The academy for teachers professional development program. A model to support teacher retention

Abstract There is limited robust evidence of the effectiveness of professional development (PD) on teacher retention. This study tested The Academy for Teachers’ PD program, a content-focused intensive workshop with complementary events. This program is designed to impact teacher retention, support integration of material into the classroom and teacher network-building; and to increase pride in the profession, energy/excitement for teaching, and self-efficacy. We used an exploratory mixed methods approach, including in-depth interviews with program participants (Fellows), followed by a post-only non-equivalent quasi-experimental design of intervention effects. We conducted 18 in-depth interviews with Fellows, coding Fellow responses using thematic, line-by-line coding. We further surveyed K-12 teachers (n = 757) in the New York City metropolitan area who participated in The Academy program compared to those that applied but didn’t participate. Qualitative data illustrated mechanisms underpinning success of the program. Quantitative findings demonstrated significant positive effects on teacher retention (OR = 3.71) independent of years of teaching, school-level factors, and classroom autonomy. There were positive statistically significant findings for integration of material into the classroom, perceived pride in the profession, and teacher efficacy with small to moderate effect sizes. This study illuminates how a unique and holistic approach to PD can have measurable positive impacts on teacher outcomes and teacher retention.


Introduction
Attracting and retaining high-quality teachers has been an ongoing problem in the education sector since the 1980s (Ingersoll, 1994;Nguyen et al., 2020), and projections through 2025 estimate the teacher shortage will continue to persist, exceeding 110,000 teachers annually (Sutcher et al., 2016). This predicament will continue to affect low-performing, low-income areas with greater proportions of minority students, as they disproportionately experience higher staffing turnover and an influx of inexperienced teachers compared to high-performing, high-income, and predominately white schools (Borman & Dowling, 2008;Boyd et al., 2011;Johnson et al., 2005).
In New York City, teacher attrition rates are higher than the rest of the nation (García & Weiss, 2019;Miller, 2004;Sutcher et al., 2016). New York City experiences a 15% annual turnover rate on average, with rates as high as 31% in low-income community school districts (New York State Education Department, 2018). When indicators of teacher quality (certification, relevant training, and years of experience) are taken into account, the shortage is even more acute, particularly impacting high-poverty schools (García & Weiss, 2019;Guarino et al., 2006).
Keeping high-quality teachers in classrooms is critical, as they are arguably the greatest in-school influence on student achievement (Aaronson et al., 2007;Chetty et al., 2014;Kane & Staiger, 2008). The retention of high-quality teachers not only has bearing on what and how much students learn, but also has longitudinal effects on student achievement (Cheng & Zamarro, 2018;Goe, 2007;Kyriakides & Creemers, 2008;Shen et al., 2004). Furthermore, teacher retention can improve the health of the school by increasing teacher and instructional quality, promoting staff cohesion, teacher morale, and curriculum coherence (Ronfeldt et al., 2013), while minimizing the financial burden to schools and the education system writ large with the need to fill vacancies (Haynes, 2014).
There are many factors associated with teacher attrition and retention that are important in shaping teachers' decisions to stay in or leave the profession (Borman & Dowling, 2008; Boyd et al., 2011;Nguyen et al., 2020). Such factors range from teacher characteristics and demographics to individual factors (e.g., self-efficacy) that can be modified/targeted with professional development to external factors beyond a teacher's control (e.g., workforce trends and education policies).
Over the past two decades, a substantial corpus of research has demonstrated that professional development (PD) can have small but significant impacts on improving teacher retention (Allen & Sims, 2017;Coldwell, 2017;Reynolds & Wang, 2005;See, et al., 2020), particularly when it invigorates and excites teachers about teaching (Gore et al., 2017), increases teachers' impact on their students through quality content delivery (Carter & Keiler, 2009), and fosters the development of a professional community that expands beyond the teacher's school (Darling-Hammond et al., 2017;Zhang & Zeller, 2016). Professional development can also impact teacher retention by improving teacher self-efficacy (Tzivinikou, 2015). Teachers who have high selfefficacy have greater impacts on student achievement, as they are able and willing to persist through difficulties using innovative strategies, and exhibit greater energy and enthusiasm for teaching (Bandura, 1977). Since over 90% of teachers in the United States participate in PD annually, PD can be an effective method to reach teachers (Taie & Goldring, 2020).
The general consensus around effective professional development is that it should be long-term, content-focused, inquiry-based, learner-centered, and incorporate active learning. Such PD utilizes adult learning theory; provides opportunities for feedback and reflection; supports collaboration, typically in job-embedded contexts; and supports teachers with learning communities (Avalos, 2011;Borko et al., 2010;Darling-Hammond et al., 2017). Many of these characteristics have been embedded into the Every Student Succeeds Act (2015), which defines professional development as "sustained (not stand-alone, 1-day, or short term workshops), intensive, collaborative, jobembedded, data-driven, and classroom focused" (p. 11).
Evidence suggests that participation in alternative structures may provide effective teacher PD. Such structures include intensive single workshops (Guskey & Yoon, 2009), experiences that involve teachers as both learners and teachers (Darling-Hammond & McLaughlin, 1995), as well as experiences that incorporate formal and informal learning opportunities (Lantz-Andersson et al., 2018;Richter et al., 2014;Thacker, 2017). Despite the scope and opportunity that PD presents for educators, many teachers believe their professional goals are not being met by current PD offerings (Taie & Goldring, 2020). Teachers want professional development that is interactive, engaging, and relevant to their students and themselves (Matherson & Windle, 2017).
In this report, we assess the outcomes of a specific professional development program, The Academy for Teachers Program (henceforth referred to as The Academy). The Academy's programs, described in detail below, provide a unique constellation of experiences designed to elevate both participants and the teaching profession. Programming includes intensive, full-day structured courses based on subject-specific content of particular interest to teachers and led by renowned experts that allows teachers to be students within each session. The program honors teachers through the provision of intimate and exclusive experiences and aims to build pride in the profession, as well as renewed excitement and energy for teaching. Select group excursions to cultural events, regular social activities, and teacher-driven explorations into practice round out the program offerings. In addition, The Academy aims to facilitate the creation of a diverse professional community and the sharing of resources that increase high-quality content brought into the classroom to improve teacher retention and ultimately, student outcomes. Figure 1, presents a holistic framework identifying factors that influence teacher retention across four broad levels: (1) Individual factors, such as demographics characteristics (e.g., age, gender, relationship status, race/ethnicity), teacher characteristics (e.g., entry into teaching, years of teaching experience), teacher beliefs/perceptions/abilities (e.g., self-efficacy, energy/enthusiasm, perceived pride in the profession, feeling like a student); (2) Interpersonal factors, such as interactions with other people which can provide social and professional support; (3) School factors, characteristics of the school and its environment, such as administrative support, classroom autonomy, school resources, student body composition, and working conditions; and (4) External factors related to workforce trends (e.g., salaries, employment rates), education trends, and policies.
In the study reported here, we address the following interconnected research questions: (1) To what extent, if any, does participation in The Academy for Teachers professional development program build pride in and excitement for the teaching profession?
(2) To what extent, if any, does participation in The Academy for Teachers professional development program promote teaching confidence, relative to similar teachers who don't participate in the program?
(3) To what extent, if any, does participation in The Academy for Teachers professional development program lead to integration of material into the classroom relative to similar teachers who don't participate in this program?
(4) To what extent, if any, does participation in The Academy for Teachers professional development program build community relative to similar teachers who don't participate in this program?
To what extent, if any, does participation in The Academy for Teachers professional development program lead to greater teacher retention relative to similar teachers who don't participate in this program?

Overview of the academy for teachers professional development program
The Academy for Teachers (The Academy) is a non-profit organization that provides professional development programming for full-time K-12 teachers with at least three years of teaching experience and who are working in either a public, charter, private, or religious/parochial school in the New York City metropolitan area. The Academy's programming aims to honor teachers by providing them with elevated enrichment opportunities with leading experts across disciplines, exposing them to diverse communities of exceptional teachers, and building their pride in and excitement for the teaching profession.
The Academy provides master classes, events, activities, and workshops. The Academy receives approximately 420 nominations per year and accepts 63% of these nominations. Once accepted to a master class, teachers become Fellows of the program and are eligible to participate in the other program offerings (events, activities, and workshops). Teachers are nominated for a master class by principals and peers, based on the following criteria: (1) passion for the subject, (2) creativity in the classroom, and (3) devotion to students. A nomination, application, and reference, all requiring open-ended support statements focused on the criteria, are rated by multiple Academy reviewers and Fellow scorers to determine admission into the program.
The master classes, the organization's flagship offering, provide full-day sessions that bring educators together with a "Master," a prominent artist/thinker (including Pulitzer and Nobel Prize winners, academic experts, MacArthur and Guggenheim fellows, policy makers, politicians, and business leaders) in distinguished, intimate settings. During the morning, teachers spend three hours working with the accomplished guest speaker. In the afternoon the teachers exchange ideas and strategies with other educators who share their passion for the day's topic. The Academy offers an average of 16 master classes per year (each approximately eight hours in length), and Fellows typically participate in 1 to 2 classes per year.
Master classes are intentionally structured to support an increase in (1) teacher pride in the profession, (2) excitement and enthusiasm for teaching, (3) integration of program material in the classroom, (4) the development of professional networks among teachers, and (5) the diversity of teacher networks.

Teacher pride in the profession
Master classes typically provide 18 teachers with an intimate experience in beautiful and often exclusive spaces where efforts are made to make teachers feel valued and honored (e.g., through introductory speeches highlighting their accomplishments, distributing handouts sharing their qualifications, providing teachers elevated food offerings, and imparting direct messages of encouragement from Masters).

Excitement/enthusiasm for teaching
The master classes allow teachers to be immersed in their subject for the day, while having the opportunity to experience being a student again. Teachers can share their ideas, experiences, and passions for the content which, supported by a peer effect, is included in programming to build enthusiasm and excitement for teaching.

Student impact through integration of materials in the classroom
To support greater integration of material into the classroom, teachers are provided with content to inform teaching units and/or electives. Time is also designated within the session for teachers to meet with peers teaching similar grade levels and/or subjects and collaboratively generate ideas to incorporate content learned in the Master class into their classroom. Furthermore, teachers are also encouraged to contribute to a "Pedagogical Potluck," an online repository of lesson plans specific to the content that teachers can access after the master class for inspiration.

Community network building and diversity of networks
At the beginning of each master class, teachers meet in small heterogeneous groups (sorted by grade level, type of school, and/or subject) to discuss a preassigned homework assignment designed to foster communication, deepen content knowledge, and build passion and commitment to the content area. Throughout the master class, teachers have numerous opportunities to network with their peers, including during a catered unstructured lunch and in concluding exercises with peers to generate strategies to incorporate material into their classrooms. To increase opportunities to network, teachers attending the master class are also provided with their peers' contact information and are added to a listserv with updates and invitations to future activities and events.
The Academy brings together a diverse group of teachers and provides a number of opportunities to meet other teachers from different types of schools, teaching different subjects and grade levels, and with varying levels of teaching experience.
Events are special one-off opportunities for Fellows to connect to the cultural and intellectual life of New York City alongside other excellent educators. Events aim to increase perceived pride in the profession. Activities are sustained social offerings, including a Fellows' chorus and softball team. Events and activities also serve to expand the diversity and breadth of teachers' professional networks. Workshops typically focus on content or pedagogy and are opportunities for teachers to teach one another. Fellows participate in an average of two events, activities, and/or workshops per year.
Collectively, The Academy offers inclusion in an exclusive "club" that gives Fellows a sense of belonging and that extends its reach beyond a conventional one-off PD event by also providing ongoing access to resources, a community of similar professionals, encouragement, and skill-andcommunity building events, activities, and workshops.

Study Design
The study used an exploratory mixed methods approach (Creswell & Clark, 2017), including indepth interviews with Fellows, followed by a post-only quasi-experimental design to evaluate the effects of The Academy's professional development program on the following target outcomes: • teacher confidence • pride in the profession • excitement/enthusiasm for teaching • student impact through increased integration of PD materials into the classroom • community network building • diversity of teacher networks • teacher retention

Interview protocol
The lead investigators created an interview script with five open-ended questions and probes. These questions aimed to understand initial and sustained interest in the program and potential mediators driven by the program related to teacher retention and classroom integration. Sample questions included: What initially interested you in the program? For what reasons have you stayed involved/not remained involved? In what ways has being a Fellow impacted you? How, if at all, has The Academy encouraged you to stay in the teaching profession?

Recruitment and sampling
Researchers had access to The Academy's database of Fellows, which included demographic characteristics collected through the application process and contact information. The researchers used a purposeful sampling strategy in recruitment to match the sample of interview participants on the following key demographic characteristics: grade-level taught; school location within the five New York City boroughs; type of school (public or private); Title 1/low-income serving; teaching experience (5 years, 10-19 years, 20+ years); and program engagement (high, medium, and low). Thirty-six of the 1,912 Fellows in the database (1.88%) were initially contacted via e-mail over a 2-week period in November 2020, to serve as interview participants that represented the population of Fellows across the aforementioned demographic characteristics.

Qualitative data collection and analysis
The two lead researchers, each experienced in conducting interviews and from an organization independent from The Academy for Teachers, held 30-minute online interviews with each of the 36 the participants. All interviews were audio-recorded and transcribed verbatim by an outside source. Transcriptions were then reviewed with their original audio-recordings a minimum of two times to ensure accuracy. One researcher reviewed the transcripts beginning with an open-coding process, using line-by-line coding to identify themes, patterns of words, perceptions, and ideas which were then classified into categories. The initial themes included emergent themes and a priori themes identified from The Academy's program logic model. The two researchers reviewed the coding scheme with two transcripts and discrepancies between coding pairs were resolved through conversations between these researchers to generate the final coding scheme. A percent agreement of 85% across coding pairs indicated acceptable levels of reliability. The remaining transcripts were imported into the qualitative software Dedoose (Dedoose Version 8.0.35, 2018) and themes were extracted and coded. Findings from the qualitative analysis were consistent with the program logic model and were used to inform the quantitative instrument development

Recruitment and sampling
The treatment group consisted of Fellows; specifically, those who were nominated and participated in at least one Master class. The comparison group was made up of non-fellows, applicants who were nominated but not accepted. Groups were recruited via e-mail and all potential participants were asked to complete an online survey. The total population included all Fellows and non-fellows who applied to The Academy's programming between 2011 and 2020. A total of 1,895 Fellows and 994 non-fellow applicants were approached via e-mail to participate in the study. Ultimately, 711 Fellows (37.5% response rate) and 114 non-Fellows (11% response rate) consented to participate in the study. Differing response rates were primarily driven by outdated contact information for non-fellows as indicated by a higher proportion of e-mail bounce backs (6%; n = 112 in Fellows and 9%; n = 91 in non-Fellows). Survey data from 68 participants were removed from analysis due to missing or duplicate data, rendering 757 participants (675 treatment; 82 comparison) for analysis.
To understand the study outcomes, we used propensity score weights to reduce selection bias and balance the treatment and comparison groups on key baseline participant characteristics in our analyses. A propensity score is a single score representing a participant's probability of receiving treatment given a set of observed characteristics (Austin, 2011;Rosenbaum & Rubin, 1983). Propensity score weights were estimated using logistic regression, through an iterative balance-checking approach. The researchers used propensity scores as sampling weights in a regression modelling approach to create equivalent treatment and comparison groups, allowing for more credible causal inference than standard regression approaches without such adjustments.

Survey
The online self-reported survey (60 items, requiring an average of 9 minutes to complete) was administered between November 2020 and January 2021. The survey asked teachers (Fellows and non-Fellows) to reflect on their last year of full-time teaching, prior to the onset of the COVID-19 pandemic. All scales were either taken directly or modified from previously validated instruments, as described below. Constructs were represented by scales using a 5-point Likert scale with scores ranging from 0-4 (higher scores reflecting more of each construct). The survey measured seven categories: Enthusiasm/Energy for Teaching, Perceived Pride in the Profession, Teacher Efficacy, Integration of Material into the Classroom, Community building, Network Diversity, and Teacher Retention. Each category is described further below. (1) Enthusiasm/Energy for Teaching (3 items; Klassen et al., 2012;National Center for Educational Statistics, 2018) and includes responses such as., "I feel energized about my job;" "I feel a deep connection to my content area." (2) Perceived Pride in the Profession (7 items; Klimek, 2019) and includes responses such as, "I have a positive image of the teaching profession;" "I feel respected as a teacher." (3) Teacher Efficacy (3 items; Bandura, 1998;Ingvarson et al., 2005b) and includes responses such as, "I felt confident incorporating content that I learned through professional development experiences in the classroom;" with the stem "To what extent do you agree or disagree with the following statements about teaching." (4) Integration of Material into the Classroom (5 items), which was measured with the following stem, "To what extent did participation in The Academy/PD lead you to . . . " For example, "To what extent did participation in The Academy/PD lead you to incorporate what you learned as a Fellow/ from PD into your teaching, in terms of content?" and used a 5-point Likert scale with scores ranging from 0-4 with the following choice options (not at all to-a very great extent).
(5) Community building was determined based on the number of educational professionals in a teacher's professional network and (6) the diversity of that network. Participants were asked to report the number of people in various categories using the following stem: "During your LAST school year (2019-2020, prior to the Covid-19 pandemic), how many teachers and other education professionals in the following categories were in your professional network?" (7) Teacher Retention was measured by assessing full-time teaching status in grades K-12 or comparable ungraded levels, using two items to capture the main activity that participants engaged in the last school year. Covariates for teacher retention included two scales. One scale for classroom autonomy (5 items; National Center for Educational Statistics, 2018), such as "How much control did you have in your classroom over the following areas of your planning and teaching (selecting textbooks and other instructional materials; selecting teaching techniques; evaluating and grading students." The other scale assessed individual and school-level factors (6 items) such as, "Necessary materials such as textbooks, supplies, and copy machines are available as needed by staff;" "There is a great deal of cooperative effort among the staff members;" (National Center for Educational Statistics, 2018). Such scales have been demonstrated to be strong predictors of teacher attrition (Borman & Dowling, 2008;Boyd et al., 2011;Nguyen et al., 2020).
The survey captured the three personal demographic characteristics (i.e., age, gender, race/ ethnicity) and the following teaching characteristics: years of teaching experience, type of school (public, private, charter, religious/parochial), Title 1/low-income, grade level (elementary, middle, high, a combination of elementary and middle, a combination of middle and high), and entry into teaching (whether through an alternative certification program or not). We incorporated propensity scores in the regression analyses, including all the aforementioned demographic characteristics in the propensity score model, to account for differences in group composition and balance groups in an effort to equate the Fellow sample with the non-Fellow sample.

Quantitative data analysis
We analyzed data using the statistical software R (R Core Team, 2019). Cronbach's alpha coefficient values for scales ranged from 0.79 to 0.87. We calculated descriptive statistics and compared Fellow and non-Fellow participants on relevant demographic variables both before and after propensity score weighting via the standardized bias method. As a result, we calculated and assessed both unweighted and weighted standardized mean differences for each demographic variable, providing a proxy for baseline equivalence. Guidelines consider standardized mean differences values within ±0.1 as balanced (Stuart et al., 2013). We included stabilized propensity score weights to account for nonrandomization in either group. Given an appropriate specification of the propensity score model, propensity scores balance the two grouping conditions (i.e., Fellows and non-Fellows) on key demographic information and create an approximation of a random experiment. Furthermore, the stabilization of the propensity scores helped to address extreme propensity score weights that could otherwise have biased the results.
Outcome means for the dependent variables were calculated by group status to determine if descriptive differences existed for enthusiasm/energy for teaching, perceived pride in the profession, teacher efficacy, the extent to which teachers reported integrating material from PD into their classroom, and the number and diversity of education professionals in a teacher's professional network. Subsequently, we performed weighted simple linear regression analyses using the estimated propensity scores as sampling weights. Later, we accounted for relevant covariates by building weighted multiple linear regression models, through a step-by-step iterative process.
To examine the effect of The Academy's programming on teacher retention, we used a logistic regression model for the dichotomous dependent variable (teacher retained-yes/no) and included years of teaching experience, the classroom autonomy scale, and teacher, and school-level factors scale as covariates. We built logistic regression models both with and without stabilized propensity score weights. In Table 2, we present themes and illustrative quotes. Fellows revealed that the following outcomes were generated from their participation in The Academy's professional development: network building, bringing content to the classroom and the school, pride in the profession, teacher efficacy and professional affirmation, and feeling like a student. Additionally, many Fellows identified which program experiences mediated their decisions to continue teaching. We describe each of these outcomes in the following paragraphs.

In-depth interviews
Network Building. Fellows indicated that building a community of teachers and providing teachers with an exposure to a diverse community was the most prominent outcome of being a Fellow of The Academy for Teachers. Fellows felt that exposure to a community of teachers provided them with opportunities to connect with others with similar experiences, which in turn provided inspiration and support. Fellows were able to share best practices and novel ideas to improve their teaching practice, had a social forum with other teachers that disrupts the isolation of the job, and found a support network to share frustrations and affirm experiences. Exposure to a diversity of teachers (e.g., those teaching at different types of schools, teaching in different boroughs, with different grade levels and subjects) provided a unique forum to learn from and gain insights from different perspectives, philosophies, and experiences, which was something that Fellows highly valued, particularly private/independent school Fellows.
Bringing it Back to the Classroom and School. Almost all Fellows spoke about bringing elements of the Fellow experience back to their students and schools. Specific features of the Academy's programming facilitated easy integration, including the content-focus of the master classes, ideas and inspirations from the Pedagogical Potluck, opportunities to engage in planning and idea generation in group settings, and interstitial conversations in social settings with other teachers. In some cases, Fellows were able to support policy and program changes at the school level based on their experiences (e.g., instituting a Pedagogical Potluck and other resource-sharing practices, creating courses, and modifying curricula/syllabi).
Pride in the Profession. Fellows indicated that The Academy builds pride in teaching by making educators feel special, giving them opportunities to engage with exceptional speakers, and inviting them to settings that are distinguished and rarified. Specific examples include being nominated and accepted through a competitive process, gourmet catered breakfast and lunch, personalized VIP attention (staff referring to teachers by name as they enter the room, formal introductions at the start of the program by the program president, printed name cards on quality stock, etc.), the status and prestige of the speakers, and experiences in elevated spaces with beauty and history, among others that aren't typical to a normal teacher's work day or professional development experiences.
Teacher Efficacy and Affirmation. For some Fellows, The Academy helped promote confidence in teaching and affirmed their decision to become a teacher. These feelings were primarily driven by the inclusion and acceptance to The Academy programming and the cadre of other exceptional teachers.
Feeling Like a Student. Fellows discussed how "feeling like a student" for the day was valuable and impactful to their teaching practice by giving them greater empathy for and connection to the student experience. They indicated that being in a room alongside other accomplished peers in a learning setting can cultivate feelings similar to those that students experience in a classroom.
Keeping Teachers in the Profession. While The Academy's PD was described as a contributing factor to staying in the profession, most Fellows indicated it wasn't the only or the most prominent factor in their decision to remain in the profession. According to Fellows, The Academy contributes to their retention by providing a community of teachers to share ideas and resources with, by

Feeling Like a Student
It's about being brave. You are going to be in a class, not that there's any kind of competition, but you're going to be in a class of people who are smarter than you, who are more experienced than you are and you have to be able to walk into the room and not know what's coming and I'm going to bring my best self. And that's what I ask of my students every day. So it lets me walk the walk, not just talk the talk whenever I'm in one of those classes, and I think it's so important to have that perspective.

Public High School Art-Cinema Teacher
It is a day for you, as teachers, to really just be students and to learn and to make art and that was so cool.

(Continued)
energizing and exciting them about teaching and their content area, by respecting them, and building a sense of pride for the profession. Subgroup analysis by school type demonstrated that for public school teachers, increasing pride in the profession was a meaningful driver of teacher retention compared to private school teachers. Table 3 presents differences between the eight demographic and teaching characteristics in the two groups. There were statistically significant differences in the proportion of teachers at religious/parochial schools, with a larger proportion of non-Fellows, n = 6 (7%) than Fellows, n = 7 (1%) teaching at this type of school.

Theme Illustrative Text
Providing a Community That community that occurs with a group of teachers and that community that occurs in a classroom with students, that's really important to me and it's part of the reason that I love teaching, that interaction with people. And, if I get opportunities to do that, with a wider base, I feel like that's broadening my community and making me even bigger and that's something that Academy for Teachers allowed me to do.

Public High School Science Teacher
One of the things I like the most about teaching or one of the reasons that I am a science teacher and not working in a lab is because I like to interact with people very much. I think it [The Academy] contributes to those factors that I wouldn't perhaps be exposed to . . . a sense of community, that wouldn't exist if I were working in a lab or doing something else relating to science.

Public High School Science Teacher
Generating Motivation and Excitement For me, it's being able to take that energy and excitement back to the classroom [from The Academy's PD experience]. I haven't done the empirical studies on it, but I would think that a more interested and excited teacher would be able to translate that to the students a bit more and hopefully make them more interested and excited to learn whatever it is.

Public Middle School Science Teacher
Feeling Appreciation and Value as a Professional I know that there are people out there that appreciate us. That's big. I hear the rhetoric so much, thank you for what you do and we're all over the newspaper, especially in the fall, and I guess the academy actually put their money where their mouth is. They're like, we appreciate you and we want to give you this, we're here to give you this to get you together in this course and here you go, thank you for doing what you're doing.

Public High School Art Teacher
Being treated as a professional keeps me teaching. I think that's really, really, important. Being given the time and space to improve your practice and your content knowledge, which is a requirement for many other profession[s], like doctors and lawyers. And the acknowledgement that that takes time and shouldn't be something that you're just cramming in on the weekends or during your summer, rather during the time that you're working, and electing to participate in some activities that will increase your content knowledge and pedagogy.

Public High School Science Teacher
Participants were asked a series of questions pertaining to their teaching and professional development experiences. The majority of teachers (n = 538; 71%) taught at public school, followed by 22% (n = 167) at private school; the remaining taught at public-charter or religious/ parochial school. The mean total years of teaching was 15 (±9.1) years, with a range from 3 to 53 years. The majority of participants taught at the secondary level, with 61% (n = 465) teaching in high school only, and 18% (n = 137) and 11% (n = 86) teaching middle school and elementary, respectively. The remaining teachers taught multiple groups (e.g., combination of elementary and middle, middle and high, or other). All other demographic characteristics are presented in Table 3. Results suggest using propensity score analysis reduced bias by improving covariate balance.
Findings from the linear and logistic regression models are presented in Tables 4 and 5 respectively.

Enthusiasm/energy for teaching
Generally, most teachers did not have a lot of enthusiasm and energy for teaching. Fellows were slightly more enthusiastic/had more energy for teaching (M = 2.89, SD = 1.24) than non-Fellows (M = 2.34, SD = 1.59), however, these findings were not statistically significant (p = 0.78) with small effect sizes (−0.02), after controlling for years of experience, school type, individual level factors, classroom autonomy, and age. See, Table 4.

Perceived pride in profession
There were statistically significant differences between Fellows (M = 2.34, SD = 1.10) and non-Fellows (M = 1.89, SD = 1.36). Fellows showed greater perceived pride with a significant moderate effect size of 0.23 (p = 0.03), after controlling for school type. See, Table 4.

Integration of material into the classroom
Fellows (M = 1.88, SD = 1.11) reported incorporating more material (e.g., content, pedagogy, resources) from their PD experience with The Academy than non-Fellows (M = 1.49, SD = 1.28) did with alternative PD when controlling for years of teaching and classroom autonomy. This effect was statistically significant (p < 0.01) with a moderate effect size (0.45). See, Table 4.

Professional network size
Fellows had an average of 24 (±26.11) people in their professional networks, while non-Fellows had 16 (±27.61). There were no statistically significant differences between Fellows and non-Fellows in the number of education professionals in their professional network (p = 0.11) when controlling for years of teaching, type of school, low-income designation, individual-level factors, and classroom autonomy. See, Table 4.

Professional network diversity
Similarly, there were no statistically significant differences (p = 0.27) between Fellows (M = 47.46, SD = 145.06) and non-Fellows (M = 25.38,SD = 45.69) in the diversity of their professional networks when controlling for years of teaching, type of school, individual-level factors, classroom autonomy, and teacher gender. See, Table 4.

Teacher retention
Logistic regressions with and without weights demonstrated that there is a statistically significant relationship between teacher retention and participating as a Fellow (see , Table 5). Fellows are 3.71 times as likely to be retained than non-Fellows. The predicted probability of staying in teaching for Fellows is 0.95 compared to non-Fellows 0.84, demonstrating that Fellows have an 11% higher probability of being retained in the teaching profession that non-fellows. See, Table 5.

Discussion
As other studies have demonstrated, teacher retention has been a persistent issue in the education sector for decades (Ingersoll, 1994;Nguyen et al., 2020;Sutcher et al., 2016). Furthermore, the COVID-19 pandemic has highlighted the importance and value of teachers, as well as the profound retention challenges the profession continues to face. Paying attention to teacher development across a teacher's career life cycle and providing a host of different supports to account for the varied factors that contribute to teacher retention are critical.
The Academy for Teachers provides content-focused intensive workshops led by renowned experts held in exclusive/beautiful spaces, as well as bespoke social events, activities, and workshops designed to provide desirable experiences and build a diverse community of exceptional professionals. The Academy for Teachers' program improves teacher retention, increases integration of useful program materials into classrooms, and builds teacher efficacy and pride in the profession. Furthermore, The Academy presents a unique model that combines multiple elements of formal and informal professional development to focus on elevating teacher experiences holistically-treating teachers as resources to invest in rather than problems to be solved. In doing so, this PD approach not only builds upon the principles of effective PD but also considers personal, interpersonal, external, and school factors related to teacher retention. Most existing professional development research is limited to assessing the outcomes of formal experiences and no studies to date have looked at holistic PD models and their impacts on teacher retention.
Of the PD studies that link directly to teacher retention and have found positive effects, the PD has typically been limited to select groups of teachers, such as early career or public school teachers (DeAngelis & Presley, 2011;Reynolds & Wang, 2005) or a single-subject area teachers (Allen & Sims, 2017). This study demonstrates that The Academy's model for professional development can be successfully employed across a large and diverse group of teachers (from different types of schools, grade levels, subjects, and career spans); fulfills teachers' needs for membership in a robust community of teachers and exposes teachers to different perspectives.
Professionalization has long been a source of both hope and frustration for teachers. Teaching is often perceived as undesirable by teachers and their surrounding community due to a lack of prestige, commensurate wages, and limited career advancement opportunities and professional development (Glennie et al., 2016). Other jobs that are considered prestigious offer access to stately environments with gourmet catering and other exclusive experiences, and employee perks to signal appreciation. However, within professional development programs, intentionally promoting teacher respect is often an ignored variable. This study provides insight into the alchemy needed to build a sense of pride in the profession among teachers, namely through investments in teachers. As demonstrated in this study, Fellows identified that feeling greater pride for teaching, through The Academy's programming, was a contributing factor to their intention to remain in the teaching profession. Other studies (Easley, 2006(Easley, , 2008Rots et al., 2010) have similarly demonstrated that respect for teachers as professionals is an important contributing factor to their remaining in the teaching profession because it nurtures their sense of professional value and worth.
As with other causal inference research, this study provides evidence for the effectiveness of a PD model. However, it is also important to understand how and why this approach worked, which is not typically investigated alongside causal inquiries. Although the PD was described as a contributing but not driving factor in Fellows' intentions to remain in the profession, which aligns with other research (Coldwell, 2017;DeAngelis & Presley, 2011), this study provides potential mediators for teacher retention that are worth exploring further. Namely, professional development experiences that honor and respect teachers, energize and excite them about their content area, provide teachers with opportunities to be students for a day, and provide them access to a community of diverse colleagues. The results reported here are subject to a few limitations. Despite applying weighted propensity scores to equalize groups, these results are constrained by the use of a post-only study design which is vulnerable to internal validity threats. Also, the study pertains to a specific population of teachers, nominated because they are considered exceptional amongst their colleagues and therefore may not be generalizable to the broader population of teachers. Although The Academy presents a model for professional development that is open to all teachers (teaching all grade levels, from different types of schools, etc.), the subject-specific design may also render the program less relevant to elementary school teachers. Deeper investigation, through subgroup analysis, was not possible due to the small proportion of teachers from elementary schools and warrants further exploration.
This study provides a foundation for further research to explore with teachers at different career stages, across a national context, and to investigate the impact on student outcomes. In addition, further research is needed to assess how each teacher outcome assessed here explains variance in teacher retention. Finally, these findings have implications for policy makers, urban school reformers, teacher educators, professional developers, and administrators. Combining multiple professional supports to intentionally treat teaching professionals holistically, considering personal, interpersonal, school-level, and external factors, appears to be a promising approach.

Author's contributions
HBL & JEL conceptualized the study (leading instrument development, data collection and analysis) and led the development and drafting of the manuscript providing their expertise in outcome evaluations and education research. MJC was involved in manuscript development and review, providing insight into the study's implications for policy. JL & KGD supported study design efforts, analyzed data, and also contributed to the manuscript development and review providing their expertise in education causal inference research.