Teachers’ Social Networks,College-Going Practices,and the Diffusion of a School-Based Reform Initiative

Abstract

This study examines the association between high school teachers’ collegial networks and their own familiarity with, and practices related to, supporting students’ admission to college. Using survey data collected at two time points from 163 teachers in four mid-Michigan high schools in 2011-2012, this study (a) employs cluster analysis to map collegial networks by school, and (b) uses social network influence modeling to examine how teachers may be influenced by their closest colleagues. Results of the cluster analysis find that all four high schools surveyed show evidence of distinct clustering by subgroups. Results from influence modeling show that while positive in almost all cases, the impact of teachers’ exposure to colleagues is not significantly associated with a change in familiarity or practices related to college-going. However, interaction with program staff from a college-advising program is found to be related to a small positive change in some teacher practices.

Keywords

sociology of education sociology social sciences social network analysis teaching education higher education organizations occupation work sociology of work

Introduction

Many students dream of careers that require a college degree. Although these students may realize that their career of choice requires attending college, many are unfamiliar with factors relating to the college-going process. According to Schneider and Stevenson (1999), approximately 60% of high school students have beliefs about college that do not align with their career goals. That is, they do not understand how much education they will need to pursue a given career or which colleges might offer them the best opportunities to do so. Low-income students seem to be particularly at-risk for such a misalignment between educational expectations and occupational aspirations (Schneider, 2015; Schneider & Stevenson, 1999). Whereas students growing up in middle-class or affluent families have more exposure to people who have attended college (e.g., parents, neighbors) as well as greater access to private tutors and/or college-entrance exam preparatory materials (Schneider & Stevenson, 1999), students in low-income families may lack these forms of capital.

Furthermore, low-income students may not have the resources to visit colleges and may be unaware of many of the nuances of applying for financial aid and matriculating (see, for example, Schneider, 2015). Given that attending college is associated with greater lifetime income, qualification for more prestigious jobs, and better health outcomes (Baum, Ma, & Payea, 2013; Cohen & Syme, 2013; National Science Board, 2007; Oreopoulos & Petronijevic, 2013), promoting college-going for low-income and otherwise underrepresented students is critical as we strive to become a more equitable nation. Indeed, recent reports from both the National Academies of Sciences, Engineering, and Medicine (2016) and the National Research Council (2011) stress the need for reducing many of the barriers that low-income and underrepresented students face in pursuit of in-demand science, technology, engineering, and mathematics (STEM) careers and educational opportunities.

Borne out of a desire to address these inequities, the College Ambition Program (CAP) is a whole-school intervention program intended to promote a culture of college-going in high schools with low rates of college enrollment—typically those with high proportions of low-income students (Schneider, 2015; Schneider, Broda, Judy, & Burkander, 2013; Schneider, Judy, Mazuca, & Broda, 2014). CAP has been implemented in multiple schools and provides resources and support to students, staff, parents, and community stakeholders. The CAP model was developed based on empirical research on transitions to adulthood (Csikszentmihalyi & Schneider, 2000; Schneider, 2007; Schneider & Stevenson, 1999), which found that an adolescent’s decision to enroll in a postsecondary institution was associated with being able to visualize oneself as a college student, transform interests into realistic actions, and create strategic plans. CAP centers are typically staffed by graduate students in education who work with students and school personnel to increase awareness about the college-going process. The program is composed of four components designed to improve the college-going culture of the school. The four major components of CAP are (a) course counseling, (b) financial aid advising, (c) mentoring and tutoring, and (d) college visits.

Ultimately, the goal of CAP is to increase college-going in high-school students. Several journal articles, book chapters, and other publications describe the impact of CAP in this respect (e.g., Schneider, 2015; Schneider et al., 2013; Schneider et al., 2014), with results suggesting a moderate but meaningful positive impact. However, another important consideration is that CAP is intended to be a sustainable program that persists in schools after researchers complete the initial trials (Schneider, 2015). If this program, and others like it, are to continue long term, school personnel need to be familiar with the program’s components and deeply invested in its success. Without such a community of practitioner buy-in, CAP may not endure, regardless of its success in advancing student college-going (see, for example, Bryk & Schneider, 2002). Therefore, better understanding patterns of the diffusion of knowledge related to college-going practices among teachers and other school personnel are central to the long-term tenability of CAP. In the current study, we used social network analysis and influence modeling to examine the impact of CAP on teachers’ beliefs, practices, and familiarity related to the college-going process. Specifically, this study analyzed teachers’ collegial networks, their levels of interaction with CAP staff, and the impact of both on the flow of knowledge and practices that promote college-going. In the subsequent sections, we review how social network theory has been used to investigate educational reform initiatives, describe literature pertinent to developing a college-going culture in high schools, and advance a theory regarding how teachers’ collegial networks relate to the development of this culture.

Social Network Theory in Education

Social network theory is a by-product of social capital theory (Bourdieu & Wacquant, 1992; Coleman, 1988), which seeks to understand patterns and structures within organizations through which social capital is transmitted (Burt, 2000; Frank, Zhao, & Borman, 2004; Moolenaar, 2012). Three assumptions that underpin social network theory are that (a) people who have relationships with one another exchange resources, (b) people within a network are interdependent, and (c) social networks both provide opportunities for action and constrain action (Degenne & Forse, 1999). Relatively recently, scholars have begun to employ social network theory to frame interactions within educational networks (e.g., Bidwell & Yasumoto, 1999; Daly, 2012; Daly, Moolenaar, Bolivar, & Burke, 2010; Frank et al., 2004; Penuel, Riel, Krause, & Frank, 2009; Siciliano, 2016; Spillane, Shirrell, & Sweet, 2017). Although these researchers have investigated a wide range of educational issues, they share the belief that social interactions between teachers, administrators, and other stakeholders are critical to understanding and reforming our educational system.

In theory, social networks have no a priori bounds. Teachers can easily share ideas with each other across district, state, and even national borders. However, researchers often impose boundaries on the networks they examine, often at the school or district level (Frank et al., 2004; Moolenaar, 2012). Indeed, recent research suggests that physical proximity between people within a network plays a role in the diffusion of knowledge (Spillane et al., 2017), so imposing boundaries on networks may be a way to capture close-proximity relationships.

Furthermore, the phenomena of interest to researchers are often themselves bounded within certain social structures. School reforms are one such phenomenon. That is, reforms are often implemented at the school level, and within-school networks are likely most relevant for studying these reform efforts. A growing body of research suggests that social networks within schools play a meaningful role in the success or failure of school-wide reform efforts (e.g., Daly, 2012; Daly, Liou, & Moolenaar, 2014; Daly et al., 2010; Penuel et al., 2009). In a contrasting case study of two elementary schools, Penuel and colleagues (2009) found that, despite having similar financial resources and principals devoted to reform, the differences in the social structures of the two schools led to differences in the successes of their reform efforts. The school with shared commitment toward reform and a more connected social structure was more successful in meeting its goals, whereas the school with a more fragmented social structure was less successful. Similarly, in a mixed-methods case study examining English Language Arts (ELA) curriculum and instructional reform in five underperforming elementary schools, Daly and colleagues (2010) found considerable differences in the schools’ social network structures, which was associated with differences in the enactment of reform efforts. The authors found that not only was the density of instrumental interactions (e.g., collaborative lesson planning, reviewing student data together) positively associated with progress toward reform and self-efficacy beliefs, but density of affective interactions was also related to positive outcomes. These findings suggest that various types of interactions within teacher networks are related to reform efforts.

Although often used to study reform efforts related to improving student achievement (e.g., Daly et al., 2010; Penuel et al., 2009), social network analysis allows researchers to study other types of educational reform as well. For example, Frank et al. (2004) found that school social network patterns, such as access to people with expertise, influenced the extent to which teachers implemented new instructional technologies in their classrooms. Teachers who were encouraged by peers to incorporate new technologies and who had ties to people proficient with those technologies more often implemented them in their own classrooms. In a literature review exploring teachers’ use of data, Daly (2012) noted that teachers often analyze and interpret data with colleagues and are, thus, influenced by patterns of collegial relationships. Collectively, these studies suggest that teacher and administrator social networks can facilitate or obstruct a plethora of educational initiatives. In this study, we further explore the relationship between social network structures and the implementation of an intervention designed to improve the college-going culture of schools.

College-Going Interventions and School-Wide College Culture

Recent reports indicate that people who complete a college degree earn higher wages, have higher wage growth, experience lower rates of unemployment, and enjoy more job stability than people who do not (Baum et al., 2013; Buddin, 2012; Day & Newburger, 2002). However, not all schools are able to ensure that students have access to the support they need to prepare for college while still in high school. More specifically, schools with majority populations of historically underrepresented groups may lack such supports. While White college-going rates have steadily increased since the 1970s, members of historically underrepresented groups have not seen a similar increase. Despite the presence of college preparation programs throughout the country for several decades, these trends have not changed significantly (Gandara & Bial, 2001; Perna & Titus, 2005). These programs typically offer some combination of mentoring, tutoring, course counseling, and financial advising, but they vary in implementation from programs that target specific underrepresented groups to school-wide interventions that are available to the entire student body. Evaluations of many of these programs are mixed. While most college outreach programs show modest effects, school-wide interventions seem to be more effective than those that target individual students (Domina, 2009). In a review of research on four widely used college outreach programs (Gear Up, Talent Search, Upward Bound, and Quantum Opportunities Program), Domina (2009) suggests that larger, school-wide programs have more potential for “spillover effect” (p. 127), which could mean that unmotivated students (the most difficult group to impact) are more likely to experience changes in belief or actions related to college. In contrast, most targeted interventions identify students who are motivated to attend college but may lack enough knowledge or resources, effectively leaving a portion of students out of the treatment.

School-wide interventions have the potential to be more effective than targeted programs because they can leverage school-wide college-going culture in supporting students to manage the college search, application, and matriculation process. Particularly for students whose parents did not attend college, the school resources dedicated to college guidance are critical supports. Often, these students rely heavily on school networks, where administrators, counselors, and teachers with college experience can supplement the students’ own family or local networks with limited college experience (Bryan, Farmer-Hinton, Rawls, & Woods, 2017; Farmer-Hinton, 2008). School size and location, course offerings and tracks, and racial and socioeconomic status of students together impact the college-going culture of the school, which is reinforced by counselors and teachers (Holland & Farmer-Hinton, 2009). Other factors include district expectations and priorities regarding college-going, relationships that counselors have with postsecondary institutions, and the prevalence of counselors dedicated solely to college preparation rather than scheduling, discipline, or other administrative tasks (McDonough, 1997, 2005). Several studies suggest that the presence of a strong college-going culture at the high school is an important factor for first-generation students, as well as minority students and students from low-income backgrounds (Farmer-Hinton, 2008; O’Connor, 2000; Plank & Jordan, 2001).

While most research on promoting college access for low-income youth is based on work in urban settings, rural communities can also face similar issues. Research suggests that rural students’ educational and occupational ambitions are often lower than those of students in nonrural settings (Chenoweth & Galliher, 2004; Cobb, 1989), which may lead to lower expectations of educational attainment (Cobb, 1989). Students in rural schools also face several other factors that may also contribute to the development and attainment of lower educational goals including poverty, lack of family and professional role models, lack of self-confidence, valuation of work over education, and lack of support/encouragement from influential individuals (Chenoweth & Galliher, 2004; Cobb, 1989). Furthermore, the occurrence of lower ambitions may be magnified during difficult economic conditions, especially in rural communities when top students leave for college, resulting in a perceived brain drain (Sherman & Sage, 2011).

The Role of Teacher Networks in School-Wide Reforms

Although school counselors are often seen as those primarily responsible for facilitating the college-going process in schools, the literature is consistent that they are not the only ones providing college guidance to students and encouraging (or discouraging) their ambitions. This may be due to the notably high ratio of students to school counselors in schools today. While the American School Counselor Association’s suggested ratio is 250 students to one counselor, the national average in 2015-2016 was 464:1 (American School Counselor Association, 2018). Ideally, counselors can devote time to meet with all students individually and provide advice on course selection, college applications, scholarships, and financial aid. However, as a counselor’s caseload increases, the feasibility of significant one-on-one contact decreases. McDonough (1997) suggests that this case overload, coupled with extra administrative responsibilities such as discipline and scheduling, mean that few counselors get to spend much one-on-one time with students. Consequently, students often turn to teachers for this support. Students who successfully transition to college often credit one or two helpful teachers or school personnel who guided them through the process (Farmer-Hinton, 2008; Freeman, 1997; Levine & Nidiffer, 1996). However, little is known about how teachers get the knowledge needed to support these students and how resources to support students are shared among colleagues within a school.

Because teachers play a critical role in implementing school-level reforms, particularly those that target school culture, one must consider how to measure the influence of school-wide reform initiatives on teachers’ practices and beliefs, as well as the extent to which teachers are influenced by their respective collegial networks. Previous research has identified multiple factors that contribute to how teachers respond to reforms, including organizational structure and dissemination of information (Bryk & Schneider, 2002; Daly et al., 2010), professional development (Kearns et al., 2010; Wilson, 1990), and ideological alignment with reform initiatives (Kennedy, 2005; Olsen & Sexton, 2009). While we acknowledge the influence of all these factors, this study focuses on knowledge, norms, and behaviors (specifically, college-going knowledge and behaviors), given that recent research suggests that teachers adapt reforms through experimentation and by drawing on what is learned from colleagues (Frank, Zhao, Penuel, Ellefson, & Porter, 2011).

Influence modeling

Drawing on educational research employing social network theory (e.g., Cobb, 1989; Daly, 2012; Daly et al., 2010; Frank et al., 2004; Moolenaar, 2012; Spillane et al., 2017), this study argues that knowledge in schools is transmitted in networks of teachers, often in the form of knowledge flow from teachers with high expertise in an area to those with less expertise. Thus, an influence model (Frank, 1998) may be the best way to understand the diffusion of information within collegial networks in CAP schools. An influence model is a type of statistical model that includes a network effect (the knowledge and/or behaviors of one’s colleagues) as a potential predictor of an individual’s own knowledge and behaviors. Given that teachers will likely have varying levels of knowledge about college-going strategies and differing mindsets about their students’ ability to succeed in college, we used influence modeling to better understand how teachers’ interactions with colleagues and CAP program staff influenced their beliefs and practices over time.

While teachers’ beliefs and practices are not the direct outcome of interest of the CAP intervention, teachers have a role in how the intervention evolves in each of the treatment sites. How teachers interact with their closest colleagues, their administrators, and with the CAP intervention may influence how the intended enhancement of the college-going culture for each treatment school evolves. A shared focus and vision among teachers created through a strong foundation of trust also contributes to the effectiveness of school-wide reforms (Bryk & Schneider, 2002). The structure of these networks also has considerable bearing on how CAP might persist in schools over time. The design of the CAP intervention was built on these principles, which link the complex nature of reform implementation in a school-wide setting and the critical role that teacher social networks play in proposed changes to their school culture.

Research Questions

Because a primary goal of the CAP model was to support schools in developing a strong college-going culture, we hypothesized that teachers and their social context would play a critical role in culture change. This hypothesis drove the following research questions:

Research Question 1: What are the underlying collegial social structure(s) within CAP high schools?

Research Question 2: Given these social structures, how do teachers’ colleagues influence their own practices, familiarity, and beliefs related to college-going?

Research Question 3: How do teachers’ interactions with CAP staff influence their own practices, familiarity, and beliefs related to college-going?

Theory and Hypotheses

CAP was designed to be a collective intervention, based on the diffusion of knowledge and practices (Broda, 2015; Rogers, 2003; Schneider, 2015; Schneider et al., 2013). Using CAP as a hub for resources about college, the program aimed to distribute knowledge and resources that impact the college-going mindsets of both students and teachers. It had been observed at the CAP treatment schools that within a faculty, teachers possess varying attitudes and levels of information related to the college search, and application and selection processes. We, therefore, hypothesized that teachers with similar mindsets would cluster together (homophily), creating subgroups within the faculty that were supportive of the aims of our project and the college-going culture at each school. Furthermore, we hypothesized that, over time, teachers should be influenced by the behaviors and practices of their peers, resulting in a positive relationship between teacher outcomes and those of their collegial grouping. Finally, interactions with CAP staff should have driven growth in teachers’ college-going mindsets and practices. Therefore, we hypothesized that teachers who interacted more often with CAP staff should have seen growth in these areas, even after controlling for their Time 1 beliefs and practices.

Method

School Sample

CAP was designed to be a whole-school intervention that influences not only students but also teachers, staff, and families. For the 2011-2012 school year (the year this study was conducted), the CAP intervention was implemented, for the first time, in four public secondary schools (two urban and two rural) in mid-Michigan.¹ Table 1 below provides additional descriptive statistics on the four schools included in the study. The two urban schools were classified as “midsize,” serving between 800 and 1,200 students, were in the same district, and served a racially diverse population of students. The student composition at Middlebrook High School was 40% White, 35% Black, 20% Hispanic, and 5% Other race/ethnicities. At Drew High School, 50% of students were Black, 25% White, 10% Hispanic, and 15% Other race/ethnicities. The rural treatment schools served between 400 and 600 students, more than 95% of whom were White. The urban schools included a large percentage of economically disadvantaged students, with more than 50% of their students eligible for free and reduced price lunch. At the rural schools, 20% of the students were eligible for free and reduced price lunch. In addition, all treatment schools had above-state-average percentages of students who would be the first in their family to go to college. This was a precondition for participation in the CAP intervention. In terms of school curriculum, all four schools offered Advanced Placement coursework, but only Kelly High School offered International Baccalaureate (IB) coursework. Eligible schools were recruited by starting first with school district leadership, and then approaching school principals who then made the ultimate decision of whether or not to participate after consulting with their faculty.

Table 1.

Relevant School Characteristics for Study Sample.

	Urban schools		Rural schools
	Middlebrook	Drew	Scherzer	Kelly
Total Enrollment	800	1,200	400	550
FRPL	56.0%	70.0%	20.0%	20.0%
ELL	18.0%	25.0%	<1%	<1%
Minority	84.0%	75.0%	<1%	3.0%
Offer AP?	Yes	Yes	Yes	Yes
Offer IB?	No	No	No	Yes

Note. FRPL = free or reduced price lunch; ELL = English language learner; AP = Advanced Placement; IB = International Baccalaureate.

Teacher Sample

Table 2 provides additional context on the sample of teachers included in this study, provided by school to facilitate comparison. Teacher experience varied from an average of 10.93 years at Drew High School to an average of 14.05 years at Scherzer High School. Percentage of teachers with 10 or fewer years of experience varied from a low of 35% at Kelly High School to a high of 50% at Scherzer High School. Percentage of teachers with 20 or more years of experience varied from a low of 10% at Drew High School to a high of 32% at Scherzer High School.

Table 2.

Teacher Descriptive Statistics, by School.

Variable	N	M/%	SD	Minimum	Maximum
Drew HS
Years teaching experience	29	10.93	6.61	1	33
10 years or less	29	48%		0	1
20 years or more	29	10%		0	1
MA or higher	29	59%		0	1
Noninstructional role	29	38%		0	1
Science or math teacher	29	21%		0	1
Interact with CAP Staff (Yes)	29	86%		0	1
Female	29	76%		0	1
Non-White	29	3%		0	1
Scherzer HS
Years teaching experience	22	14.05	11.24	2	35
10 years or less	22	50%		0	1
20 years or more	22	32%		0	1
MA or higher	22	60%		0	1
Noninstructional role	22	23%		0	1
Science or math teacher	22	27%		0	1
Interact with CAP Staff (Yes)	22	95%		0	1
Female	21	67%		0	1
Non-White	22	5%		0	1
Middlebrook HS
Years teaching experience	16	13.38	7.50	1	27
10 years or less	16	38%		0	1
20 years or more	16	19%		0	1
MA or higher	16	81%		0	1
Noninstructional role	16	31%		0	1
Science or math teacher	16	31%		0	1
Interact with CAP Staff (Yes)	16	62%		0	1
Female	16	63%		0	1
Non-White	16	13%		0	1
Kelly HS
Years teaching experience	33	13.42	8.14	2	35
10 years or less	34	35%		0	1
20 years or more	34	21%		0	1
MA or higher	34	66%		0	1
Noninstructional role	34	21%		0	1
Science or math teacher	34	26%		0	1
Interact with CAP Staff (Yes)	34	85%		0	1
Female	32	56%		0	1
Non-White	34	15%		0	1

Note. CAP = College Ambition Program; HS = high school; MA = Master’s degree.

The percentage of teachers who reported earning a Master’s degree or higher ranged from 59% at Drew High School to 81% at Middlebrooks. Between 21% and 38% of teachers reported being assigned to a noninstructional role (e.g., content coach, mentor, counselor, administrative intern). Between 21% and 31% of teachers reported teaching a math or science course. Teachers were also asked to report their level of interaction with CAP staff over the course of the year. Response options ranged from “never” to “frequently.” Overall, 90% of teachers reported some interaction with CAP staff. This percentage ranged from 62% at Middlebrook High School to 95% at Scherzer High School.

Finally, in terms of demographic characteristics, proportion of female teachers ranged from a low of 56% at Kelly High School to a high of 76% at Drew High School. Proportion of non-White teachers ranged from a low of 3% at Drew High School to a high of 15% at Kelly High School.

Development of Survey Instruments

Data used in this study came from teachers in the four schools participating in the CAP intervention in 2011-2012 (henceforth, “CAP schools”). The survey instruments were developed using survey items from the Survey of Chicago Public Schools—High School Teacher Edition (UChicago Consortium on School Research, 2009) and the High School Longitudinal Study of 2009 (Ingels et al., 2011) teacher questionnaire with a few items specifically developed to measure each teacher’s interaction with the CAP program (e.g., student referrals, using CAP mentors, participating in professional development). Items used in the teacher questionnaire measured three areas related to the school’s college-going culture: (a) teacher’s knowledge about the process; (b) teacher behaviors (e.g., writing letters of recommendation); and (c) perceptions of school-wide norms. Many of the items included here are part of the “Expectations for Postsecondary Education” subscale of the 5 Essentials survey administered annually by the UChicago Consortium for Chicago School Research. This subscale has been shown to have an internal reliability of α = .79 in previous administrations (Levenstein, 2016) and has been used extensively in previous research. In addition, teachers were asked to list their five closest colleagues in the building and indicate the frequency with which they interact. A sample of the “Wave 1” survey is available in the appendix.

Data Collection

Data were collected from teachers and instructional staff in two waves. In Wave 1, teachers completed a one-page paper survey during the Fall semester of the first year about their prior knowledge, behaviors, and norms related to college-going. The survey was given at the first mandatory staff meeting for each school. Any absent teachers received a follow-up contact and were given additional time to complete the survey. In Wave 2, administered at the end of the Spring semester of the first year, teachers completed an online survey using SurveyMonkey that measured their knowledge, behaviors, and norms related to college-going. In addition, teachers provided background information and teacher network data in the survey administered in Wave 2. Teacher names were coded with a unique identifier and then removed from the survey responses to protect the identity of each participant. The response rate for the first wave was 75% of the total teaching faculty across all schools. For the second wave, the response rate was 70% of all teachers.

Measures

Study outcomes and prior measures: Teacher-reported practices, familiarity, and beliefs

The outcomes in this study fell into three general categories: teachers’ practices related to college-going, their familiarity with components of the college-going process, and their beliefs about themselves and their students related to college-going. Within the practices category, we measured two separate outcomes: PRAC2 and T2_RECLETTERS. The first outcome, PRAC2, represents a score weighted by constituent items’ standardized factor loadings of a teacher’s score at Time 2 on a six-item survey asking how often they engage in a variety of teaching practices related to college-going (see the appendix for a copy of the teacher survey). All practice items were developed on a 1 to 5 scale: 1 = never, 2 = a few times a year, 3 = monthly, 4 = weekly, and 5 = daily. The second practice outcome, T2_RECLETTERS, represents teacher reports of how many letters of recommendation they wrote over the course of the current school year. Responses here were also placed on a 1 to 5 scale: 1 = 0 letters, 2 = 1 to 3 letters, 3 = 4 to 6 letters, 4 = 7 to 9 letters, and 5 = 10 or more letters. Both measures were collected at both Time 1 and Time 2. Time 1 responses were included as prior measures in final path analysis models.

In addition to the practice outcomes described, this study also examined teachers’ self-reported familiarity (FAM2) with five key components of the college-going process: (a) the Free Application for Federal Student Aid (FAFSA); (b) deadlines for college applications; (c) dates for college admissions exams (e.g., ACT and SAT); (d) online resources for the college search process; and (e) resources for scholarship and grant opportunities. For each area above, teachers ranked their familiarity on a scale of 1 to 5, with 1 representing not familiar and 5 representing very familiar. Scores on the FAM2 variable were weighted by constituent items’ standardized factor loadings. As with the practice outcomes above, scores for the familiarity outcome were collected at both Time 1 and Time 2. Time 1 responses were included as prior covariates in the final model.

Finally, we investigated a third outcome, teachers’ beliefs about the college-going process. We measured teachers’ beliefs using a four-item scale. Example items on this scale included, “My students have the capacity to do college preparatory work” and “I feel that it is part of my job to prepare students to succeed in college.” Teachers responded to these items on a scale ranging from 1 = strongly disagree to 4 = strongly agree. Scores for the belief outcome were collected at both Time 1 and Time 2.

For all the scale scores described above, we calculated reliability coefficients for both Time 1 and Time 2 using coefficient H (Hancock & Mueller, 2001). Coefficient H is a measure of maximal reliability that is appropriate for estimating the reliability of optimally weighted scales (McNeish, 2018). In such scales, items with higher standardized factor loadings make larger contributions to the overall scale score than do items with lower standardized factor loadings. Similarly, Coefficient H provides a measure of reliability that is commensurately weighted by standardized factor loadings. As with Cronbach’s alpha, values of Coefficient H can range from 0 to 1, with higher values indicating greater reliability. H values for scores on the scales of interest were as follows: Practice Time 1 = .90, Practice Time 2 = .89, Familiarity Time 1 = .80, Familiarity Time 2 = .94, Beliefs Time 1 = .59, Beliefs Time 2 = .46. Because we measured recommendation letters using a single item, we cannot provide an estimate of reliability. Because our reliability coefficients for the beliefs outcome were below the threshold for acceptable use, we did not conduct any further analyses using this variable. Table 3 presents standardized factor loadings for practices, familiarity, and beliefs at both time points.

Table 3.

Standardized Factor Loadings for Outcome Variables at Time 1 and Time 2.

	Factor loading (Time 1)	Factor loading (Time 2)
Practices^a
Talk about your experience in college with students	0.64	0.64
Talk about college admissions requirements and application deadlines with students	0.90	0.85
Talk with students about scholarship and financial aid opportunities	0.81	0.85
Help students with their college application essays or personal statements	0.74	0.76
Talk to students about what classes they should take to prepare for certain colleges	0.71	0.70
Integrate information about college into regular classroom lessons	0.55	0.50
Familiarity^b
Free Application for Federal Student Aid (FAFSA)	0.58	0.77
Deadlines for college applications (including early decision)	0.80	0.96
Dates for college admissions exams (e.g., ACT and SAT)	0.66	0.83
Resources for scholarship and grant opportunities	0.71	0.77
Beliefs
Teachers expect most students in this school to go to college	0.56	0.15
Teachers at this school help students plan for college outside of class time	0.43	0.08
My students have the capacity to do college preparatory work	0.61	0.45
I feel that it is part of my job to help prepare students to succeed in college	0.35	0.61

All items in this scale begin with the stem, “In this school, how often do you . . .”

Directions for this scale ask participants to rate how familiar they are with each item.

Focal covariates

This study was particularly focused on how teachers’ colleagues influence their college-going beliefs and practices. As such, a key covariate was each teacher’s exposure to the beliefs and practices of their colleagues at Time 1. To approximate this, the Wave 2 survey asked teachers to identify their five closest colleagues in the building, and the frequency with which they interacted with them using a Likert-type scale ranging from 1 = monthly to 4 = several times per day. Exposure terms were calculated for each of the six outcomes listed above by calculating the mean values of up to five teachers that are listed as important colleagues by teacher i. For example, if teacher i listed teachers j, k, l, m, and n as colleagues, teacher i’s estimated network exposure was calculated as the mean of responses for teachers j through n at Time 1. In addition to network exposure, this study tested the impact of teachers’ interactions with CAP program staff on their beliefs and practices at Time 2. Here, interactions with CAP staff were operationalized by the survey item “How often do you interact with CAP center staff?” Responses were on a scale from 1 to 4: 1 = never, 2 = rarely, 3 = sometimes, and 4 = frequently.

Other covariates

Along with the focal covariates described above, a series of additional covariates were included as statistical controls. These included dichotomous predictors of gender (FEMALE), race/ethnicity (NONWHITE), and school assignment (MIDDLEBROOK, SCHERZER, and KELLY). Several additional dichotomous predictors were included to account for teachers’ experience (YEAR10LESS, YEAR20MORE) and subject area assignment (SCIMATH, NONINSTRUCT). Table 4 below presents means and standard errors for all variables used in the analysis.

Table 4.

Descriptive Statistics for the Analytic Sample.

Variable name	n	M	SE	Minimum	Maximum
Outcomes
psum_time2	104	14.54	0.533	0	28
pmean_time2	99	2.555	0.075	1	4.667
t2_recletters	99	2.919	0.137	1	5
t2_capacity	96	2.833	0.074	1	4
t2_myjob	99	3.404	0.052	2	4
familmean_time2	99	3.021	0.115	1	5
Focal Covariates
exposure_sum	104	5.47	0.433	0	16
exposure_pmean	104	1.065	0.076	0	2.667
exposure_letters	104	1.371	0.115	0	4
exposure_capacity	104	1.674	0.126	0	3.5
exposure_job	104	2.025	0.147	0	4
exposure_famil	104	1.895	0.146	0	5
t2_capstaff	99	2.465	0.092	1	4
Time 1 Prior Measures
psum_time1	104	5.923	0.541	0	24
pmean_time1	101	1.017	0.091	0	4
t1_2letters	104	1.25	0.136	0	4
t1capacity	101	1.931	0.138	0	4
t1job	104	2.308	0.162	0	4
familmean_time1	101	2.2	0.171	0	5
Other Covariates
year10less	101	0.426	0.049	0	1
year20more	101	0.198	0.040	0	1
Noninstruct	101	0.277	0.045	0	1
Female	98	0.653	0.048	0	1
Nonwhite	101	0.0891	0.028	0	1
Scimath	101	0.257	0.043	0	1
knowt1	101	0.208	0.040	0	1
Middlebrook	104	0.154	0.035	0	1
Scherzer	104	0.212	0.040	0	1
Kelly	104	0.327	0.046	0	1

Within-School Cluster Analysis

To better understand the dynamics of teachers’ social structures in the four high schools, the teacher network data described above were used to construct a sociogram for each school, which illustrates each teacher (represented by dots) and their respective collegial ties (represented by lines from nominator to nominee). Along with a visual representation, a cluster analysis was performed using Kliquefinder (Frank, 1995, 1996), which tested potential teacher subgroups against a null hypothesis that the observed groups could have formed at random. Using a Monte Carlo simulation, a distribution and test statistic (theta) were generated to evaluate this null hypothesis. This analysis is a critical first step in understanding whether teachers’ collegial groups are relevant to understanding the social structure of each school. If distinct clustering is not found, this might suggest that further study of collegial subgroups would not be appropriate.

Modeling Teacher Network Effects

Cluster analysis provided one lens on the social structure of the schools in this sample by providing a picture of the overall social structure, including subgroups of teachers, and also by indicating the likelihood that a school has a robust social structure with distinct subgroups of teachers. Next, a series of influence models (Frank, 1998) was estimated to examine the impact of teachers’ collegial networks and their interaction with CAP program staff on their college-going-related familiarity and practices. We estimated these in a path analytic framework using the sem command in Stata 14 (StataCorp, 2015). For each outcome variable, we estimated a series of three models. Each influence model takes the functional form of a multiple linear regression, with the outcome of teacher familiarity and practices at Time 2 predicted by a series of covariates, including network exposure and Time 1 prior measures. The formal models (Equations 1-2) and explanations of components are below, specified in terms of teacher i.

In the first model for each outcome, we included Time 1 prior measures of the outcome, our exposure term, and the reported level of interaction with CAP program staff as predictors, such that,

\begin{matrix} Y_{i} = β_{0} + β_{1} (T I M E 1 P R I O R_{i}) + β_{2} (E X P O S U R E_{i}) \\ + β_{3} (T 2_C A P S T A F F_{i}) + ε_{i} . \end{matrix}

(1)

This model allowed us to estimate the influence of our focal covariates on each outcome before accounting for other covariates.

In the second model for each outcome, we kept all predictors from the first model and added other covariates to control for possible confounding variables, such as school assignment and other teacher characteristics. The equation for Model 2 for each outcome is found below:

\begin{array}{l} Y_{i} = β_{0} + β_{1} (T I M E 1 P R I O R_{i}) + β_{2} (Y E A R 10 L E S S_{i}) \\ + β_{3} (Y E A R 20 M O R E_{i}) + β_{4} (N O N I N S T R U C T_{i}) \\ + β_{5} (F E M A L E_{i}) + β_{6} (N O N W H I T E_{i}) + β_{7} (S C I M A T H_{i}) \\ + β_{8} (M I D D L E B R O O K_{i}) \\ + β_{9} (S C H E R Z E R_{i}) + β_{10} (K E L L Y_{i}) + β_{11} (E X P O S U R E_{i}) \\ + β_{12} (T 2_C A P S T A F F_{i}) + ε_{i}, \end{array}

(2)

where, $Y_{i}$ = teacher i’s Time 2 measure of outcome (practices or familiarity); $T I M E 1 P R I O R_{i}$ = teacher i’s Time 1 measure of outcome; $Y E A R 10 L E S S_{i}$ = 1 if teacher i reports 10 or fewer years of teaching experience; $Y E A R 20 M O R E_{i}$ = 1 if teacher i reports 20 or more years of teaching experience; $N O N I N S T R U C T_{i}$ = 1 if teacher i works in a noninstructional teaching role (guidance counselor, content coach); $F E M A L E_{i}$ = 1 if teacher i identifies as female; $N O N W H I T E_{i}$ = 1 if teacher i identifies as non-White; $S C I M A T H_{i}$ = 1 if teacher i reports teaching math or science; $M I D D L E B R O O K_{i}$ = 1 if teacher i works in Middlebrook High School; $S C H E R Z E R_{i}$ = 1 if teacher i works in Scherzer High School; $K E L L Y_{i}$ = 1 if teacher i works in Kelly High School; $E X P O S U R E_{i}$ = mean of Time 1 outcomes for teacher i’s reported colleagues; $T 2_C A P S T A F F_{i}$ = teacher i’s reported level of interaction with CAP program staff; and $ε_{i}$ = person-specific error term for teacher i, independent and identically normally distributed with common variance, $σ_{ε}^{2}$ .

Finally, in the third model for each outcome, we replaced our unweighted exposure term with a weighted exposure term. In this term, more popular teachers (i.e., those who were more often identified as a close colleague) contributed more weight to the term. For example, assume teacher j was identified as a close colleague by 10 peers and teacher k was identified as a close colleague by three peers. If teacher i identifies both teacher j and teacher k as close colleagues, teacher j’s score would contribute more toward teacher i’s exposure term than would teacher k’s score. The equation for Model 3 is otherwise identical to the equation presented for Model 2.

This three-step approach allowed us to examine the relative impact of the focal covariates before and after accounting for other covariates by examining changes in Akaike information criterion (AIC) and Bayesian information criterion (BIC) values, which are indicators of model fit. Furthermore, the third model in each series allowed us to examine an alternative specification of exposure such that the peers most frequently identified by others have more influence in the model.

Missing Data

Given the limited sample size of the study and the relatively high degree of missingness (27%-31%, depending on the outcome being measured), we adjusted for missing data using path analysis with Full Information Maximum Likelihood (FIML) estimation. FIML is an efficient estimation method that accounts for missing data within the analysis model and requires satisfying fewer assumptions than other methods for estimating missing data, such as multiple imputation (Allison, 2012). To estimate missing data, we used the mlmv estimator in Stata 14’s (StataCorp, 2015) sem command. This method assumes multivariate normality of all continuous variables and that missing values are missing at random (MAR).

Results

Within-School Cluster Analysis and Sociograms

Figures 1 to 4 are sociograms for each of the four high schools in the study sample. As the diagrams illustrate, each school is characterized by a unique social clustering structure. Monte Carlo simulations produced strong evidence that the observed subgroups in all schools were not formed by random chance (Frank, 1995).² In other words, all four schools showed evidence of clustering among teachers.

Figure 1.

Sociogram of collegial ties at Middlebrook HS. Ties are indicated by an arrow and solid line from nominator to nominee. Different shapes represent statistically significant within-school clusters, as determined by Frank’s (1995, 1996) clustering algorithm in Kliquefinder. Participants indicated by an “X” received nominations, but made not nominations themselves, and were therefore not a part of any specific subgroup.

Figure 2.

Sociogram of collegial ties at Drew HS. Ties are indicated by an arrow and solid line from nominator to nominee. Different shapes represent statistically significant within-school clusters, as determined by Frank’s (1995, 1996) clustering algorithm in Kliquefinder. Participants indicated by an “X” received nominations, but made not nominations themselves, and were therefore not a part of any specific subgroup.

Figure 3.

Sociogram of collegial ties at Scherzer HS. Ties are indicated by an arrow and solid line from nominator to nominee. Different shapes represent statistically significant within-school clusters, as determined by Frank’s (1995, 1996) clustering algorithm in Kliquefinder. Participants indicated by an “X” received nominations, but made not nominations themselves, and were therefore not a part of any specific subgroup.

Figure 4.

Sociogram of collegial ties at Kelly HS. Ties are indicated by an arrow and solid line from nominator to nominee. Different shapes represent statistically significant within-school clusters, as determined by Frank’s (1995, 1996) clustering algorithm in Kliquefinder. Participants indicated by an “X” received nominations, but made not nominations themselves, and were therefore not a part of any specific subgroup.

Middlebrook High School (Figure 1) shows evidence of four distinct teacher clusters, with high levels of within-group ties and somewhat lower levels of between-group ties. Drew High School (Figure 2) has five teacher clusters, with one cluster isolated from the other four. This means that teachers in this cluster only nominate each other, and no other faculty. Furthermore, it means that teachers in this cluster were only nominated by others in the same cluster. Scherzer High School (Figure 3) also appears to have five distinct clusters, but a different arrangement, with one cluster of teachers (represented by black circles) both giving and receiving high levels of nominations from the other subgroups. Finally, Kelly High School (Figure 4) has nine distinct teacher clusters, the most, by far, of any school in the sample. Two subgroups, denoted by white circles and black squares in the sociogram, appear highly central to the social structure, and both give and receive the majority of nominations.

Influence Modeling for Teacher Practices and Beliefs

Moving from cluster analysis to influence modeling using path analysis, we estimate the impact of teachers’ colleagues and CAP staff interactions on their practices and familiarity related to college-going. Tables 5 to 7 present results for each series of models for the practices, recommendation letters, and familiarity outcomes, respectively.

Table 5.

Regression Estimates Predicting Teachers’ Time 2 Familiarity.

	Model 1		Model 2		Model 3
Parameter	Β	SE	β	SE	β	SE
Time 1 familiarity	.485***	0.106	.394***	0.0884	.402***	0.088
Exposure	.001	0.209	.095	0.230
CAP staff interaction	.127	0.0781	.064	0.077	.057	0.078
Fewer than 10 years teaching			.101	0.156	.093	0.159
More than 20 years teaching			.306**	0.171	.300**	0.179
Noninstructional position			.124	0.169	.131	0.168
Female			−.053	0.130	−.049	0.131
Non-White			−.206*	0.178	−.212*	0.177
Science/math teacher			−.179*	0.163	−.183**	0.163
Middlebrook HS			−.134*	0.193	−.128	0.193
Scherzer HS			.138	0.178	.135	0.185
Kelly HS			−.061	0.198	−.091	0.239
Exposure (weighted)					.027	0.046
N	163		163		163
AIC	1,028.4		2,127.0		2,379.6
BIC	1,071.8		2,448.7		2,701.3

Note. CAP = College Ambition Program; HS = high school; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion.

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

Table 6.

Regression Estimates Predicting Teachers’ Time 2 Practices.

	Model 1		Model 2		Model 3
Parameter	β	SE	β	SE	β	SE
Time 1 practices	.418***	0.068	.406***	0.066	.402***	0.064
Exposure	.022	0.154	−.031	0.161
CAP staff interaction	.209**	0.041	.167	0.046	.165	0.047
Fewer than 10 years teaching			.09	0.093	.09	0.093
More than 20 years teaching			.073	0.104	.078	0.105
Noninstructional position			.173*	0.101	.173*	0.101
Female			.069	0.080	.075	0.079
Non-White			−.007	0.113	−.02	0.115
Science/math teacher			−.05	0.099	−.049	0.099
Middlebrook HS			−.127	0.118	−.129	0.118
Scherzer HS			−.137	0.111	−.13	0.111
Kelly HS			−.102	0.097	−.089	0.103
Exposure (weighted)					.014	0.037
N	163		163		163
AIC	727.6		1,886.4		2,115.7
BIC	770.9		2,208.2		2,437.5

Note. CAP = College Ambition Program; HS = high school; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion.

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

Table 7.

Regression Estimates Predicting Teachers’ Time 2 Recommendation Letters Written.

	Model 1		Model 2		Model 3
Parameter	β	SE	β	SE	β	SE
Time 1 recommendation letters	.718***	0.0697	.762***	0.0696	.767***	0.0702
Exposure	.086	0.371	.094	0.417
CAP staff interaction	.211**	0.109	.119	0.120	.119	0.122
Fewer than 10 years teaching			.086	0.222	.089	0.224
More than 20 years teaching			.031	0.253	.032	0.257
Noninstructional position			.026	0.237	.031	0.240
Female			.051	0.204	.043	0.204
Non-White			−.041	0.372	−.027	0.380
Science/math teacher			.026	0.220	.024	0.222
Middlebrook HS			−.261***	0.313	−.258***	0.317
Scherzer HS			−.048	0.270	−.057	0.270
Kelly HS			−.080	0.259	−.087	0.273
Exposure (weighted)					.055	0.0957
N	152		163		163
AIC	1,009.3		2,165.0		2,393.5
BIC	1,051.6		2,486.8		2,715.3

Note. CAP = College Ambition Program; HS = high school; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion.

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

In our first series of models, we estimated the influence of various predictors on teachers’ teaching practices related to college-going. In the first model, previous practices ( $β = . 42$ ) and level of reported interaction with CAP staff ( $β = . 21$ ) significantly predicted teachers’ practices related to college-going. After including other covariates in the second model, interaction with CAP staff was no longer a significant predictor of these practices. In this model, working in a noninstructional teaching role (e.g., school counselor) positively predicted use of college-going practices ( $β = . 17$ ), and previous practices remained a significant predictor ( $β = . 41$ ). In the final model, in which we weight our exposure term by teacher popularity, noninstructional teaching role ( $β = . 17$ ) and previous practices ( $β = . 40$ ) remained significant predictors of college-going practices.

In our second series of models, we estimated the influence of predictors on the amount of college reference letters teachers reported writing for students. In the first model, previous number of letters written ( $β = . 72$ ) and level of reported interaction with CAP staff ( $β = . 21$ ) emerged as significant predictors of reference letters written. After adding other covariates to the second model, interaction with CAP staff no longer predicted the amount of reference letters written. However, previous number of letters written remained a significant predictor in both the second ( $β = . 76$ ) and the third ( $β = . 77$ ) models. Furthermore, working at Middlebrook High School negatively predicted reference letter writing in both the second ( $β = - . 26$ ) and third ( $β = . - 26$ ) models.

In our third series of models, we estimated the influence of predictors on teachers’ familiarity with components of the college-going process. In all three models, previous familiarity was a significant predictor of current familiarity ( $β = . 46, . 39, . 40,$ respectively). In the second model, during which we added other covariates, the NONWHITE variable ( $β = - . 21$ ), the SCIMATH variable ( $β = - . 18$ ), the YEAR20MORE variable ( $β = . 31$ ), and the MIDDLEBROOK variable ( $β = - . 13$ ) emerged as significant negative predictors of familiarity with components of the college-going process. That is, teachers who identified as non-White, those who reported teaching math or science, and those who taught at Middlebrook tended to have lower familiarity with components of the college-going process, whereas teachers who had taught for more than 20 years tended to have higher familiarity. The NONWHITE ( $β = - . 21$ ) variable, the SCIMATH variable ( $β = - . 18$ ), and the YEAR20MORE variable ( $β = . 30$ ) remained significant predictors in the third model.

Discussion

The present study examined the relationships between teachers’ collegial networks, a school-wide intervention (CAP) designed to foster a college-going culture in high schools, and teacher practices and familiarity related to college-going. To explore these relations, we first used social network analysis to better understand the social structures existing in the four schools that received the CAP intervention. Applying Frank’s (1995, 1996) clustering algorithm, we found that all four schools demonstrated evidence of distinct teacher subgroups. The size and structure of subgroups varied between school sites, suggesting wide potential variation in teacher network exposure. Because we believed teacher networks play a central role in the success of the CAP initiative, we further investigated how variability in teacher interaction patterns related to our outcomes of interest. More specifically, we used data from teachers’ collegial networks to develop an influence model predicting teachers’ practices and familiarity related to college-going. We did not find that either a weighted or unweighted exposure term—representing the mean scores of the colleagues with whom teachers reported interacting on each outcome of interest—significantly predicted teachers’ familiarity with or practices related to college-going. Teachers’ reported level of interaction with CAP staff significantly predicted their use of practices related to college-going as well as the amount of college reference letters they wrote for students in models including focal covariates. However, once additional covariates were added in subsequent models, interaction with CAP staff was no longer a significant predictor of either outcome.

These findings connect and extend recent work on teacher social networks by focusing on teachers’ college-going practices and beliefs, an area that has not been much explored using social network analysis. Previous studies have examined how teachers’ collegial networks influence the adoption of new instructional technology (Frank et al., 2004), or the role networks play in teachers’ implementation of school and district policies (Daly, 2012; Daly et al., 2010), while this study examined the role of teacher networks in shaping teachers’ beliefs and practices related to college-going. Thus, this work represents an extension of social network research in a new context.

Furthermore, this work provides additional insight into the malleability (or perhaps lack thereof) of teacher familiarity with and practices related to college-going. In all models, teachers’ previous practices or familiarity were the strongest predictors of later practices or familiarity, which suggests that they are difficult to change. However, interaction with CAP staff did emerge as a significant predictor of these outcomes in some models, even after controlling for the influence of teachers’ peer networks, which provides some cautionary evidence that changes in teachers’ knowledge and practices are possible even over a relatively limited period. Given the strong emphasis on developing college-going cultures in schools (Domina, 2009; Holland & Farmer-Hinton, 2009) and the key role that teachers are believed to play in the process (Bryan et al., 2017), more research seems needed to understand how to better effect such changes.

Limitations

Finally, this study has several limitations that should be addressed. First, because the subjects of interest (teachers) are a part of already formed social cliques and subgroups, the sample cannot be random; in fact, it is intentionally not so. To help alleviate the risk of selection bias and/or other latent variables influencing our outcome, two waves of data were collected, which provided a prior measure to control for when estimating the teacher network effect on beliefs and practices. Second, this study makes use of teacher self-reports of both familiarity and practices. Familiarity, at least in this case, seems on one hand to be straightforward (teachers know about financial aid resources or they do not), but, on the other hand, it also encompasses broader views and perspectives on building-wide culture, which could be much harder to understand systematically. Finally, even after controlling for pretest measures at Time 1,³ it is important to note that the findings should not be interpreted as causal, but rather associational.

Conclusion

In sum, this study offers insight into the nature of teachers’ familiarity and practices related to college-going and how they may be influenced by those of their peers over the course of a school year. We find both constructs to be stable over time and not significantly impacted by interaction with peers or with intervention staff. Future work might examine the impact of additional exposure within a larger and more heterogeneous sample of schools and teachers across multiple years, to see whether practices or beliefs may be more susceptible to change over time. Regardless, we would urge caution to those who might expect sweeping changes in teaching practices or beliefs over a short period of time. Many of these ideas seem to be deeply held and may require sustained effort to change.

Footnotes

Appendix

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this work is provided by the National Science Foundation under grant number DRL-1316702, P.I. Barbara Schneider. The opinions expressed here are those of the authors and do not represent the views of the funding agency.

Notes

ORCID iD

Michael Broda

Author Biographies

Michael Broda, PhD, is an assistant professor of quantitative methods and statistics in the Department of Foundations of Education in the School of Education at Virginia Commonwealth University. He received his PhD in educational policy from Michigan State University. His recent publications have appeared in the Journal of Research on Educational Effectiveness, AERA Open, and the Journal of Research in Science Teaching.

Eric Ekholm, MT, is a graduate student in the Educational Psychology program in the School of Education at Virginia Commonwealth University. Some of his recent research has been published in Educational Psychology Review, Early Childhood Research Quarterly, and School Psychology Quarterly. Before returning to school to pursue his PhD, Eric taught middle school English for four years.

Barbara Schneider, PhD is the John A. Hannah Chair and University Distinguished professor in the College of Education at Michigan State University. Her research focuses on how the social contexts of schools and families influence the academic and social well-being of adolescents as they move into adulthood.

Amy C. Hutton, PhD, is the director of Enrollment Research and Evaluation for the Division of Strategic Enrollment Management at Virginia Commonwealth University. She received her PhD in Educational Research and Evaluation and MFA in Theatre Pedagogy from Virginia Commonwealth University. Her work has been featured in the New York Times, the Bulletin of Science, Technology, and Society, several textbooks, and internationally through the American Association of Collegiate Registrars and Admission Counselors.

References

Allison

P. D.

(2012). Handling missing data by maximum likelihood. In Proceedings of the SAS Global Forum 2012 Conference. Cary, NC: SAS Institute Inc. Retrieved from http://support.sas.com/resources/papers/proceedings12/312-2012.pdf.

American School Counselor Association. (2018). Student-to-school-counselor ratio 2015-2016. Available from www.schoolcounselor.org

Baum

Payea

(2013). Education pays, 2013: The benefits of higher education for individuals and society. New York, NY: College Board.

Bidwell

C. E.

Yasumoto

J. Y.

(1999). The collegial focus: Teaching fields, collegial relationships, and instructional practice in American high schools. Sociology of Education, 72, 234-256.

Bourdieu

Wacquant

L. J.

(1992). An invitation to reflexive sociology. Chicago, IL: The University of Chicago Press.

Broda

M. D.

(2015). Three perspectives on postsecondary attainment: Students, teachers, and institutions (Doctoral dissertation). East Lansing, MI: ProQuest.

Bryan

Farmer-Hinton

Rawls

Woods

C. S.

(2017). Social capital and college-going culture in high schools: The effects of college expectations and college talk on students’ postsecondary attendance. Professional School Counseling, 21, 95-107.

Bryk

Schneider

(2002). Trust in schools. New York, NY: Russell Sage Foundation.

Buddin

(2012). Implications of educational attainment trends for labor market outcomes. ACT Research Report Series, 2012(7): 1-40.

10.

Burt

(2000). The network structure of social capital. Research in Organizational Behavior, 22, 345-423.

11.

Chenoweth

Galliher

R. V.

(2004). Factors influencing college aspirations of rural West Virginia high school students. Journal of Research in Rural Education, 19(2), 1-14.

12.

Cobb

R. A.

(1989). Vocational and educational aspirations of high school students: A problem for rural America. Research in Rural Education, 6(2), 11-16.

13.

Cohen

A. K.

Syme

S. L.

(2013). Education: A missed opportunity for public health intervention. American Journal of Public Health, 103, 997-1001.

14.

Coleman

J. S.

(1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95-S120.

15.

Csikszentmihalyi

Schneider

(2000). Becoming adult: How teenagers prepare for the world of work. New York, NY: Basic Books.

16.

Daly

A. J.

(2012). Data, dyads, and dynamics: Exploring data use and social networks in educational improvement. Teachers College Record, 114(11), 1-38.

17.

Daly

A. J.

Liou

Y. H.

Moolenaar

N. M.

(2014). The principal connection: Trust and innovative climate in a network of reform. In Van Maele

Forsyth

P. B.

Van Houtte

(Eds.), Trust and school life: The role of trust for learning, teaching, leading, and bridging (pp. 285-311). Dordrecht, The Netherlands: Springer.

18.

Daly

A. J.

Moolenaar

N. M.

Bolivar

J. M.

Burke

(2010). Relationships in reform: The role of teachers’ social networks. Journal of Educational Administration, 48, 359-391.

19.

Day

J. C.

Newburger

E. C.

(2002). The big payoff: Educational attainment and synthetic estimates of work-life earnings (Current Population Reports, Special Studies). Washington, DC: U.S. Department of Commerce, Economics and Statistics Administration, U.S. Census Bureau.

20.

Degenne

Forsé

(1999). Introducing Statistical Methods Series: Introducing social networks. London, England: Sage.

21.

Domina

(2009). What works in college outreach: Assessing targeted and schoolwide interventions for disadvantaged students. Educational Evaluation and Policy Analysis, 31, 127-152.

22.

Farmer-Hinton

R. L.

(2008). Social capital and college planning: Students of color using school networks for support and guidance. Education and Urban Society, 41, 127-157.

23.

Frank

K. A.

(1995). Identifying cohesive subgroups. Social Networks, 17, 27-56.

24.

Frank

K. A.

(1996). Mapping interactions within and between cohesive subgroups. Social Networks, 18, 93-119.

25.

Frank

K. A.

(1998). Chapter 5: Quantitative methods for studying social context in multilevels and through interpersonal relations. Review of Research in Education, 23, 171-216.

26.

Frank

K. A.

Zhao

Borman

(2004). Social capital and the diffusion of innovations within organizations: Application to the implementation of computer technology in schools. Sociology of Education, 77, 148-171.

27.

Frank

K. A.

Zhao

Penuel

W. R.

Ellefson

Porter

(2011). Focus, fiddle, and friends: Experiences that transform knowledge for the implementation of innovations. Sociology of Education, 84, 137-156.

28.

Freeman

(1997). Increasing African-Americans’ participation in higher education: African-American high-school students’ perspectives. The Journal of Higher Education, 68, 523-550.

29.

Gandara

Bial

(2001). Paving the way to postsecondary education: K-12 intervention programs for underrepresented youth (NCES 2001-205). Washington, DC: U.S. Department of Education, National Center for Education Statistics.

30.

Hancock

G. R.

Mueller

R. O.

(2001). Rethinking construct reliability within latent variable systems. In Cudeck

DuToit

Sorbom

(Eds.), Structural equation modeling: Present and future (pp. 195-216). Lincolnwood, IL: Scientific Software International, Inc.

31.

Holland

N. E.

Farmer-Hinton

R. L.

(2009). Leave no schools behind: The importance of a college culture in urban public high schools. The High School Journal, 92(3), 24-43.

32.

Ingels

S. J.

Pratt

D. J.

Herget

D. R.

Burns

L. J.

Dever

J. A.

Ottem

. . . Leinwand

(2011). High School Longitudinal Study of 2009 (HSLS:09): Base-year data file documentation (NCES 2011-328, U.S. Department of Education). Washington, DC: National Center for Education Statistics.

33.

Kearns

D. M.

Fuchs

McMaster

K. L.

Sáenz

Fuchs

L. S.

Yen

. . . Smith

T. M.

(2010). Factors contributing to teachers’ sustained use of kindergarten peer-assisted learning strategies. Journal of Research on Educational Effectiveness, 3, 315-342.

34.

Kennedy

M. M.

(2005). Inside teaching. Cambridge, MA: Harvard University Press.

35.

Levenstein

(2016). 2014 My Voice, My School survey quality profile: Technical report. Chicago, IL: UChicago Consortium on School Research.

36.

Levine

Nidiffer

(1996). Beating the odds: How the poor get to college. San Francisco, CA: Jossey-Bass.

37.

McDonough

P. M.

(1997). Choosing colleges: How social class and schools structure opportunity. Albany: State University of New York Press.

38.

McDonough

P. M.

(2005). Counseling and college counseling in America’s high schools. Alexandria, VA: National Association for College Admission Counseling.

39.

McNeish

(2018). Thanks coefficient alpha, we’ll take it from here. Psychological Methods, 23, 412-433. doi:10.1037/met0000144

40.

Moolenaar

N. M.

(2012). A social network perspective on teacher collaboration in schools: Theory, methodology, and applications. American Journal of Education, 119, 7-39.

41.

National Academies of Sciences, Engineering, and Medicine. (2016). Barriers and opportunities for 2-year and 4-year STEM degrees: Systemic change to support students’ diverse pathways. Washington, DC: The National Academies Press.

42.

National Research Council. (2011). Expanding underrepresented minority participation: America’s science and technology talent at the crossroads (Committee on Underrepresented Groups and Expansion of the Science and Engineering Workforce Pipeline, Hrabowski

F. A.

Henderson

P. H.

Psalmonds

, Eds.). Washington, DC: Board on Higher Education and the Workforce, Division on Policy and Global Affairs, The National Academies Press.

43.

National Science Board. (2007). A national action plan for addressing the critical needs of the U.S. science, technology, engineering, and mathematics education system (Publication No. NSB-07-114). Washington, DC: U.S. Government Printing Office.

44.

O’Connor

(2000). Dreamkeeping in the inner city: Diminishing the divide between aspirations and expectations. In Danziger

Lin

A. C.

(Eds.), Coping with poverty: The social contexts of neighborhood, work, and family in the African American community (pp. 105-140). Ann Arbor: University of Michigan Press.

45.

Olsen

Sexton

(2009). Threat rigidity, school reform, and how teachers view their work inside current education policy contexts. American Educational Research Journal, 46, 9-44.

46.

Oreopoulos

Petronijevic

(2013). Making college worth it: A review of the returns to higher education. The Future of Children, 23(1), 41-65.

47.

Penuel

W. R.

Riel

Krause

Frank

K. A.

(2009). Analyzing teachers’ professional interactions in a school as social capital: A social network approach. Teachers College Record, 111, 124-163.

48.

Perna

L. W.

Titus

M. A.

(2005). The relationship between parental involvement as social capital and college enrollment: An examination of racial/ethnic group differences. The Journal of Higher Education, 76, 485-518.

49.

Plank

S. B.

Jordan

W. J.

(2001). Effects of information, guidance, and actions on postsecondary destinations: A study of talent loss. American Educational Research Journal, 38, 947-980.

50.

Rogers

E. M.

(2003). Diffusion of innovations. New York, NY: Free Press.

51.

Schneider

(2007). Forming a college-going community in U.S. public high schools. Seattle, WA: Bill & Melinda Gates Foundation.

52.

Schneider

(2015). 2014 AERA Presidential Address: The College Ambition Program: A realistic transition strategy for traditionally disadvantaged students. Educational Researcher, 44, 394-403.

53.

Schneider

Broda

Judy

Burkander

(2013). Pathways to college and STEM careers: Enhancing the high school experience. New Directions for Youth Development, 2013(140), 9-29.

54.

Schneider

Judy

Mazuca

Broda

(2014). Trust in elementary and secondary urban schools: A pathway for student success and college ambition. In Van Maele

Van Houtte

Forsyth

(Eds.), Trust relationships and school life (pp. 37-56). New York, NY: Springer.

55.

Schneider

Stevenson

(1999). The ambitious generation: America’s teenagers motivated but directionless. New Haven, CT: Yale University Press.

56.

Shadish

W. R.

Clark

M. H.

Steiner

P. M.

(2008). Can nonrandomized experiments yield accurate answers? A randomized experiment comparing random and nonrandom assignments. Journal of the American Statistical Association, 103, 1334-1344.

57.

Sherman

Sage

(2011). Sending off all your good treasures: Rural schools, brain-drain, and community survival in the wake of economic collapse. Journal of Research in Rural Education, 26(11), 1-14.

58.

Siciliano

M. D.

(2016). It’s the quality not the quantity of ties that matters: Social networks and self-efficacy beliefs. American Educational Research Journal, 53, 227-262.

59.

Spillane

J. P.

Shirrell

Sweet

T. M.

(2017). The elephant in the schoolhouse: The role of propinquity in school staff interactions about teaching. Sociology of Education, 90, 149-171.

60.

StataCorp. (2015). Stata Statistical Software: Release 14. College Station, TX: Author.

61.

UChicago Consortium on School Research. (2009). Survey of Chicago Public Schools—High school teacher edition. Chicago, IL: Author.

62.

Wilson

S. M.

(1990). A conflict of interests: The case of Mark Black. Educational Evaluation and Policy Analysis, 12, 293-310.