rate of a typical student, calculated across a range of observed values of responsiveness. 7 For the first model of the effect of student-level perceptions of responsiveness, the difference in the expected absence rates between a student at the tenth percentile in perceived responsiveness and one at the ninetieth percentile is 0.7 percentage points (6.9% absent compared with 6.2%). 8 Similarly, in evaluating the estimated effect of school-level responsiveness, we see a small difference in predicted absence rates for a typical student in the tenth percentile of school-level responsiveness compared with a student in the ninetieth percentile (7.0% absent compared with 6.1%). 9 The next two columns in table 3 display the results for the models of chronic absenteeism. Consistent with expectations, the estimated effect of responsiveness on chronic absenteeism is negative in both models, indicating that students in more responsive schools tend to be less likely to have missed 15 or more days of school. However, although the estimated effect of student-level responsiveness is statistically significant ( p ! .05), the estimated effect of school-level responsiveness is not. Table 5 reports predicted probabilities of being chronically absent that are calculated from these models across a range of values in responsiveness. The estimates based on the first model indicate that students' perceptions of the responsiveness of their teachers and school leaders are associated with a notable reduction in chronic absenteeism: a student in the tenth percentile has a predicted probability of being chronically absent of 26.2%, and this probability declines to 22.9% for a student in the ninetieth percentile. Similarly, the probability of chronic absenteeism is 2.6 percentage points higher for a student attending a school in the tenth percentile of responsiveness than for a student at a school in the ninetieth percentile (23.4% compared with 20.8%, though this effect is not statistically significant). For an average ninth-grade class of 211 students, this latter estimate translates into a difference of five fewer chronically absent ninth graders in the more responsive school. The final two columns of table 3 report the models of ninth graders' annual GPAs. As expected, we find that greater responsiveness tends to be associated with higher GPAs. This effect is statistically significant for both the model that employs student-level perceptions of responsiveness and the model that uses the school-level averages. Table 6 shows how expected GPA varies across the range of responsiveness. For the student-level model, the difference between a student in the tenth percentile in perceived responsiveness and one in the ninetieth percentile is 0.14 points (2.69 vs. 2.83). In the school-level model, we observe a similar difference: expected GPA increases from 2.65 for a student in a school in the tenth percentile to 2.85 for a student in the ninetieth percentile. Put somewhat differently, we would expect an average student in the tenth percentile of school responsiveness who took six courses to receive four Bs and two Cs on their report card, whereas an otherwise identical student in a highresponsiveness school would be expected to receive five Bs and just one C.
The appendix (available online) presents additional models that serve as robustness checks of the observed association between schools' responsiveness to student voice and academic outcomes. These supplemental, cross-sectional models employ data for all CPS students in grades 7-12 and include additional schoollevel variables drawn from CPS's annual parent survey. Specifically, we include measures of parent's/guardian's sense of their school community, parent's/ guardian's sense of partnership with their child's teacher, and parent's/guardian's ratings of the school facilities. If the school-level measures of responsiveness were picking up on the effects of school "quality" more generally, then we would expect that its relationship to our outcome measures would be attenuated by the inclusion of these additional school-level measures. However, we find that the relationships between responsiveness and our three academic outcomes are all statistically significant and that they are not affected by the inclusion of these three additional variables. These results provide additional evidence that it is a school's responsiveness to students' concerns, not some other attributes that one might associate with a "good" school, that is correlated with variation in students' academic outcomes.

Discussion
Although positive rhetoric often accompanies statements about the importance of student voice, there are very few rigorous studies examining the relationship between student voice and academic outcomes. This article responds to that gap. We believe it is the first large-scale quantitative study employing panel data to assess the relationship between student voice and academic outcomes. Drawing on a unique data set from a large and diverse school district, we observe a consistent, positive relationship between a school's responsiveness to student voice and students' grades and attendance. While controlling for prior academic outcomes, responsiveness to student voice, whether measured at the student level or school level, was related to higher GPAs, fewer absences, and less chronic absenteeism. However, although the student-level measure of responsiveness was a statistically significant predictor of all three academic outcome variables, the school-level results were mixed. Responsiveness was a statistically significant predictor of grades, but it was only marginally significant with respect to absences and was not a statistically significant predictor of chronic absences. The limited statistical significance of the school-level analyses may well have been due to a lack of statistical power, as we have only 86 schools in our sample. Although the magnitudes of the estimated effects are modest, they could have a lasting impact on students' academic trajectories, given the importance of the ninth-grade year to students' subsequent academic achievement in high school and rates of college attendance (Allensworth and Easton 2007;Jackson 2018;Neild et al. 2008). Moreover, improved attendance and GPA are not only important ends in their own right: such improvements, because they are associated with student judgments of responsiveness, may well also signal a more positive sense of attachment to the school. Although these findings provide support for policy and reform efforts designed to promote student voice and strengthen our understanding of the relationship between student voice and academic outcomes, it is clear that a great deal of work remains to be done. This study relies on observational data and therefore cannot test the causal relationship between responsiveness to student voice and academic outcomes with the same certitude that would be possible with an experimental design. There is a clear need for field trials and studies that focus on specific initiatives. Such studies would make causal claims more possible, and they would also provide an evidence base regarding the desirability of particular policies and programmatic features.
Additional studies are also needed to help us better understand the mechanisms through which student voice promotes desired outcomes. Specifically, our review of the literature suggests there are dynamics at the organizational, student, and classroom levels through which responsiveness to student voice might support the improved academic outcomes that we observe. Fine-grained analysis will be needed if we are to examine mechanisms associated with this model. For example, our measure of responsiveness to student voice currently combines an item assessing the responsiveness of teachers with an item assessing the responsiveness of administrators. In future studies, it would be valuable to include more items and assess teacher and administrator responsiveness as separate constructsconstructs that might, potentially, have differing relations to desired outcomes. The responsiveness of school leaders, for example, might be more likely to have an impact on attendance, whereas the responsiveness of teachers might be more likely to have an impact on grades. In addition, although it was not possible in the current study, it would be worthwhile to associate the responsiveness of a particular teacher with grades in that teacher's classroom, rather than measuring teacher responsiveness in the aggregate. Furthermore, our model suggests that responsiveness will influence students' agency, belonging, and competence. Measures of those mediators and examination of their relationship to responsiveness will be essential for determining which of these pathways are most important for student outcomes. In short, both qualitative and quantitative studies are needed to clarify when and why these and other hypothesized dynamics related to student voice may yield desired results.
It is also important to note that this study's findings are consistent with those of related studies focusing more broadly on supportive school contexts. Studies by the University of Chicago Consortium on School Research, for example, consistently find that supportive school environments promote desired academic outcomes (Hart et al. 2020). Similarly, Daniel Hamlin's (2021) large-scale study of students in New York City finds that a positive school climate is associated with modest improvement in attendance. Scholars have also found that students who feel a sense of belonging have better attendance and are less likely to drop out (Christle et al. 2007). Because responsiveness to student voice is, in many respects, an indicator of a supportive school context, the fact that supportive school contexts are positively associated with academic outcomes adds to our confidence regarding the findings presented in our study.
At the same time, we believe that responsiveness to student voice is different from other forms of a supportive school context in some important ways. Measures of supportive school contexts generally highlight factors such as student safety, attachment to one's school, and trusting relations with others (Thapa et al. 2013). We constructed our assessment to embody responsiveness to student critique instead. By focusing on whether educators respond to young people's critiques, we aim to firmly center democratic rights by recognizing that students have agency, insights, and a right to be heard. This measurement of responsiveness to student voice, consistent with the critical synthesis provided by Cook-Sather (2006), aims to highlight not only the need for educators to support students but also that educators should value students' perspectives and critiques and be open to their visions regarding necessary change. Scholarship that examines this asset-based and critically oriented stance is therefore also needed.
In addition, although we have argued that evidence highlighting the impact of responsiveness to student voice on academic outcomes may increase mainstream support for student voice, there is a risk that focusing on the potential academic benefits of attending to student voice will orient teachers and educational leaders to do so in a narrow or performative way, solely to achieve mainstream academic goals. It is also possible that support for responsiveness to voice will only be conferred when it relates to minor or convenient issues and will not extend to forms of student voice that embody sizable critiques. Indeed, absent the kind of student mobilization, as well as the support of the assistant principal, that occurred in the opening example from Burroughs High School, students' critiques regarding significant issues such as institutionalized racism will often receive little response. Scholarship is needed that examines ways that responsiveness to critiques may vary depending on the context and the concern being voiced (see Mayes 2020). Moreover, further examination of the relationships between responsiveness to student voice and student engagement in school can contribute to more expansive views of learning in context.
The generalizability of our findings is limited, of course, by the fact that they are based on the study of a single school district. At the same time, given that a large proportion of students in CPS come from economically disadvantaged backgrounds and racially minoritized groups, these findings signal that responsiveness to student voice can make a difference among young people who face marginalization in their lives both inside and outside of school. Scholarly work is needed to specify strategies, conceptualize how such efforts respond to the realities of youth across contexts, and document their impacts. Furthermore, although this study focuses on academic priorities, we do not believe these are the only priorities to which scholars or policy makers should attend. Scholars have found that when marginalized youth experience a lack of responsiveness within schools, and when school contexts do not create space for critique or acknowledge systemic inequality, students' sense of belonging and their trust and engagement with civic and political institutions diminish (Bruch and Soss 2018;Gray et al. 2018;Hope and Jagers 2014). Thus, in addition to traditional academic goals, promoting responsiveness to student voice may well advance commitments to both civic and political engagement. Moreover, to honor students' full humanity, examinations of responsiveness to student voice should also prioritize identification of supports for student voice that enable marginalized youth to draw on their critiques and imaginations to create learning environments where more youth thrive and experience love and joy (King 2020; Love 2019; Muhammad 2020).

Conclusion
Often, educational approaches designed to promote traditional academic goals are thought of as distinct from those designed to promote democratic aims. Scholarship has long recognized that this is a false choice, but the distinction is often made nevertheless. Therefore, one of the particularly promising aspects of student voice as a reform priority is its potential to better integrate these educational aims.
When educators ignore students' voices, one impact may well be democratic and educational disengagement. Unfortunately, school reform efforts all too often ignore student voice completely or only consult students in superficial ways. Even when responding to crises in education-such as the debates over how to reopen schools amid the COVID-19 pandemic-students' voices and, perhaps especially, their critiques are often not given due weight. Further, policy and reform efforts focusing on student voice are particularly salient during a period in which there is increased recognition of the demand for educational institutions to respond to the needs and concerns of youth from marginalized backgrounds. We are hopeful that evidence indicating academic benefits in school environments where students feel their concerns are attended to will prompt additional study of these practices and help pave the way for integrating student voice in more frequent and significant ways into school reform efforts.
Clearly, scholars, policy makers, and reformers still have much to learn about ways to promote responsiveness to student voice. We suspect the benefits of such efforts will be wide ranging. What starts as responsiveness to critique might lead to the kinds of partnerships between students and educators that generate meaningful and lasting changes to school climate and culture and, as a result, support academic and democratic aims.

Notes
We wish to acknowledge valuable support in conceptualizing the form and focus of this study, as well as in facilitating access to the data, from Heather Van Benthuysen and the Social Science and Civic Engagement Department of Chicago Public Schools, and from John Easton and Holly Hart at the UChicago Consortium on School Research. Funds for this study were provided by the Robert R. McCormick Foundation and the Spencer Foundation. We would also like to thank the three anonymous reviewers for their insightful feedback on an earlier draft of this article. Of course, all responsibility for the content of the article rests with us.
3. Details about the 2019 5Essentials Survey can be found on the University of Chicago Consortium on School Research's website: https://consortium.uchicago.edu/publications /supporting-school-improvement. For 2019, the response rate for students across all CPS schools was 81%. Because the CPS civic initiative has not directly supported charter schools, we focus exclusively on schools that are run by the district. We also exclude students who did not attend a CPS school in eighth grade, because we do not have any data on these students' school attendance or academic performance before ninth grade. Finally, we include only ninth-grade students who were in eighth grade in the previous academic year, because they are likely to be systematically different from the group of students who were repeating the ninth grade. For the 86 high schools that we analyze, we estimate that the school-level response rates (calculated as the percentage of enrolled students as of the twentieth day of class in the fall that took the survey in the spring) range from 36% to 96%, with a median of 75%. 4. Respondents for whom data were missing for any of the student-level variables included in a particular model were excluded from that analysis. Because the responsiveness questions were randomly assigned to half of the students taking the survey, the sample used in the estimation of model 1 is a little less than half the size of that used in the estimation of model 2.
5. For students who attended multiple schools during the 2018-19 academic year, we use their records from the school in which they finished the year.
7. The predicted values of each dependent variable are calculated using the coefficients reported in table 3 by varying the value of responsiveness while holding all other variables constant at their sample mean (for continuous variables) or mode (for categorical variables). 8. To put these percentiles into more context, the tenth percentile for the student-level measure of responsiveness roughly corresponds to a student who answered "disagree" to both questions, whereas the ninetieth percentile corresponds to a respondent who answered "agree" to one question and "strongly agree" to the other. 9. For the school-level measure of responsiveness, the tenth percentile corresponds to a school in which approximately three-quarters of students answered "agree" to one of the responsiveness items and "disagree" to the other, and one-quarter answered "agree" to both questions. The ninetieth percentile corresponds to a school where these ratios are reversed, with roughly three in four students answering "agree" to both questions and a quarter splitting their answers between "agree" and "disagree."