Since the release of Equality of Educational Opportunity, researchers have emphasized the importance of applying the results of research to policies for school improvement. Policies tell educators to do something, but not how to enact specific laws. This study analyzes data from 347 schools in 21 districts to identify variables that support the enactment of policies for parental engagement. We address research questions on how school and district practices affect the quality of school-based partnership programs. Our results indicate that a policy on parental involvement may be a good first step, but other factors—principals’ support for family and community engagement and active facilitation of research-based structures and processes by district leaders—are important for establishing a basic partnership program. These factors promote programs that engage all students’ families. Schools that take these steps have higher percentages of engaged families and report higher rates of average daily attendance among their students.


district leadership, school leadership, family and community involvement, partnership program development

There are interesting questions to ask about the role of federal, state, and local policies in improving programs of school, family, and community partnerships. For instance, though policies are important for promoting school improvement, how much do policies affect school change? This study explores the responses of schools and districts to policy recommendations for partnership programs and the connections between these programs and family engagement and student attendance.


The Equality of Educational Opportunity (EEO) report (Coleman et al. 1966) focused attention on the importance of families in children’s education, based mainly on analyses of measures of family socioeconomic status. Its findings about the strong connections of family background and weak contributions of school resources to student achievement sparked a decade-long argument among social scientists on the question: which is more important for student learning—the school or the family? The debate spurred the field of education research into action that has continued to this day. To study influences on student learning, researchers began collecting new and better data on school and classroom environments, students’ opportunities and motivation to learn, family factors, and the connections between home, school, and community. [End Page 202]

We and our colleagues were inspired by the controversy to change our research question away from the “contest” of family versus school and away from the seemingly fixed inequities in family involvement linked to parents’ socioeconomic status (SES). We posed a new question: if families are so important for student success in school, how can all schools engage all families so that more students benefit from their parents’ support and encouragement to do their best in school? This new, more difficult question required research on school policies, school organization, leadership, and the alterable variables that might produce more equitable programs in family and community engagement for the success of more students.

Historically, family engagement has been treated as about the parents, that is, as external to schools. Our new question asked whether and how teachers and administrators could work with all students’ parents and with community partners from the earliest years on to ensure students’ readiness for school, grade-level learning, progress and promotion to the next grade, and on-time graduation from high school. This approach, in making school, family, and community partnerships a component of school organization—one that is central to other school improvements—ultimately was about the students.


The Elementary and Secondary Education Act (ESEA) lists requirements for parent and family engagement at the school, district, and state levels, building on guidelines in ESEA reauthorizations since 1988.1 Schools are required to engage all families in ways that support student achievement. Districts are told to assist all schools in developing partnership programs. States are expected to collect and review district policies on parental involvement, and the requirements are monitored for compliance to justify the continuation of Title I funds (Cowan 2003). The policy tells educators to engage families, but does not specify how to meet these requirements or how to improve the quality of their partnership programs. There is, then, a critical gap between the intent and enactment of the law.


We have drawn on three theoretical perspectives to guide our research questions and analyses. Sociocultural and organizational learning theories posit that districts and schools learn from each other when they share leadership and responsibilities for school improvement, including family and community engagement (Honig 2006, 2008; Huber 1991; Spillane and Diamond 2007; Stein and Coburn 2008). Socio-cultural learning theory asserts that good communication between and among colleagues who gain knowledge, exchange ideas, and take actions to develop a “culture of collaboration” affects the organization as a whole (Knapp 2008; Wenger 1998).

Organizational learning theory states that organizations improve when leaders share knowledge, plan actions, conduct evaluations, gather evidence, make sense of data, and identify best practices (Elmore 2004; Senge 1990; Supovitz 2006; Weick 1995). In combination, the two theories reinforce each other with expectations that organizations and individuals will learn and advance. That is, the interpersonal exchanges at the heart of sociocultural learning theory are informed by attention to useful data, which is central to organizational learning theory (Honig 2008; Leithwood and Prestine 2002; Louis 2008; Mayrowetz 2008).

A third theory calls attention to the content of leadership for developing programs for school, family, and community partnerships. The theory of overlapping spheres of influence asserts that children learn and grow at school, at home, and in the community, and that they benefit when parents, teachers, and others in [End Page 203] the community collaborate in ways that encourage learning and development (Epstein 1987, 2011). Goal-linked involvement activities implemented by educators, parents, students, and community members should reduce the distance and potential discord between home, school, and community and increase the quality of school-based partnership programs. Out-reach and information from schools should increase the number of involved families and improve goal-linked results for students.

This interdisciplinary theory of overlapping spheres of influence specifies an external model that represents the degree of the shared interests and actions of home, school, and community concerning student learning and school success. An internal model recognizes that the student is the central actor in learning and specifies the complex relationships and interactions of parents, teachers, and community partners.

In research on the practices that occur in the overlapping contexts of home, school, and community, we identified a framework of six types of involvement that helps categorize separable practices of partnership that pose unique challenges to engaging all families and that produces different results for student achievement and behavior. The six types of involvement—parenting, communicating, volunteering, learning at home, decision-making, and collaborating with the community—can be activated to engage families with children on specific school improvement goals (such as improving students’ reading skills and attitudes, attendance, or health) (Epstein et al. 2009).

Programs based on this theory change family involvement from an external factor unrelated to schools to an essential component of school and classroom organization (Bryk et al. 2010; Epstein and Sheldon 2006). Partnership programs require leaders to set policy, select or customize and conduct practices, and evaluate progress in engaging all families. In each school, an Action Team for Partnerships (ATP) provides the structure for planning, implementing, and evaluating a site-based program of family and community engagement linked to goals for student learning and development (Epstein et al. 2009).

The three theories support the process of side-by-side leadership (Epstein, forthcoming). In contrast to top-down directives from districts to schools or bottom-up reports of good practice, side-by-side leadership recognizes the importance of multidirectional learning that is enriched—not restricted—by the dissimilar roles of participants in diverse learning communities. Rather than focusing only on prescribed procedures or narrow monitoring for compliance (as in top-down vertical networks), or only on atheoretical trial-and-error approaches (as in bottom-up networks), side-by-side leaders customize communications, develop tools, collect and analyze data, and take action to continually improve school programs (as in Continuous Progress Learning Communities; see Bryk et al. 2015). Researchers, district leaders, principals, teachers, parents, and others work side by side to exchange information and ideas and raise questions to improve research and practice.

We draw from these theories to test whether and how school-based actions, district assistance, and the simultaneous and joint work of school team members with district leaders improve the nature and extent of family and community engagement and results for parents and students.


The National Network of Partnership Schools (NNPS) at Johns Hopkins University was established in 1995 to close the gap between written policies with directives for family and community engagement and actions taken at the school, district, and state levels to engage all families in ways that contribute to student success in school. In NNPS, results of research on the structures and processes for organizing effective and equitable programs of partnership are translated into training, tools, and publications for educators and parents. Leaders who join NNPS are guided to use the research-based approaches to enact policy and improve practices so that all families are involved in their children’s education in age-and grade-appropriate ways from preschool through high school.

In NNPS, each district must identify a leader for partnerships who facilitates and [End Page 204] encourages school-based ATPs to build their capacities to plan, implement, and continually improve their programs to create a welcoming school climate and support site-specific goals for student success. At the end of each academic year, district leaders and schools in NNPS complete UPDATE surveys to evaluate their work and progress. NNPS developed reliable scales and measures on these surveys to assess district and school progress on essential elements that affect the quality of partnership programs from one year to the next. The data are collected and analyzed for annual reports for NNPS members and for the public (Epstein and Ames 2016; Sheldon and Ames 2016).


Prior studies based on data collected separately from districts and from schools in NNPS explored factors that affected the quality of school-based and district-level partnership programs and practices. Research on schools found that those with well-functioning ATPs, strong support from principals, and positive ratings of the assistance received from district leaders were more likely than other schools to have higher-quality programs of family and community involvement (Hutchins and Sheldon 2013; Sanders and Sheldon 2009; Sheldon 2005, 2008; Van Voorhis and Sheldon 2004). Research on districts found that leaders who directly and actively facilitated school ATPs were more likely than other district leaders to report that their schools were making more progress in developing and improving programs of family and community involvement (Epstein 2008; Sanders 2008, 2009).

One study went further, studying schools nested within districts to understand the simultaneous efforts of district leaders and school teams to work together to improve school-based partnership programs (Epstein, Galindo, and Sheldon 2011). In that study, which combined independently collected school and district data, schools with at least three years of assistance from district leaders developed more advanced family engagement activities and engaged more families than did schools without consistent district support.


This study builds on the prior work with extended analyses of schools nested within districts to learn whether and how district assistance to school-based teams affects partnership program development and results for parents and students. We set four research questions:

  1. 1. How do school factors affect the implementation of:

    1. a. Basic structures and processes to implement school-based programs of family and community involvement?

    2. b. Advanced outreach activities to involve families who are typically uninvolved or “hard to reach”?

  2. 2. How do district factors affect schools’ basic and advanced partnership programs?

  3. 3. How do basic and advanced partnership programs affect the percentage of parents who are good partners with the school in their children’s education?

  4. 4. How do basic and advanced programs and the percentage of parents who are good partners with their schools affect school reports of students’ average daily attendance?

Figure 1 presents the research model and hypothesized paths of influence to address the research questions. Analyses were conducted to learn whether and how school and district practices affected the quality of schools’ basic partnership program implementation and advanced outreach to involve more families, and whether the quality of these programs measurably predicted the percentage of involved parents and students’ average daily attendance.


Survey data were collected in 2014 from 347 schools in 21 districts that were members of NNPS at Johns Hopkins University. The schools were located in large urban (20.9 percent), small urban (32.2 percent), suburban (26.6 percent), and rural (20.3 percent) areas across the country. The majority of schools (67.7 percent) served elementary and K–8 students; the rest were middle and junior high schools (18.4 percent) and high schools (13.9 percent). A few schools with mixed-grade organizations were [End Page 205]

Figure 1. Theoretical Effects Model: Connections Between School and District Leadership on the Quality of School-Based Partnership Programs, Parent Participation, and Rates of Student Attendance Source: Authors’ compilation.
Click for larger view
View full resolution
Figure 1.

Theoretical Effects Model: Connections Between School and District Leadership on the Quality of School-Based Partnership Programs, Parent Participation, and Rates of Student Attendance

Source: Authors’ compilation.

[End Page 206]

excluded from the analyses. The sample closely matches the proportion of elementary and secondary schools in the nation as a whole (National Center for Education Statistics 2015).

Most schools (69.9 percent) received Title I funds. Within schools, on average, 67.8 percent of students were eligible for free or reduced-price lunches. About 39.9 percent of students were African American, 41.1 percent white, 14.9 percent Latino/Hispanic, and 2.6 percent Asian American, and a small percentage of students had other backgrounds. Across schools, an average of 5.1 languages other than English were spoken at home by students’ families, ranging from only English to over 38 languages. The schools enrolled an average of 628 students, with school size ranging from under 50 to over 3,000 students. The variation in the demographics of this sample generally reflects the diversity in the nation’s schools.

In addition to diverse demographics, districts and schools in NNPS varied in how long they had worked on research-based approaches to partnership program development, ranging from one to nineteen years. Some had joined recently, whereas others had become expert leaders on partnerships.2

Because all districts and schools join NNPS to obtain guidance and support to implement the structures and processes that enable them to organize effective partnership programs, the sample eliminates one kind of selection bias. However, each site enters the network with a unique history and works to improve its program at its own pace. In this way, the variations among sites make it possible to study whether and how district and school leaders’ actions affect the quality of partnership programs.


Dependent Variables

Two measures of the quality of partnership program implementation at the school level are of interest in this study: basic program implementation and advanced program outreach. Two measures of the results of partnership programs extend prior studies: the percentage of involved parents and students’ average daily attendance.

The Quality of Basic Program Implementation

This twelve-item scale (α = 0.92) measured whether and how well schools organized the basic components of a partnership program. The items were scored from 1 (did not do) to 4 (did very well) to reflect low-to-high implementation of structures and processes such as whether the school established an action team; wrote an action plan; implemented activities for six types of involvement; implemented involvement activities linked to school improvement goals for student success; evaluated the activities that were implemented; and conducted other basic organizational activities. Schools averaged 3.22 on this scale (standard deviation = 0.59), indicating that most teams viewed their work on the basics as “okay,” with clear variation among the schools in the sample.

Advanced Implementation

This nine-item scale (α = 0.85) measured whether and how well a school implemented activities to solve challenges to engage all families, including those who are typically hard to reach, and to improve the implementation of activities for six types of involvement. The items were scored from 1 (not yet working on this challenge) to 4 (solved this challenge) and averaged to reflect low-to-high attention to engaging all students’ families. For example, teams reported on whether they worked to get information from workshops to families who could not attend; communicated with families who did not speak English at home; involved major demographic groups of families in school decisions; recruited and trained volunteers; and addressed other challenges that might limit family and community involvement. Schools averaged 2.76 on this scale (standard deviation = 0.56), indicating that most were making fair-to-good progress in addressing challenges, whereas others were not yet focused on advanced implementations to engage uninvolved families.

Percentage of Involved Parents

This five-item scale (α = 0.82) measured school reports on the percentage of parents who attended [End Page 207] back-to-school nights, attended parent-teacher conferences, volunteered at school, monitored children’s homework, and were considered “good partners” in children’s education. The items ranged in six increments from 0 percent (none) to 100 percent (all) and were averaged to reflect low-to-high parental participation.

Students’ Average Daily Attendance

This single item reported students’ average daily attendance (ADA), which ranged from as 31 to 100 percent across schools.

Independent Variables

School Variables (Level 1): Principal’s Support

This ten-item scale (α = 0.92) measured how strongly the principal supported the work of the ATP. Items were scored from 1 to 4 to indicate whether a principal never, seldom, often, or always provided time for team meetings; encouraged families to participate in involvement activities; encouraged teachers to work on partnerships; allocated funds for involvement activities; and offered other support for the school’s program of family and community involvement. The average sum for this scale was 3.5 (standard deviation = 0.54), indicating that most principals were positive about team-work for partnerships, but varied in the kind and extent of their support.

School Variables (Level 1): School Reports of District Support

This seven-item scale (α = 0.90) measured ATP reports on the extent and helpfulness of assistance from district Leaders for Partnership in program development. The items, which were scored 1 (no district support provided), 2 (not helpful), 3 (helpful), or 4 (very helpful), focused on district leaders’ training teams for partnerships, funding, recognition, help with evaluations, ideas for best practices, and other assistance. The average school report on district assistance was 3.12 (standard deviation = 0.75), indicating that most teams acknowledged some support from district leaders, but varied in which facilitative actions they experienced and how helpful they rated these actions to have been.

District Variables (Level 2): District Facilitation

This seven-item scale (α = 0.89) measured the reports of district Leaders for Partnerships on whether and how well they facilitated school-based Action Teams for Partnerships to organize and improve their partnership programs. The items, which were scored 1 (not conducted), 2 (need to improve), 3 (okay), or 4 (conducted very well), focused on assistance to schools on basic actions to establish their partnership programs. This included help forming an Action Team for Partnerships, understanding and using the framework of six types of involvement, writing a One-Year Action Plan for Partnerships, collecting the schools’ plans to follow work and progress, helping to develop a budget, meeting with the principal about teamwork and partnerships, and helping to evaluate work and progress. The average facilitation score was 3.05 (standard deviation = 0.80), indicating that most district leaders provided some direct assistance to schools in ways they deemed “okay,” but varied in how well facilitation was progressing.

Background Measures

Analyses statistically controlled school demographic variables, including grade level (elementary = 0, secondary = 1), size of school enrollment, percentage of students receiving free or reduced-price lunch, and percentage of English language learners (ELL). In prior studies, grade level was a significant variable, whereas the poverty level of the school and students’ language services were less consistently important for the development or quality of a partnership program (Epstein 2008; Epstein, Galindo, and Sheldon 2011; Sheldon 2008). The background variables remain of interest in studies of parental involvement and were included in all major analyses in this study.


Using Stata, we analyzed two-level hierarchical linear models (HLM) that permitted attention to the independent and simultaneous relationships [End Page 208] of key explanatory variables at the district and school levels.3 HLM accounts for the fact that schools within a district are guided by the same policies and leaders and are likely to be more similar to one another in many ways than schools selected at random. HLM techniques, which adjust for the impact of clustered errors (Raudenbush and Bryk 2002), produce less-biased and more accurate estimates than do less rigorous methods for studying school (level 1) and district (level 2) effects on the quality of partnership program implementation and outreach.

To address this study’s first two research questions, we analyzed a series of HLM models focused on two dependent variables—the quality of basic program implementation and the extent of advanced outreach to all families. First, we estimated a fully unconditional model with no predictors to identify within-and between-district variance for each outcome. Then we analyzed conditional models that tested the relationships of school-level variables and the district-level variable with each outcome.

In all models, the intercept was defined as random, slopes were fixed, and continuous measures were grand-mean-centered so that each level 1 coefficient represented the average effect across schools. Significant coefficients for level 1 variables indicate that school-based actions and experiences affected the quality of basic program implementation and advanced outreach to involve all families. Significant coefficients for the level 2 variables indicate that district leaders’ assistance to ATPs independently affected the quality of schools’ basic organization and the advanced outreach of their partnership programs.

To address the third and fourth research questions, we conducted ordinary least squares (OLS) regression analyses in step-wise progression to examine whether the school variables in the HLM analyses (principal and school reports of district support for partnerships) were associated with parents’ participation in their children’s education and students’ average daily attendance. We also explored a constellation of district variables to see whether the duration (years) of districts’ membership in NNPS was associated with leadership qualities and with district leaders’ reports about their schools’ progress on partnership program development.


The initial HLM unconditional analyses showed that 12.1 percent of the variance in basic program implementation and 11.4 percent of the variance in advanced outreach to involve all families were between districts. There was, then, enough between-district variance to explore the relationships between district characteristics and the two program outcome variables at the school level.

Basic Program Implementation

In table 1, school-level data in model 1 show that schools with a greater percentage of students receiving free or reduced-pricemealsre-ported lower levels of basic program implementation (B = 0.003, p ≤ 0.006). Model 1 also shows that, with demographic characteristics statistically controlled, schools were more likely to implement a partnership program at a basic level when there was strong principal support for partnerships and when schools’ ATPs reported receiving helpful support from their district leaders (B = 0.466, p ≤ 0.000, and B = 0.201, p ≤ 0.000, respectively). With poverty level, principal support, and schools’ reports of district support taken into account, elementary and secondary schools in NNPS, large and small schools, and those with more or fewer ELLs did not differ significantly on the nature and extent of their basic program implementation.

Model 2 shows the two-level model for basic partnership program implementation that adds district leaders’ reports of the extent and quality of their facilitation of schools’ ATPs to the equation in place of the schools’ reports of district support. The analysis confirms the results reported for model 1 indicating that schools had stronger basic partnership program [End Page 209] implementation when there was greater principal support for family and community engagement (B = 0.550, p ≤ 0.000). Model 2 extends knowledge by showing that district leaders’ reports of the nature and quality of their active facilitation of schools’ partnerships programs were associated with schools’ stronger implementation of basic partnership program elements (B = 0.108, p ≤ 0.003). These findings spotlight the independent important impact of both school leadership and district leadership on the extent to which schools organize, plan, and implement efforts to engage families in their children’s education.

Table 1. Multilevel Models Predicting Basic and Advanced Partnership Program Implementation
Click for larger view
View full resolution
Table 1.

Multilevel Models Predicting Basic and Advanced Partnership Program Implementation

Advanced Program Implementation

In table 1, models 3 and 4 report results of HLM analyses predicting advanced implementation of partnership programs that aimed to engage all students’ families, especially those who were previously uninvolved or “hard to reach.” In these models, the demographic characteristics of school populations are not significant explanatory variables. In model 3, principal support and schools’ reports of district support are significantly and positively associated with more active outreach to engage all families (B = 0.361, p ≤ 0.000, and B = 0.177, p ≤ .001, respectively). In model 4, only principal support remains an important variable for meeting challenges to engage all families (B = 0.465, p ≤ 0.000).

Results for Parents and Students

Table 2 uses OLS regression to extend the results of the HLM analyses by exploring the temporal nature of schools’ partnership program development. Column 1 reproduces the school-level HLM analyses indicating that low-poverty schools reported stronger basic partnership program implementations. With demographic variables accounted for, schools with strong principal support for partnerships and school teams’ reports of helpful district support were significantly associated with strong partnership programs.

In column 2, analyses indicate that in addition to principal support (β = 0.140, p ≤ 0.016), the nature and quality of schools’ basic program implementation was positively and significantly linked to advanced implementation to engage all students’ families in productive ways (β = 0.471, p ≤ 0.000).

Following potential influence paths, column 3 shows that elementary schools and low-poverty schools had greater percentages of parents engaged than did secondary schools (β = −0.351, p ≤ 0.000, and β = −0.463, p ≤ 0.000, respectively). Additionally, with all background variables statistically controlled, schools with more advanced program implementations had more parents engaged in their children’s education (β = 0.286, p ≤ 0.000).

Column 4 indicates that when more families were engaged, students attended school on [End Page 210] a more regular basis, as reported in rates of average daily attendance (β = 0.172, p ≤ 0.023). Table 2 suggests the stage-wise development of partnership programs. Plans and actions to implement a basic partnership program to address challenges to engage all families were likely to have taken place before advanced work and to have contributed to it. More parents in these schools became engaged and more engaged parents contributed to better student attendance.

Table 2. OLS Regression Analyses of School-Level Predictors of Partnership Program Implementation, Parent Participation, and Student Average Daily Attendance
Click for larger view
View full resolution
Table 2.

OLS Regression Analyses of School-Level Predictors of Partnership Program Implementation, Parent Participation, and Student Average Daily Attendance

The results reported in tables 1 and 2 suggest that programs of partnership can be strengthened and improved over time. Some schools do more than others to organize the structures and processes of a basic partnership program. The stronger the basic program, the more likely a school is to address challenges to engage parents who typically are under-or uninvolved, including fathers and parents who speak languages other than English.

Qualities of District Leadership

To delve deeper into the finding that district leadership matters for schools’ programs and progress on partnerships, we used an independent measure from NNPS records that was separate from the survey data. District leaders had been members of NNPS for one to nineteen years, with an average of 6.8 years. Using this indicator for all districts that reported UPDATE data in 2014 (N = 39, including the twenty-one districts with four or more schools in the HLM sample), we explored whether the number of years in NNPS affected the nature and extent of district leadership and reports of whether schools were making good progress versus no or little progress in their programs of family and community engagement.4

Table 3 reports a constellation of informative correlates. Years in NNPS was not significantly related to general leadership qualities (for example, for establishing a district office for partnership program development, planning a budget, and writing work plans, r = 0.127 NS). This reflects the fact that districts joined NNPS with a common goal to get some assistance in improving their partnership programs. By contrast, years in NNPS was significantly correlated with major measures of how [End Page 211] well district leaders facilitated school teams (r = 0.483, p < 0.01), emphasized the evaluation of progress on partnerships (r = 0.373, p < 0.05), and worked with colleagues on the partnership agenda (r = 0.395, p < 0.05). The duration variable also was related to district leaders’ reports of the percentage of schools making good progress on partnership program development (r = 0.384, p < 0.01). The consistent pattern of correlates in table 3 suggests that NNPS plays a role in guiding district leaders to increasingly take action to assist their schools in implementing research-based approaches to program development with the goal of increasing the number and diversity of families who become engaged in their children’s education.

Table 3. Constellation of Correlates of Years of Membership in NNPS with District Leadership, Facilitation, Evaluation, Support, and Schools’ Progress on Partnerships
Click for larger view
View full resolution
Table 3.

Constellation of Correlates of Years of Membership in NNPS with District Leadership, Facilitation, Evaluation, Support, and Schools’ Progress on Partnerships


We drew from organizational learning and leadership theories to study the development of school-based programs of family and community engagement. The study extended prior research with new evidence that school and district attention to partnership program development is associated with outreach to families, responses from parents, and results for students.

Independent reports from school-based ATPs and from district leaders identified a likely progression in program development. Results suggest that strong principal support and active guidance from district leaders—for example, in forming teams and writing plans—helped school teams establish their basic partnership programs. School teams with strong principal support and helpful district leaders progressed to more advanced implementation activities to engage families who often are hard to reach (for example, multicultural families, fathers, volunteers). Stronger school programs increased the prevalence of parents who were partners in education, and schools with more engaged families reported higher rates of students’ average daily attendance. These well-organized, goal-linked partnership programs that increased the involvement of more and different parents helped improve student attendance as well as other academic and behavioral outcomes.

The interactions of school principals, partnership teams, and district Leaders for Partnerships reflected the assumptions of socio-cultural and organizational learning theories and activated the content of the theory of overlapping spheres of influence for school, family, and community partnerships. The nonsurvey measure of district leaders’ years of NNPS membership suggested that these leaders benefited over time from participating in the research-based network. Longer experience was linked to various leadership qualities—active facilitation of school teams, strong collegial support, attention to evaluation—and to reports that more schools were making good progress in developing programs of family and community engagement.

It should be noted that all schools in NNPS are guided by handbooks, tools, evaluation requirements, professional development workshops and webinars, and frequent communications [End Page 212] to help them establish, implement, and continually improve their programs of family and community engagement. District leaders are guided to prepare and motivate school-based teams to organize and improve their school-based programs to engage students’ families and community partners.

The variable “years in NNPS” reminds us that districts and schools join NNPS with different histories and experiences conducting family and community engagement activities. No school—whether an NNPS member or not—is at “ground zero” in connecting with families. All schools conduct at least a few traditional activities (for example, sending home report cards, having some parent-teacher conferences, distributing a school newsletter, conducting a favorite activity such as a spring fair or a family reading night). Most schools, however, do not plan, evaluate, and improve their programs and practices of partnerships in systematic ways. Districts and schools join NNPS to learn to work systematically to organize partnership programs as a component of good school organization and to engage all students’ families—not just a few—as partners in children’s education. Good plans, evaluations, and continuous improvements are needed by district leaders and school teams to change what may be haphazard or random family and community engagement activities (Weiss, Lopez, and Rosenberg 2010) into more effective and equitable partnership programs (Epstein et al. 2009).

This study supports and extends other studies that have linked family engagement with improved student attendance (Sheldon 2007; Sheldon and Jung, 2015). Student attendance is a leading indicator of student learning, achievement test scores, and graduation from high school (Balfanz et al. 2007; Mac Iver and Messel 2013). This study suggests that systematic efforts by district leaders and school teams to engage families in their children’s education can help, albeit indirectly, to improve attendance.

Countless other studies, reviews, and meta-analyses have shown that goal-linked family engagement significantly improves student learning in specific academic subjects across the grades (Catsambis 2001; Fan and Chen 2001; Galindo and Sheldon 2012; Jeynes 2003, 2012; Sénéchal and LeFevre 2002; Van Voorhis 2011; Van Voorhis et al. 2013). This study extends knowledge by identifying a sequence of program development actions that help schools strengthen their programs to engage more and different families and contribute to results for students.

Tables 1 and 2 raise an interesting question about whether schools face challenges in engaging subgroups of families. Results show that the percentage of students who are English language learners is not significantly associated with schools’ basic or advanced program implementations. Results also show that the measure of poverty—the percentage of students eligible for free or reduced-price meals—is a more serious challenge for programs of family engagement. There are typically fewer ELL students in schools—under 10 percent, except in a few states—than students in poverty, who make up about 48 percent of the student body in the average school in the United States (National Center for Education Statistics 2015) and over 65 percent, on average, in this sample.

Federal law (U.S. Department of Education 2015) requires educators to communicate with all families in languages they understand. Most districts and schools in NNPS are taking this regulation seriously and report that immigrant and refugee families are responsive to outreach activities (Epstein and Ames 2016; Sheldon and Ames 2016). Having relatively few ELL students and specific goals to communicate with families who speak languages other than English at home may help to explain why the ELL variable is not a significant determinant of schools’ partnership programs.

By contrast, schools with a large number of students in families with low income often face many challenges and are themselves underresourced. Educators may be challenged to give attention to family and community engagement at the same time that they are working to improve academic, behavioral, and health-related programs for children in poverty. This is understood, but the most important coefficients in table 1 show that with all other variables statistically controlled, schools that serve diverse students in any community can organize basic and advanced partnership [End Page 213] programs if they have strong principal support for family engagement and helpful assistance from district leaders.

The results of this study are of interest because, across the country, districts and schools are becoming more diverse economically, racially, culturally, and linguistically (Fortuny and Hernandez 2010). Most educators still struggle to communicate with and engage all students’ families at all grade levels (Markow and Pieters 2012). At the same time, most parents still struggle with remaining engaged in their children’s education at each grade level to help them do their best in school (Harris and Robinson, this issue). Schools in this study served diverse populations of students and families, with the majority in high-poverty communities. It is important to note that the schools that did more to apply research-based structures and processes to their work on partnerships were more likely to conduct basic and advanced activities that engaged more families in their children’s education, regardless of the families’ demographics.5


Although this study extends knowledge on partnership programs with new analyses of nested data from schools and their district leaders, it has limitations that need attention in future research. The limited number of districts in the sample seriously restricted the HLM analyses at the district level in table 1 to one key variable—district facilitation of schools. Future studies with larger samples of districts will be able to more fully explore the dynamics of district leaders’ influence on schools’ partnership programs. Table 3 introduced other measures of district leadership that may affect the quality of schools’ programs of family and community engagement.

The data in this study were cross-sectional. It was possible to frame analyses with measures that represented early and later time periods to explore potential paths of influence from initial basic organization of programs to advanced outreach, to parents’ responses, and to rates of student attendance. However, these were proxies for measures taken over time. The suggested step-wise progress in program development must be tested in future studies with longitudinal school, district, parent, and student data to confirm or correct the reported results. Future longitudinal studies of districts and their schools will permit more complex and better-specified multilevel models to understand the temporal order of actions in the development of district leadership and school programs of family and community engagement.

This study relied on survey data, which, like all research methods, have strengths and weaknesses. One strength, which permitted our HLM analyses, was having data from multiple reporters at the district and school levels. Another strength was having data at the school level on the percentage of engaged parents and rates of average daily student attendance, which extended our ability to focus on the results of partnership programs beyond prior studies. We also introduced an independent measure, years in NNPS, to explore how the duration of their efforts affected district leaders’ support for and guidance of schools’ partnership programs. Still, the reported results would be strengthened by nonsurvey data, such as site visits for independent observations of district leaders’ work with schools and action team meetings, or in-depth interviews with purposeful samples of district leaders, school principals, teachers, parents, and students to confirm or refute the survey results.6 Similarly, the survey data on students’ average daily attendance would be stronger if official records on attendance were collected, and the scope of the study could also be enlarged with data on student achievement, behavior, and other indicators of students’ success in school.

The study worked to minimize shared reporter bias and mono-source bias (Spector 2014) [End Page 214] with independent reports from multiple reporters—namely, district leaders and school-based ATPs. This permitted us to check whether school reports about district leaders’ assistance matched the reports from the district leaders themselves. These features of the multilevel sample strengthened the credibility of results indicating the importance of district leadership for schools’ progress on partnerships. However, although districts and schools in this study varied in their demographics and the duration of their work on partnerships, they all were active members of NNPS. Thus, they could carry a shared bias to emphasize their positive work on partnerships.

Future studies should guard against shared reporter bias by collecting multiple sources of information (for example, school records of parent participation in parent-teacher conferences, workshops, or other events) to double-check school and district reports on the extent of parents’ participation. Comparisons of the patterns of program development in NNPS-member schools with non-NNPS sites will clarify whether research-based approaches result in more effective and more equitable partnership programs. Multiple reporters, multiple records, case and control situations, and useful artifacts will minimize the biases that are inherent in survey data and help to validate or refute the findings of this study.


At the time of the EEO report, family involvement was reported for and expected from those with high income and more formal education. This study suggests that family involvement is not a prescribed or “fixed” behavior, but a matter of school and district organization to promote equitable connections between the home, the school, and the community that benefit more—or all—students. Our results show that when schools and districts successfully plan and implement programs of partnerships, they can change old patterns that limit involvement to some parents and engage more and different families in children’s education.

We started research on the design, development, and implementation of partnership programs in the early 1980s by recognizing the social fact repeated in EEO and other studies in the 1960s and1970s that children whose parents are involved in their education tend to do better in school. If families are so important in children’s education, we asked, how can all schools engage all families in ways that increase student success in school? How can all schools engage those families with fewer educational and economic advantages in their children’s education across the grades? And in schools that succeed in engaging these families, do the families respond and are there improvements in their children’s attendance, achievement, and other indicators of success in school?

Although there is value in ongoing research on what parents do on their own to increase their children’s learning and development, such studies are likely to continue to report that some parents are engaged and others are confused or waiting for guidance on how to be productively engaged in their children’s education. We believe that it is critical to extend research on whether and how district and school leadership and programs of partnership increase the number and diversity of involved parents and whether and how their engagement affects student success in school. This study suggests that when effective and equitable school organizational practices are in place, more parents become involved and students benefit.

The results of this study have implications for policy and practice. Education policy is not an end in itself. In complex, multilevel systems, an official policy cannot be enacted without establishing a leadership structure, professional development, a budget, evaluations, incentives, and consequences. This study suggests that even though it may be necessary to have a policy on the books to encourage attention to parental involvement, it is not enough to tell schools or districts just to do something to engage families. Rather, it is critical to have knowledgeable leaders, research-based structures and processes, and strong content in place at the school and district levels to establish and improve plans and practices that promote more equitable and meaningful partnerships with all students’ families. When policy is accompanied by factors that support enactment, more schools do more to engage all [End Page 215] families in goal-linked practices that may contribute to improved attendance and student success in school.

Since the time of the EEO report—and indeed, spurred by it—sociologists, psychologists, and education researchers have built a field of study that goes well beyond the “contest mentality” of the 1960s and 1970s. The question once phrased as a competition—which is more important, the family or the school?—is no longer useful and instead has become one about an ongoing research agenda: what are the contributions of collaborative relationships among schools, families, and the community to student learning and development?

There are some in research and in education who still hold the old view that family engagement is about the parents and that it is up to parents to get involved—or not—in their children’s education. This view omits the concept of partnership and ignores the benefits of a strong agreement among educators, parents, and policy leaders that education is a shared responsibility of home, school, and community. NNPS aims to help districts and schools understand that partnerships are a means to help more students succeed and as such should be part of school organization—not external to the school. With this view, it is possible for districts and schools to fulfill the intent of written policies for parental involvement with effective actions that promote more equal educational opportunities for more students.

It is a social fact that families are important in children’s lives. The new question that we pose for research, policy, and practice focuses on whether it is possible to change the distribution of involved families from a few economically advantaged families to all families, so that more students benefit from family support, encouragement, and participation in their education. This study suggests that this challenge is being met, incrementally, by districts and schools that apply research-based approaches to organize and continually improve programs and practices of family and community engagement as an essential component of school organization.

Joyce L. Epstein

Joyce L. Epstein is research professor of education and sociology at Johns Hopkins University and director of the Center on School, Family, and Community Partnerships.

Steven B. Sheldon

Steven B. Sheldon is associate professor of education at Johns Hopkins University and associate director of the Center on School, Family, and Community Partnerships.

Direct correspondence to: Joyce Epstein at, 2701 N. Charles St., Suite 300, Baltimore, MD 21218; and Steven Sheldon at, 2701 N. Charles St., Suite 300, Baltimore, MD 21218.


Balfanz, Robert, Liza Herzog, and Douglas J. Mac Iver. 2007. “Preventing Student Disengagement and Keeping Students on the Graduation Path in Urban Middle-Grades Schools: Early Identification and Effective Interventions.” Educational Psychologist 42(4): 223–35.
Bryk, Anthony S., Louis M. Gomez, Alicia Grunow, and Paul G. LeMahieu. 2015. Learning to Improve: How America’s Schools Can Get Better at Getting Better. Cambridge, Mass.: Harvard Education Press.
Bryk, Anthony S., Penny B. Sebring, Elaine Allen-sworth, Stuart Luppescu, and John Q. Easton. 2010. Organizing Schools for Improvement: Lessons from Chicago. Chicago: University of Chicago Press.
Catsambis, Sophia. 2001. “Expanding Knowledge of Parental Involvement in Children’s Secondary Education: Connections with High School Seniors’ Academic Success.” Social Psychology of Education 5(2): 149–77.
Coleman, James S., Ernest Q. Campbell, Carol J. Hobson, James M. McPartland, Alexander M. Mood, Frederick D. Weinfeld, and Robert L. York. 1966. Equality of Educational Opportunity. Washington: U.S. Department of Health, Education, and Welfare, Office of Education.
Cowan, Kristen Tosh. 2003. “Parental Involvement.” In The New Title I: The Changing Landscape of Accountability by Kristen Tosh Cowan. Washington, D.C.: Thompson Publishing.
Elmore, Richard F. 2004. School Reform from the Inside Out: Policy, Practice, and Performance. Cambridge, Mass.: Harvard Education Publishing Group.
Epstein, Joyce L. 1987. “Toward a Theory of Family-School Connections: Teacher Practices and Parent Involvement.” In Social Intervention: Potential and Constraints, edited by Klaus Hurrelmann, Franz-Xaver Kaufmann, and Friedrich L. Lösel. New York: DeGruyter.
———. 2008. “Research Meets Policy and Practice: How Are School Districts Addressing NCLB Requirements for Parental Involvement?” In No Child Left Behind and the Reduction of the Achievement Gap: Sociological Perspectives on Federal Educational Policy, edited by Alan R. Sadovnik, Jennifer O’ Day, George Bohrnstedt, and Kathryn Borman. New York: Routledge. [End Page 216]
———. 2011. School, Family, and Community Partnerships: Preparing Educators and Improving Schools. 2nd ed. Boulder, Colo.: Westview Press.
———. Forthcoming. “Creating a National Network on School, Family, and Community Partnerships: Multi-level Goals, Challenges, and Successes.” In Proceedings of the Projecto ESCXEL—Rede de Escolas de Excelência. First International Conference, New University of Lisbon, Portugal (November 2014).
Epstein, Joyce L., and R. Tyler Ames. 2016. Annual NNPS Report: 2015 District Data. Baltimore: Johns Hopkins University, Center on School, Family, and Community Partnerships.
Epstein, Joyce L., Claudia Galindo, and Steven B. Sheldon. 2011. “Levels of Leadership: Effects of District and School Leaders on the Quality of School Programs of Family and Community Involvement.” Educational Administration Quarterly 47(3): 462–95.
Epstein, Joyce L., and Steven B. Sheldon. 2006. “Moving Forward: Ideas for Research on School, Family, and Community Partnerships. In SAGE Handbook for Research in Education: Engaging Ideas and Enriching Inquiry, edited by Clifton F. Conrad and Ronald C. Serlin. Thousand Oaks, Calif.: Sage Publications.
Epstein, Joyce L., et al. 2009. School, Family, and Community Partnerships: Your Handbook for Action. 3rd ed. Thousand Oaks, Calif.: Corwin.
Fan, Xitao, and Michael Chen. 2001. “Parental Involvement and Students’ Academic Achievement: A Meta-analysis.” Educational Psychology Review 13(1): 1–22.
Fortuny, Karina, and Donald J. Hernandez. 2010. “Characteristics of Children of Immigrants.” Paper presented to the conference “Young Children in Immigrant Families and the Path to Educational Success.” Urban Institute, Washington, D.C. (June 28).
Galindo, Claudia, and Steven B. Sheldon. 2012. “School and Home Connections and Children’s Kindergarten Achievement Gains: The Mediating Role of Family Involvement.” Early Childhood Research Quarterly 27(1): 90–103.
Harris, Angel L., and Keith Robinson. 2016. “A New Framework for Understanding Parental Involvement: Setting the Stage for Academic Success.” RSF: The Russell Sage Foundation Journal of the Social Sciences 2(5). doi: 10.7758/RSF.2016.2.5.09.
Honig, Meredith I. 2006. “Street-Level Bureaucracy Revisited: Frontline District Central Office Administrators as Boundary Spanners in Education Policy Implementation.” Educational Evaluation and Policy Analysis 28(4): 357–83.
———. 2008. “District Central Offices as Learning Organizations: How Sociocultural and Organizational Learning Theories Elaborate District Central Office Administrators’ Participation in Teaching and Learning Improvement Efforts.” American Journal of Education 114(4): 627–64.
Huber, George P. 1991. “Organizational Learning: The Contributing Processes and the Literatures.” Organizational Science 2(1): 88–115.
Hutchins, Darcy J., and Steven B. Sheldon. 2013. Annual Report: 2012 School Update Data. Baltimore: Johns Hopkins University. Center on School, Family, and Community Partnerships.
Jeynes, William H. 2003. “A Meta-analysis: The Effects of Parental Involvement on Minority Children’s Academic Achievement.” Education and Urban Society 35(2): 202–18.
———. 2012. “A Meta-analysis of the Efficacy of Different Types of Parental Involvement Programs for Urban Students.” Urban Education 47(4): 706–42.
Knapp, Michael S. 2008. “How Can Organizational and Sociocultural Learning Theories Shed Light on District Instructional Reform?” American Journal of Education 114(4): 521–39.
Leithwood, Kenneth, and Nona A. Prestine. 2002. “Unpacking the Challenges of Leadership at the School and District Level.” In The Educational Leadership Challenge: Redefining Leadership for the 21st Century, edited by Joseph Murphy. Chicago: University of Chicago Press.
Louis, Karen Seashore. 2008. “Learning to Support Improvement: Next Steps for Research on District Practice.” American Journal of Education 114(4): 681–89.
Mac Iver, Martha A., and Matthew Messel. 2013. “The ABCs of Keeping on Track to Graduation: Research Findings from Baltimore.” Journal of Education for Students Placed at Risk 18(1): 50–67.
Markow, Dana, and Andrea Pieters. 2012. The MetLife Survey of the American Teacher: Teachers, Parents, and the Economy. New York: MetLife, Inc. [End Page 217]
Mayrowetz, David. 2008. “Making Sense of Distributed Leadership: Exploring the Multiple Usages of the Concept in the Field.” Educational Administration Quarterly 44(3): 424–35.
National Center for Education Statistics (NCES). 2015. “Fast Facts: Educational Institutions.” Digest of Educational Statistics 2013 (NCES 2015-011). Available at: (accessed October 12, 2015).
Raudenbush, Steven, and Anthony Bryk. 2002. Hierarchical Linear Models: Applications and Data Analysis Methods. Thousand Oaks, Calif.: Sage Publications.
Sanders, Mavis G. 2008. “Using Diverse Data to Develop and Sustain School, Family, and Community Partnerships: A District Case Study.” Education Management, Administration, and Leadership 36(4): 530–45.
———. 2009. “Collaborating for Change: How an Urban School District and Community-Based Organization Supports and Sustains School, Family, and Community Partnerships.” Teachers College Record 111(7): 1693–1712.
Sanders, Mavis G., and Steven B. Sheldon. 2009. Principals Matter: A Guide to School, Family, and Community Partnerships. Thousand Oaks, Calif.: Corwin Press.
Sénéchal, Monique, and Jo-Anne LeFevre. 2002. “Parental Involvement in the Development of Children’s Reading Skill: A Five-Year Longitudinal Study.” Child Development 73(2): 455–60.
Senge, Peter. 1990. The Fifth Discipline: The Art and Practice of the Learning Organization. New York: Doubleday.
Sheldon, Steven B. 2005. “Testing a Structural Equations Model of Partnership Program Implementation and Parent Involvement.” The Elementary School Journal 106(5): 171–87.
———. 2007. “Improving Student Attendance with a School-Wide Approach to School-Family-Community Partnerships.” Journal of Educational Research 100(5): 267–75.
———. 2008. “Getting Families Involved with NCLB: Factors Affecting Schools’ Enactment of Federal Policy.” In No Child Left Behind and the Reduction of the Achievement Gap: Sociological Perspectives on Federal Educational Policy, edited by Alan R. Sadovnik, Jennifer O’ Day, George Bohrnstedt, and Kathryn Borman. New York: Routledge.
Sheldon, Steven B., and R. Tyler Ames. 2016. Annual NNPS Report: 2014 School Data. Baltimore: Johns Hopkins University, Center on School, Family, and Community Partnerships.
Sheldon, Steven B., and Sol Bee Jung. 2015. “Exploring How School-Family Partnerships Improve Attendance: Principals, Teachers, and Program Organization.” Paper presented to the annual conference of the American Educational Researchers Association. Chicago (April).
Spector, Paul E. 2014. “Survey Design and Measure Development.” In Oxford Handbook of Quantitative Methods in Psychology, edited by Todd D. Little. Oxford: Oxford University Press.
Spillane, James P., and John B. Diamond. 2007. Distributed Leadership in Practice. New York: Teachers College Press.
Stein, Mary Kay, and Cynthia E. Coburn. 2008. “Architectures for Learning: A Comparative Analysis of Two Urban School Districts.” American Journal of Education 114(4): 583–626.
Supovitz, Jonathan A. 2006. The Case for District-Based Reform. Cambridge, Mass.: Harvard University Press.
Thomas, Brenda G., Marsha D. Greenfeld, Courtney R. Sender, and Joyce L. Epstein. 2015. Promising Partnership Practices 2015. Baltimore: Johns Hopkins University, National Network of Partnership Schools.
U.S. Department of Education. 2015. Elementary and Secondary Education Act (ESEA), also known as Every Student Succeeds Act (ESSA). Public Law 114-95. Washington: U.S. Government Printing Office.
Van Voorhis, Frances L. 2011. “Costs and Benefits of Family Involvement in Homework.” Journal of Advanced Academics 22(2): 220–49.
Van Voorhis, Frances L., Michele Maier, Joyce L. Epstein, and Chrishana M. Lloyd. 2013. The Impact of Family Involvement on the Education of Children Ages 3–8. New York: MDRC.
Van Voorhis, Frances L., and Steven B. Sheldon. 2004. “Principals’ Roles in the Development of U.S. Programs of School, Family, and Community Partnerships.” International Journal of Educational Research 41(1): 55–70.
Weick, Karl E. 1995. Sensemaking in Organizations. Thousand Oaks, Calif.: Sage Publications.
Weiss, Heather, M. Elena Lopez, and Heidi Rosenberg. 2010. Beyond Random Acts: Family, School, [End Page 218] and Community Engagement as an Integral Part of Education Reform. Cambridge, Mass.: Harvard Family Research Program. Information available at: (accessed October 6, 2015).
Wenger, Etienne. 1998. Communities of Practice: Learning, Meaning, and Identity. Cambridge: Cambridge University Press. [End Page 219]


1. Section 1010 on “parent and family engagement” specifies that districts ([local education agencies] LEAs) must have a policy and must “(B) provide the coordination, technical assistance, and other support necessary to assist and build the capacity of all participating schools in planning and implementing effective parent and family involvement activities to improve student academic achievement and school performance” (U.S. Department of Education 2015, emphasis added).

2. See the award-winning sites in the section on “Success Stories” at

3. The authors thank Sol Bee Jung, graduate student at Johns Hopkins University’s School of Education, for her assistance with these analyses.

4. In addition to the district facilitation scale explained in the section on measures and used in the HLM analyses in table 1, measures explored in table 3 include a fourteen-item leadership scale (α = 0.66) on efforts to establish an office for partnerships; a seven-item emphasis on evaluation scale (α = 0.66) on whether district leaders evaluated their own and schools’ partnership programs; and a thirteen-item collegial support scale (α = 0.84) on cooperation for partnership program development from district, school, family, and community members (Epstein and Ames 2016).

5. See evidence of how these processes work in economically advantaged and disadvantaged districts and schools that serve families with diverse socioeconomic, cultural, racial, and linguistic backgrounds in Thomas et al. (2015) and in “Success Stories” at the NNPS website,

6. See Sanders (2008, 2009) for prior qualitative studies on how district leaders’ facilitative actions helped ATPs improve the quality of their partnership programs and engagement of families.

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.