View in EDS HTML Full Text PDF Full Text

Subject-Area Specialization and Teacher Retention: An Elementary School Story

Saved in:
Bibliographic Details
Title: Subject-Area Specialization and Teacher Retention: An Elementary School Story
Language: English
Authors: Bastian, Kevin C., Fortner, C. Kevin, Caton, Kate
Source: Elementary School Journal. Dec 2023 124(2):343-366.
Availability: University of Chicago Press. Journals Division, P.O. Box 37005, Chicago, IL 60637. Tel: 877-705-1878; Tel: 773-753-3347; Fax: 877-705-1879; Fax: 773-753-0811; e-mail: subscriptions@press.uchicago.edu; Web site: http://www.press.uchicago.edu
Peer Reviewed: Y
Page Count: 24
Publication Date: 2023
Document Type: Journal Articles
Reports - Evaluative
Education Level: Elementary Education
Descriptors: Elementary School Teachers, Elementary Schools, Intellectual Disciplines, Specialization, Teacher Persistence, Specialists, Teacher Role
Geographic Terms: North Carolina
DOI: 10.1086/727503
ISSN: 0013-5984
1554-8279
Abstract: School leaders need effective, affordable approaches to retain their teacher workforce. We investigated a promising, low-cost option for school leaders to encourage teacher retention: subject-area specialization in elementary grades (K-5). Using data on North Carolina elementary grades teachers and schools in the 2011-2012 through 2015-2016 academic years, we track the incidence of subject-area specialization, assess whether teaching in a specialist role promotes retention, and examine whether subject-area specialization is an effective retention strategy for certain schools and teachers. Descriptive analyses show specialization is common in upper elementary grades and has become a more widely used assignment strategy over time. Retention analyses indicate that elementary grades teachers are more likely to return to the same school after becoming a specialist. These results vary by school and teacher characteristics, suggesting that specialization may be a more effective retention strategy in urban schools, in non-high-need schools, and for Black teachers.
Abstractor: As Provided
Entry Date: 2023
Accession Number: EJ1401768
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwEXi0a1Y-gKNaIDcKRAzqNeAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDLb6J7DCh92vXTbJDQIBEICBm8K-GudQaDFk_gh2D5ObsFTjYH9coXDUzL1ChzLm5D4n5ptU65v3XJEo9p6sBqG18nkoaBUCmSMIwi2aeRVNi4Zzx2HJRjFCA8xqYaqBiSDg6EzCKYdKVKmSn-4-3sWW30wGJGVs4VkYOYvAbMS3ThnedRlUcdeUOv_cp9XhfAiDkD3ncT8h9_EmprLJ4XHRTLVcz7ORWNaIasQL
Text:
  Availability: 1
  Value: <anid>AN0174564218;esj01dec.23;2024Jan04.05:29;v2.2.500</anid> <title id="AN0174564218-1">Subject-Area Specialization and Teacher Retention: An Elementary School Story </title> <p>School leaders need effective, affordable approaches to retain their teacher workforce. We investigated a promising, low-cost option for school leaders to encourage teacher retention: subject-area specialization in elementary grades (K–5). Using data on North Carolina elementary grades teachers and schools in the 2011–2012 through 2015–2016 academic years, we track the incidence of subject-area specialization, assess whether teaching in a specialist role promotes retention, and examine whether subject-area specialization is an effective retention strategy for certain schools and teachers. Descriptive analyses show specialization is common in upper elementary grades and has become a more widely used assignment strategy over time. Retention analyses indicate that elementary grades teachers are more likely to return to the same school after becoming a specialist. These results vary by school and teacher characteristics, suggesting that specialization may be a more effective retention strategy in urban schools, in non-high-need schools, and for Black teachers.</p> <p>Each year, approximately 13% of the US teacher workforce—450,000 teachers—either switches schools or leaves the teaching profession (Alliance for Excellent Education, [<reflink idref="bib1" id="ref1">1</reflink>]). This attrition has significant financial costs for school districts and disruptive effects on schools and students. Estimates indicate that teacher turnover costs billions annually, as districts must pay to recruit, hire, and train new personnel (Alliance for Excellent Education, [<reflink idref="bib1" id="ref2">1</reflink>]). Furthermore, several studies show that teacher turnover is associated with reduced student achievement in mathematics and reading (Hanushek et al., [<reflink idref="bib21" id="ref3">21</reflink>]; Ronfeldt et al., [<reflink idref="bib41" id="ref4">41</reflink>]). Teacher turnover often weakens school culture and teacher peer relationships and results in schools reassigning teachers to different grades/subjects and hiring less qualified teachers (Hanushek et al., [<reflink idref="bib21" id="ref5">21</reflink>]; Sorensen & Ladd, [<reflink idref="bib46" id="ref6">46</reflink>]). These consequences of turnover are especially concerning as attrition rates are higher in high-poverty, high-minority, and low-performing schools (Holme et al., [<reflink idref="bib24" id="ref7">24</reflink>]) and among early career and minority teachers (Goldring et al., [<reflink idref="bib19" id="ref8">19</reflink>]; Sun, [<reflink idref="bib48" id="ref9">48</reflink>]). This suggests that low-income and minority students are disproportionately affected by teacher attrition.</p> <p>School districts and schools possess a range of options to ameliorate these teacher turnover concerns. These include induction and mentoring programs for beginning teachers (Ronfeldt & McQueen, [<reflink idref="bib42" id="ref10">42</reflink>]; Smith & Ingersoll, [<reflink idref="bib45" id="ref11">45</reflink>]), teacher salary increases (Borman & Dowling, [<reflink idref="bib7" id="ref12">7</reflink>]; Imazeki, [<reflink idref="bib26" id="ref13">26</reflink>]), improving the quality of school leadership and school working conditions (Boyd et al., [<reflink idref="bib8" id="ref14">8</reflink>]; Ladd, [<reflink idref="bib31" id="ref15">31</reflink>]), and strategic teacher assignments (Bastian & Janda, [<reflink idref="bib5" id="ref16">5</reflink>]; Donaldson & Johnson, [<reflink idref="bib10" id="ref17">10</reflink>]; Ost & Schiman, [<reflink idref="bib38" id="ref18">38</reflink>]). Although research supports the promise of these approaches, districts and schools must be mindful of the costs and ease with which solutions can be implemented. From this perspective, strategic teacher assignments stand out as low-cost, readily implementable approaches to retain teachers.</p> <p>In the present study, we use statewide administrative data from North Carolina to focus on a specific strategic assignment practice: school leaders assigning elementary grades teachers to be subject-area specialists. Although subject-area specialization is common for teachers in middle and high school grades, less is known about specialization in elementary grades, where state licensure policies and assignment practices often result in teachers serving as generalists in self-contained classrooms.[<reflink idref="bib2" id="ref19">2</reflink>] Specialization offers elementary grades teachers the opportunity to teach fewer subject areas—presumably subjects in which they are relatively more effective (Bastian & Fortner, [<reflink idref="bib4" id="ref20">4</reflink>]; Fox, [<reflink idref="bib13" id="ref21">13</reflink>]; Goldhaber et al., [<reflink idref="bib18" id="ref22">18</reflink>]). This may benefit teachers' self-efficacy, job satisfaction, and retention. To help districts and schools make evidence-based decisions about specialization as a low-cost teacher retention strategy, we answer the following questions:</p> <p></p> <p>• 1.</p> <p></p> <ulist> <item> What patterns of elementary grades specialization exist, and how are these patterns changing over time?</item> <p></p> </ulist> <p>• 2.</p> <p></p> <ulist> <item> Does assignment to specialization predict an increased likelihood of returning to teach?</item> <p></p> </ulist> <p>• 3.</p> <p></p> <ulist> <item> Do teacher retention results vary by school or teacher characteristics?</item> </ulist> <p>Our work documents the incidence and patterns in elementary grades specialization in a large and diverse state and isolates the associations between specialization and retention. Importantly, we also assess whether specialization alters teacher retention in high-need schools and based on observable characteristics of teachers. These analyses are particularly relevant given the concentration of turnover in high-need schools and the need to develop a diverse, experienced, and effective teacher workforce. Overall, our analyses build upon a large body of prior work connecting school culture to teacher retention and add to a growing body of research on teacher assignments and teacher outcomes (Atteberry et al., [<reflink idref="bib3" id="ref23">3</reflink>]; Bastian & Janda, [<reflink idref="bib5" id="ref24">5</reflink>]; Blazar, [<reflink idref="bib6" id="ref25">6</reflink>]; Ost & Schiman, [<reflink idref="bib38" id="ref26">38</reflink>]).</p> <hd id="AN0174564218-2">Background</hd> <p></p> <hd id="AN0174564218-3">School Organizational Conditions and Teacher Turnover</hd> <p>Research shows that a wide array of factors are associated with teacher turnover. However, among the many explanations for turnover, recent work highlights the role of working conditions and school leaders in teacher attrition decisions (Boyd et al., [<reflink idref="bib8" id="ref27">8</reflink>]; Kraft et al., [<reflink idref="bib30" id="ref28">30</reflink>]; Ladd, [<reflink idref="bib31" id="ref29">31</reflink>]; Simon & Johnson, [<reflink idref="bib44" id="ref30">44</reflink>]). Teachers prefer to work in schools where they have classroom autonomy and the ability to influence/make decisions, where school leaders provide support and resources, where student behavior is well managed and teachers feel safe, and where there are meaningful opportunities for peer collaboration and professional development (Boyd et al., [<reflink idref="bib8" id="ref31">8</reflink>]; Ingersoll et al., [<reflink idref="bib27" id="ref32">27</reflink>]). Johnson et al. ([<reflink idref="bib29" id="ref33">29</reflink>]) posit that rather than the demographics of the K–12 student population, school working conditions—positive work environments, school culture, leadership, and collegial relationships—are the strongest predictors of whether teachers will remain at a school. School leaders have a direct effect on each of these measures of working conditions.</p> <p>The impact of school leaders is substantiated by studies connecting working conditions surveys to the retention outcomes of teachers. In New York City (Boyd et al., [<reflink idref="bib8" id="ref34">8</reflink>]; Kraft et al., [<reflink idref="bib30" id="ref35">30</reflink>]) and North Carolina (Ladd, [<reflink idref="bib31" id="ref36">31</reflink>]), teachers' perceptions of school leaders are the strongest predictor of teacher attrition. This finding is robust across a range of specification checks and validity tests, suggesting a causal relationship between school leadership quality and teacher retention. Recent work by Grissom and Bartanen ([<reflink idref="bib20" id="ref37">20</reflink>]) documents strategic teacher retention by school principals: schools with more effective principals experience less turnover among high-performing teachers and more turnover among the lowest-performing teachers. Finally, Nguyen et al. ([<reflink idref="bib36" id="ref38">36</reflink>]) have identified several determinants of teacher attrition, including teacher-principal race and gender matching, accountability policies, school improvement and reform measures, and employment opportunities within and outside the teaching profession. This work continues to underscore the influence of school leaders on school climate and the importance of relationships between school leaders and teachers to teacher retention.</p> <p>One way in which school leaders affect teacher working conditions is the assignment of teachers to classrooms. Here, it is worth noting that school leaders often make classroom assignment decisions with the input of teachers and other school personnel. As such, teachers may play a role in strategic classroom assignment decisions. A theory of action connecting classroom assignments to teacher retention is straightforward: when school leaders assign teachers to classes that are more demanding—based on student characteristics, gaps in teachers' knowledge/skills, or switching grade levels or subject areas—teachers will be more likely to leave the school or the profession. Empirical evidence supports this hypothesis. Donaldson and Johnson ([<reflink idref="bib10" id="ref39">10</reflink>]) show that Teach for America corps members are more likely to leave their school or resign from teaching when they have more challenging assignments. Likewise, Ost and Schiman ([<reflink idref="bib38" id="ref40">38</reflink>]) find that elementary school teachers with stable grade-level assignments have lower levels of turnover than teachers who frequently switch grades. For example, those who teach the same grade in their first 2 years of teaching are 20% more likely to remain in teaching than peers teaching different grades. These links between classroom assignments and retention suggest that subject-area specialization may be an effective strategy to retain elementary school teachers.</p> <hd id="AN0174564218-4">Research on Subject-Area Specialization and Outcomes for Teachers and Schools</hd> <p>To date, research on subject-area specialization has focused on which teachers specialize and whether specialization benefits teacher effectiveness, student achievement, and other school outcomes. Regarding who specializes, analyses from North Carolina show that principals assign more effective teachers to specialization. Relative to generalists in the same school, first-time mathematics specialists have lagged value-added 20% of a standard deviation higher in mathematics; first-time specialists in reading and science have lagged value-added 11% and 16% of a standard deviation higher, respectively (Bastian & Fortner, [<reflink idref="bib4" id="ref41">4</reflink>]). This finding is consistent with research showing that principals can accurately identify highly effective teachers (Jacob & Lefgren, [<reflink idref="bib28" id="ref42">28</reflink>]) and suggests principals' ability to leverage data for strategic teacher assignments.</p> <p>Economic theory posits that subject-area specialization will boost student achievement by providing teachers opportunities to acquire more job-specific human capital (Atteberry et al., [<reflink idref="bib3" id="ref43">3</reflink>]; Cook & Mansfield, [<reflink idref="bib9" id="ref44">9</reflink>]; Ost, [<reflink idref="bib37" id="ref45">37</reflink>]) and by connecting teachers who are highly effective in a particular subject area to more students. Regarding job-specific human capital, a range of studies show that teachers are more effective with greater knowledge of and experience with the content they are teaching (Cook & Mansfield, [<reflink idref="bib9" id="ref46">9</reflink>]; Hill et al., [<reflink idref="bib23" id="ref47">23</reflink>]; Myrberg et al., [<reflink idref="bib34" id="ref48">34</reflink>]). Likewise, simulation studies suggest that elementary grades specialization can result in meaningful student achievement gains (Fox, [<reflink idref="bib13" id="ref49">13</reflink>]; Goldhaber et al., [<reflink idref="bib18" id="ref50">18</reflink>]).</p> <p>Despite this theory of action, recent analyses indicate that specialization is not leading to its theorized benefits for student achievement. In a randomized experiment in Houston, Fryer ([<reflink idref="bib14" id="ref51">14</reflink>]) found that specialization had a negative impact on student achievement. Pooled across years, students attending treatment elementary schools scored 11% of a standard deviation lower on high-stakes assessments and 10% of a standard deviation lower on low-stakes assessments (Fryer, [<reflink idref="bib14" id="ref52">14</reflink>]). Analyses from North Carolina show that teachers are less effective in mathematics and reading after becoming subject-area specialists. Because principals assign more effective teachers to specialize, these drops in teacher effectiveness do not result in lower mathematics and reading achievement at the school level (Bastian & Fortner, [<reflink idref="bib4" id="ref53">4</reflink>]). Similarly, in Indiana, Hwang and Kisida ([<reflink idref="bib25" id="ref54">25</reflink>]) found that specialization leads to lower levels of teaching effectiveness in math and reading, especially when teaching students who are more likely to face obstacles in school. Increasing the number of specialist teachers does not predict measures of school quality (Hwang & Kisida, [<reflink idref="bib25" id="ref55">25</reflink>]).</p> <hd id="AN0174564218-5">Subject-Area Specialization and Potential Impacts on Teacher Retention</hd> <p>The organization of elementary grades classrooms can range from fully self-contained, where a single instructor is responsible for all academic content, to highly departmentalized models, with single-subject instruction from different teachers. Classroom organization is also influenced by philosophical considerations, child development, and the historical foundations of schooling (Parker et al., [<reflink idref="bib39" id="ref56">39</reflink>]). Each of these approaches has potential advantages and drawbacks. Self-contained classrooms may make it easier for teachers to attend to the whole child (i.e., intellectual, emotional, physical), to develop stronger relationships with students, and to facilitate connected and continuous learning for students (Parker et al., [<reflink idref="bib39" id="ref57">39</reflink>]). Specifically, generalists may find it easier to consistently enforce rules and procedures and may be more skilled at tailoring lessons to the needs of students, as they spend more sustained time with students and can better integrate content for the same students throughout the day (Anderson, [<reflink idref="bib2" id="ref58">2</reflink>]; Fryer, [<reflink idref="bib14" id="ref59">14</reflink>]).</p> <p>Whereas generalists must prepare lesson plans, assignments, and assessments and deliver instruction in all academic content areas, specialists can focus on a limited number of subjects. Specializing teachers may have more opportunities to develop subject-specific human capital through on-the-job learning, collaborating with colleagues, and targeted professional development. In particular, Markworth et al. ([<reflink idref="bib32" id="ref60">32</reflink>]) found specialization benefits the planning, instructional time, and professional development for math and science teachers. As a result, we view specialization as a nonpecuniary job benefit that may help teachers experience less job-related stress and more self-efficacy and job satisfaction. We contend that these feelings will influence teachers' retention decisions (Perrachione et al., [<reflink idref="bib40" id="ref61">40</reflink>]; Stockard & Lehman, [<reflink idref="bib47" id="ref62">47</reflink>]). Specifically, we hypothesize that specialization will encourage teachers to remain at the same school but will have less of an impact on whether individuals remain in the teaching profession. We make this hypothesis because the mechanism linking specialization to retention is teacher assignment practices within a school building. That is, specialization may be a nonpecuniary job benefit that increases a teacher's satisfaction with a specific school site but not the profession overall. Furthermore, the effects of specialization may differ across different types of schools or teachers. For example, subject-area specialization may not be a large enough nonpecuniary inducement to keep teachers in high-need schools. As such, we perform separate analyses to assess whether the effects of specialization on teacher retention differ across school environments or observable teacher characteristics.</p> <p>Overall, our work addresses an important gap in the research literature, with a recent meta-analysis arguing that there are not enough robust studies to determine whether specialization benefits teacher retention (See et al., [<reflink idref="bib43" id="ref63">43</reflink>]). We begin to answer this retention question and provide empirical evidence regarding the use of and patterns in elementary grades specialization over time. This work may be particularly timely given reports highlighting the increasing use of specialization in response to school accountability systems and more rigorous learning standards (Gewertz, [<reflink idref="bib17" id="ref64">17</reflink>]).</p> <hd id="AN0174564218-6">Data</hd> <p></p> <hd id="AN0174564218-7">Research Sample</hd> <p>The research sample for the present study includes academic content area teachers (i.e., reading, math, science, and social studies) in grades K–5 in North Carolina public schools (NCPS) during the 2011–2012 through 2015–2016 school years. To be included in the analysis, individuals must be paid as a teacher and be listed as the teacher-of-record in classroom roster files. This full sample consists of nearly 54,000 unique teachers and 170,000 teacher-year records. For our descriptive analyses on the use and patterns of subject-area specialization, we focus on teachers in grades K–2 (early elementary grades) and teachers in grades 3–5 (upper elementary grades). For our empirical analyses on specialization and retention, we estimate models across grades K–5, with separate models focused on grades K–2 and grades 3–5. By separating these samples, we assess whether the relationship between specialization and teacher retention differs across grade levels. Furthermore, our upper elementary grades sample allows us to focus on grade levels in which school leaders may be more strategic in their teacher assignment practices due to the presence of statewide assessments and school accountability systems (Fuller & Ladd, [<reflink idref="bib15" id="ref65">15</reflink>]). Our teacher sample in grades K–2 consists of approximately 32,000 unique teachers and 87,000 teacher-year records; our teacher sample in grades 3–5 consists of approximately 32,000 unique teachers and 84,000 teacher-year records.[<reflink idref="bib3" id="ref66">3</reflink>]</p> <p>Table 1 displays teacher and school-level descriptive data for our analytical sample. We present these characteristics for all K–5 teachers, for specialists in grades K–5, and for generalists in grades K–5. We define specialists as those teaching one or two of the four academic content areas in a given year; generalists teach three or four of the academic content areas in a given year. Relative to generalists, we find that specialists are less likely to be female, are older, have more years of experience, are more likely to have a graduate degree and National Board Certification (NBC), and are more likely to be teaching multiple grades and to be new to a grade level. Considering school-level characteristics, we find that specialists are much more likely to work in rural/town settings than in city/suburb settings. Although the mechanisms explaining this urban/rural difference are unclear, the higher concentration of specialists in rural environments may suggest challenges in attracting personnel to rural schools (Monk, [<reflink idref="bib33" id="ref67">33</reflink>]).</p> <p>Table 1. Teacher and School-Level Characteristics</p> <p> <ephtml> <table><thead><tr><th align="left" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Characteristics</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">K–5 Overall</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">K–5 Specialists</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">K–5 Generalists</th></tr></thead><tbody><tr><td valign="bottom" rowspan="1" colspan="1">Female (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">94.88</td><td char="." valign="bottom" rowspan="1" colspan="1">92.34**</td><td char="." valign="bottom" rowspan="1" colspan="1">95.29</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Minority (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">15.75</td><td char="." valign="bottom" rowspan="1" colspan="1">15.66</td><td char="." valign="bottom" rowspan="1" colspan="1">15.77</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Age (years)</td><td char="." valign="bottom" rowspan="1" colspan="1">40.16</td><td char="." valign="bottom" rowspan="1" colspan="1">41.28**</td><td char="." valign="bottom" rowspan="1" colspan="1">39.98</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Years of experience</td><td char="." valign="bottom" rowspan="1" colspan="1">11.20</td><td char="." valign="bottom" rowspan="1" colspan="1">12.07**</td><td char="." valign="bottom" rowspan="1" colspan="1">11.06</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Graduate degree (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">32.87</td><td char="." valign="bottom" rowspan="1" colspan="1">36.15**</td><td char="." valign="bottom" rowspan="1" colspan="1">32.33</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">NBC (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">12.10</td><td char="." valign="bottom" rowspan="1" colspan="1">13.29**</td><td char="." valign="bottom" rowspan="1" colspan="1">11.91</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Alternative entry (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">6.20</td><td char="." valign="bottom" rowspan="1" colspan="1">6.42</td><td char="." valign="bottom" rowspan="1" colspan="1">6.17</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Multigrade teacher (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">8.76</td><td char="." valign="bottom" rowspan="1" colspan="1">14.80**</td><td char="." valign="bottom" rowspan="1" colspan="1">7.77</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">New to grade (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">16.33</td><td char="." valign="bottom" rowspan="1" colspan="1">18.63**</td><td char="." valign="bottom" rowspan="1" colspan="1">15.98</td></tr><tr><td colspan="4" valign="bottom" rowspan="1">School characteristics:</td></tr><tr><td valign="bottom" rowspan="1" colspan="1"> City/suburb (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">47.19</td><td char="." valign="bottom" rowspan="1" colspan="1">35.17**</td><td char="." valign="bottom" rowspan="1" colspan="1">49.14</td></tr><tr><td valign="bottom" rowspan="1" colspan="1"> Rural/town (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">52.81</td><td char="." valign="bottom" rowspan="1" colspan="1">64.82**</td><td char="." valign="bottom" rowspan="1" colspan="1">50.86</td></tr><tr><td valign="bottom" rowspan="1" colspan="1"> School size</td><td char="." valign="bottom" rowspan="1" colspan="1">570.44</td><td char="." valign="bottom" rowspan="1" colspan="1">566.44**</td><td char="." valign="bottom" rowspan="1" colspan="1">571.10</td></tr><tr><td valign="bottom" rowspan="1" colspan="1"> Minority (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">51.75</td><td char="." valign="bottom" rowspan="1" colspan="1">51.36*</td><td char="." valign="bottom" rowspan="1" colspan="1">51.81</td></tr><tr><td valign="bottom" rowspan="1" colspan="1"> Economically disadvantaged (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">60.97</td><td char="." valign="bottom" rowspan="1" colspan="1">61.10</td><td char="." valign="bottom" rowspan="1" colspan="1">60.95</td></tr><tr><td valign="bottom" rowspan="1" colspan="1"> Test proficiency rate</td><td char="." valign="bottom" rowspan="1" colspan="1">58.92</td><td char="." valign="bottom" rowspan="1" colspan="1">58.75</td><td char="." valign="bottom" rowspan="1" colspan="1">58.95</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Teacher-year records</td><td valign="bottom" rowspan="1" colspan="1">167,853</td><td valign="bottom" rowspan="1" colspan="1">23,541</td><td valign="bottom" rowspan="1" colspan="1">144,312</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Unique teachers</td><td valign="bottom" rowspan="1" colspan="1">53,550</td><td valign="bottom" rowspan="1" colspan="1">13,020</td><td valign="bottom" rowspan="1" colspan="1">49,078</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>1 Note. This table displays descriptive data on elementary grades teachers and the schools in which they work. NBC = National Board Certification. * and ** indicate statistically significant differences between specialists and generalists at the 0.05 and 0.01 levels, respectively.</p> <hd id="AN0174564218-8">Coding Subject-Area Specialization</hd> <p>We leverage classroom roster data from the North Carolina Department of Public Instruction to determine whether an elementary-grades teacher is a subject-area specialist. Classroom roster files link students to teachers and contain fields for course codes and course titles to identify the subject areas taught by a given teacher.[<reflink idref="bib4" id="ref68">4</reflink>] To classify subject-area specialists, we begin by keeping roster records for K–5 teachers and using course codes and course titles to limit our sample to those teaching mathematics, English language arts/reading, science, and social studies. We continued to use the state course codes to count the unique number of academic subject areas (one to four) that these teachers taught each year. We define generalists as those teaching three or four subject areas and specialists as those teaching one or two subject areas. For specialists, we also created indicators for teaching one subject area only, for teaching two subject areas, and a set of indicators for the specific subject-area combinations in which teachers specialized. We consider all these indicators in our descriptive analyses of the patterns in subject-area specialization. In our regression analyses focused on specialization and retention, we consider measures for any specialization and for one- and two-subject specialists.</p> <hd id="AN0174564218-9">Outcome Measures</hd> <p>To estimate the associations between subject-area specialization and teacher retention, we use statewide salary files from the 2011–2012 through 2016–2017 academic years to identify teachers and the schools in which they work.[<reflink idref="bib5" id="ref69">5</reflink>] Our first outcome measure is whether a teacher returns to teach at the same school in the following school year. We code teachers as a "1" for this variable if they will return to the same school in the following year and as a "0" if they will not. Returning to the same school is our primary outcome because it best fits our conceptual framework—that is, the mechanism linking specialization to retention is assignment practices within the school building—and is most pertinent to school leaders who want to retain teachers in their buildings. Our second outcome is whether a teacher returns to teach at any NCPS in the following year.[<reflink idref="bib6" id="ref70">6</reflink>] We code teachers as a "1" for this variable if they will return to NCPS and as a "0" if they will not. Together, our retention measures help us assess whether subject-area specialization is related to teacher retention within schools and/or within NCPS settings.</p> <hd id="AN0174564218-10">Covariates</hd> <p>To better isolate the associations between subject-area specialization and teacher retention, we control for a rich set of teacher, classroom, and school covariates in our analyses. We control for these measures given the descriptive differences between specialists and generalists and prior work showing that retention is related to teacher characteristics and working environments (Borman & Dowling, [<reflink idref="bib7" id="ref71">7</reflink>]). The teacher covariates include time-invariant demographic measures (indicators for female and racial/ethnic identity), time-invariant and time-varying credential measures (experience, graduate degree, NBC, entering teaching through a lateral/alternative route), and time-varying teacher assignment measures.[<reflink idref="bib7" id="ref72">7</reflink>] Specifically, building from work by Ost and Schiman ([<reflink idref="bib38" id="ref73">38</reflink>]), we include indicators for teaching multiple grades during a school year and for teaching in a new grade. Because these multigrade and new-to-grade indicators may be related to assignment to specialization—that is, becoming a math specialist may mean teaching math in multiple grades—we also estimate retention models that exclude these teacher assignment covariates to assess the robustness of findings.</p> <p>We also adjust for classroom and school context. Our classroom controls include the percentage of economically disadvantaged, minority, limited English proficient, and exceptional status students taught by a teacher. At the school level, we control for school size, measures of fiscal resources (per pupil expenditures and teacher salary supplements), measures of school orderliness (short-term suspension and violent act rates), the percentage of economically disadvantaged and minority students, and indicators for school urbanicity. All models also include grade level and year fixed effects.</p> <hd id="AN0174564218-11">Analyses</hd> <p></p> <hd id="AN0174564218-12">What Patterns of Elementary Grades Specialization Exist, and How Are These Patterns Changing...</hd> <p>We examine patterns in elementary grades specialization through a series of descriptive analyses. Specifically, we begin our descriptive analyses by examining trends in the percentage of subject-area specialists in grades K–2 and 3–5 during our study period. In doing so, we assess the prevalence of subject-area specialization in each grade band and track changes in the use of specialization over time. As an extension, we also separately examine (in Fig. A1) trends in subject-area specialization in each grade level (K–5). Given that specialization is concentrated in upper elementary grades, our remaining descriptive analyses focus on grades 3–5. In these analyses, we track the percentage of single- versus two-subject specialists over time and examine trends in the subject areas in which teachers specialize.</p> <hd id="AN0174564218-13">Does Assignment to Specialization Predict an Increased Likelihood of Returning to Teach?</hd> <p>In these analyses, we aim to isolate the associations between subject-area specialization and teacher retention. To do so, our primary analytic approach is a linear probability model that includes teacher, classroom, and school covariates and a teacher fixed effect.[<reflink idref="bib8" id="ref74">8</reflink>] Rather than a comparison across individuals to determine whether specialists are more likely to return than generalist teachers, we adopt a teacher fixed-effect approach. With a teacher fixed effect, the treatment condition is estimated based on a change in a teacher's assignment—from generalist to specialist or specialist to generalist. We estimate how this change predicts teacher retention in the same school or in NCPS. However, we lose some statistical power to detect effects as only teachers switching assignments during the period of our data can contribute to estimated effects.</p> <p>We prefer a teacher fixed-effect model for three reasons. First, by comparing within teachers over time, a teacher fixed effect allows us to adjust for unobserved, time-invariant teacher characteristics that may be related to retention and assignment to specialization. Second, teacher fixed-effect models align with our conceptual framework—that is, that subject-area specialization represents a nonpecuniary incentive that may increase teachers' self-efficacy, job satisfaction, and likelihood to return. We hypothesize that it is the change in a teacher's specialization status that will make them more likely to return to the same school. Finally, to the extent that our model controls are unable to address differences between specialist and generalist teachers, a simple comparison of specialist to generalist may be biased.</p> <p>Despite these strengths, there are potential limitations to a teacher fixed-effect approach. Most notably, in a teacher fixed-effect model, our focal estimates are based on the sample of teachers who switch specialization status during our study period. This may influence the generalizability of our retention estimates if the switchers differ from the larger population of teachers whose assignment does not change over the time period of our data. In addition, estimates from a teacher fixed-effect model may be biased by time-varying teacher characteristics that are associated with specialization and retention. For example, a teacher may participate in mathematics professional development that makes them more interested in specializing in math and more committed to staying in their school. We control for time-varying teacher characteristics to reduce these threats to our focal estimates.</p> <p>To better understand the analytic sample in our teacher fixed-effect models, Table 2 displays counts of unique teachers who are always generalists, who are always specialists, or who switch specialization status during our study period. Table 2 also presents the average number of years these teachers are in the study data. During our study period, the majority of elementary grades teachers (75.69%) are always generalists. Those who are always specialists make up more than 8% of our sample and are in the data for the fewest number of years. We see 6,252 teachers (11.68%) switch from generalist to specialist during our study period. These generalist-to-specialist switchers are in the data for the longest period. Finally, the smallest number of teachers (4.29%) are those switching from specialist to generalist.</p> <p>Table 2. Generalists, Specialists, and Switchers during Our Study Period</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1" /><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Always Generalists</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Always Specialists</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Switch from Generalist to Specialist</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Switch from Specialist to Generalist</th></tr></thead><tbody><tr><td valign="bottom" rowspan="1" colspan="1">Unique teacher count</td><td valign="bottom" rowspan="1" colspan="1">40,530</td><td valign="bottom" rowspan="1" colspan="1">4,472</td><td valign="bottom" rowspan="1" colspan="1">6,252</td><td valign="bottom" rowspan="1" colspan="1">2,296</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Percentage of teachers</td><td char="." valign="bottom" rowspan="1" colspan="1">75.69</td><td char="." valign="bottom" rowspan="1" colspan="1">8.35</td><td char="." valign="bottom" rowspan="1" colspan="1">11.68</td><td char="." valign="bottom" rowspan="1" colspan="1">4.29</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Avg. number of years in data</td><td char="." valign="bottom" rowspan="1" colspan="1">3.08</td><td char="." valign="bottom" rowspan="1" colspan="1">2.17</td><td char="." valign="bottom" rowspan="1" colspan="1">4.17</td><td char="." valign="bottom" rowspan="1" colspan="1">3.85</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Female (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">95.58</td><td char="." valign="bottom" rowspan="1" colspan="1">91.71</td><td char="." valign="bottom" rowspan="1" colspan="1">93.14</td><td char="." valign="bottom" rowspan="1" colspan="1">93.57</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Minority (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">15.99</td><td char="." valign="bottom" rowspan="1" colspan="1">16.49</td><td char="." valign="bottom" rowspan="1" colspan="1">14.75</td><td char="." valign="bottom" rowspan="1" colspan="1">14.50</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Age (years)</td><td char="." valign="bottom" rowspan="1" colspan="1">40.03</td><td char="." valign="bottom" rowspan="1" colspan="1">42.21</td><td char="." valign="bottom" rowspan="1" colspan="1">40.04</td><td char="." valign="bottom" rowspan="1" colspan="1">40.09</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Years of experience</td><td char="." valign="bottom" rowspan="1" colspan="1">11.09</td><td char="." valign="bottom" rowspan="1" colspan="1">12.50</td><td char="." valign="bottom" rowspan="1" colspan="1">11.35</td><td char="." valign="bottom" rowspan="1" colspan="1">10.95</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Graduate degree (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">32.04</td><td char="." valign="bottom" rowspan="1" colspan="1">36.74</td><td char="." valign="bottom" rowspan="1" colspan="1">34.18</td><td char="." valign="bottom" rowspan="1" colspan="1">36.46</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">NBC (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">11.64</td><td char="." valign="bottom" rowspan="1" colspan="1">12.03</td><td char="." valign="bottom" rowspan="1" colspan="1">13.73</td><td char="." valign="bottom" rowspan="1" colspan="1">13.86</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Alternative entry (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">6.32</td><td char="." valign="bottom" rowspan="1" colspan="1">7.66</td><td char="." valign="bottom" rowspan="1" colspan="1">5.39</td><td char="." valign="bottom" rowspan="1" colspan="1">5.39</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Multigrade teacher (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">8.02</td><td char="." valign="bottom" rowspan="1" colspan="1">21.34</td><td char="." valign="bottom" rowspan="1" colspan="1">7.47</td><td char="." valign="bottom" rowspan="1" colspan="1">8.77</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">New to grade (%)</td><td char="." valign="bottom" rowspan="1" colspan="1">15.43</td><td char="." valign="bottom" rowspan="1" colspan="1">13.34</td><td char="." valign="bottom" rowspan="1" colspan="1">20.55</td><td char="." valign="bottom" rowspan="1" colspan="1">20.59</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>2 Note. This table displays descriptive data on elementary grades teachers who are always generalists, always specialists, switch from generalist to specialist, or switch from specialist to generalist during our study period. Data include unique teacher counts, percentages of the teacher sample, the average number of years teachers were in our analytical sample, and teacher characteristics. NBC = National Board Certification.</p> <p>As a complement to our preferred teacher fixed-effect analyses, we run models in which we substitute a school fixed effect for the teacher fixed effect. With this approach, the treatment condition is working as a specialist and we estimate whether specialists are more likely to return—to the same school or to NCPS—than generalists working in the same schools. This approach does not place the same demands on the data as a teacher fixed-effect model. Furthermore, a school fixed effect helps adjust for the possibility that teachers switch schools to become specialists by allowing comparisons across teachers compared with within teachers. Despite these benefits, we still prefer teacher fixed-effect analyses because they adjust for unobserved teacher characteristics and better test our conceptual framework linking specialization to retention.</p> <p> <ephtml> <math display="block" overflow="scroll" xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mtable displaystyle="true"><mlabeledtr><mtd><mtext>(1)</mtext></mtd><mtd><mrow><msub><mrow><mtext>WillReturn</mtext></mrow><mrow><mi mathvariant="italic">ijst</mi></mrow></msub><mo>=</mo><msub><mi>β</mi><mn>0</mn></msub><mo>+</mo><mi>β</mi><msub><mrow><mtext>Specialist</mtext></mrow><mrow><mi mathvariant="italic">it</mi></mrow></msub><mo>+</mo><mi>α</mi><msub><mrow><mtext>Teacher</mtext></mrow><mrow><mi mathvariant="italic">it</mi></mrow></msub><mo>+</mo><mi>γ</mi><msub><mrow><mtext>Class</mtext></mrow><mrow><mi mathvariant="italic">it</mi></mrow></msub><mo>+</mo><mi>δ</mi><msub><mrow><mtext>School</mtext></mrow><mrow><mi mathvariant="italic">ijt</mi></mrow></msub><mo>+</mo><msub><mi>μ</mi><mi>i</mi></msub><mo>+</mo><msub><mi>ε</mi><mrow><mi mathvariant="italic">ij</mi><mi>s</mi><mi>t</mi></mrow></msub></mrow></mtd></mlabeledtr></mtable></mrow></math> </ephtml> </p> <p>Equation (<reflink idref="bib1" id="ref75">1</reflink>) presents our preferred linear probability model with standard errors clustered at the teacher level. We cluster standard errors at the teacher level because that is the level of the treatment and the level at which we seek to make inferences. WillReturn<subs><emph>ijst</emph></subs> is an indicator for whether teacher <emph>i</emph> in classroom <emph>j</emph> in school <emph>s</emph> at time <emph>t</emph> will teach at the same school in the following year. In additional analyses, we assess whether teachers will return to NCPS in the following year. Specialist<subs><emph>it</emph></subs> is a set of time-varying indicators for whether a teacher is a subject-area specialist in a given year. We define teachers as specialists if their classroom assignments are limited to one or two subjects according to administrative data. A second analysis divides the treatment indicator into two parts (one-subject specialization or two-subject specialization) to determine whether estimates differ based on the degree of specialization. Teacher<subs><emph>it</emph></subs>, Class<subs><emph>it</emph></subs>, and School<subs><emph>ijt</emph></subs> are vectors of covariates that adjust for differences across teachers and working environments, and μ<subs><emph>i</emph></subs> is a teacher fixed effect. In specification checks, we replace the teacher fixed effect with a school fixed effect. Our primary analyses include teachers in grades K–5; we also estimate separate models for teachers in grades K–2 and for teachers in grades 3–5.</p> <hd id="AN0174564218-14">Do These Teacher Retention Results Vary by School or Teacher Characteristics?</hd> <p>Results from our teacher fixed-effect models will indicate whether teachers are more or less likely to return to the same school after becoming a subject-area specialist than when they were as a generalist. However, it is also important to understand whether subject-area specialization predicts retention in certain school settings or for certain teachers. This is particularly true because specialization is concentrated in rural environments (Table 1) and because teacher attrition is a larger problem for schools educating higher proportions of low-income, racial/ethnic minority, and low-performing students. As such, we estimate separate retention analyses for K–5 teachers working in certain school environments and based on observable teacher characteristics. These analyses control for the same vectors of teacher, classroom, and school covariates as in Equation (<reflink idref="bib1" id="ref76">1</reflink>) and include either a teacher fixed effect (preferred specification) or a school fixed effect.</p> <p>We estimate separate retention models for teachers in urban/suburban schools, rural/town schools, high-poverty schools, non-high-poverty schools, high-minority schools, non-high-minority schools, low-performing schools, and non-low-performing schools. For example, we estimate whether K–5 teachers in urban/suburban settings are more likely to return to the same urban/suburban school after becoming a subject-area specialist than when they were a generalist. Following definitions set by the National Center for Education Statistics, we classify high-poverty schools as K–12 campuses where more than 75% of students qualify for subsidized school meals (National Center for Education Statistics, [<reflink idref="bib35" id="ref77">35</reflink>]). Likewise, we identify high-minority schools as K–12 campuses where more than 75% of students are a racial/ethnic minority. We define low-performing schools as elementary schools where pass rates on state assessments are below the 25th percentile (bottom quartile).</p> <p>We estimate separate retention models for White teachers, Black teachers, early career teachers, veteran teachers, highly effective teachers, and less effective teachers.[<reflink idref="bib9" id="ref78">9</reflink>] Analyses for Black teachers are particularly relevant because the student population in NCPS is majority-minority, the teacher workforce in North Carolina is predominantly White, and recent research highlights the impacts of Black teachers on same-race students (Egalite & Kisida, [<reflink idref="bib11" id="ref79">11</reflink>]; Egalite et al., [<reflink idref="bib12" id="ref80">12</reflink>]; Gershenson et al., [<reflink idref="bib16" id="ref81">16</reflink>]). Following prior studies on teacher development (Henry et al., [<reflink idref="bib22" id="ref82">22</reflink>]), we identify early career teachers as those with less than 5 years of experience and veteran teachers as those with 5 or more years of experience.[<reflink idref="bib10" id="ref83">10</reflink>] Finally, we want to assess whether specialization encourages highly effective teachers to return to the same school. To operationalize teacher effectiveness, we create an indicator for whether teachers earn evaluation ratings of "distinguished"—the highest rating in North Carolina.[<reflink idref="bib11" id="ref84">11</reflink>] We create a measure of teacher effectiveness with evaluation ratings, rather than value-added estimates, because evaluation ratings are available for teachers in all grade levels included in our analysis. Value-added estimates are available in grades 4–5 only. Our indicator for teachers earning a distinguished rating is time-invariant and comes from the first year in which we observe teachers during our analysis period. By focusing on teachers' ratings from their initial appearance in the data, we aim to ensure that any changes in within-teacher retention are due to specialization and not because teachers earned a distinguished rating.</p> <p>These school and teacher characteristic models will indicate whether specialization significantly predicts retention for a given subgroup—that is, whether the specialization estimate significantly differs from zero. Postestimation, we also test whether specialization estimates for a given subgroup (e.g., teachers in high-poverty schools) significantly differ from estimates for the paired subgroup (e.g., teachers in non-high-poverty schools).</p> <hd id="AN0174564218-15">Results</hd> <p></p> <hd id="AN0174564218-16">What Patterns of Elementary Grades Specialization Exist, and How Are These Patterns Changing...</hd> <p>Figure 1A displays trends in the percentage of subject-area specialists in grades K–2 and in grades 3–5 during the 2011–2012 through 2015–2016 school years. Overall, we find that subject-area specialization is rare in early elementary grades. Although percentages fluctuate slightly, approximately 3%–5% of K–2 teachers are specialists. The remaining 95%–97% of K–2 teachers are generalists. Conversely, specialization is more common in upper elementary grades and has become a more widely used assignment strategy over time. In 2012, approximately 20% of upper elementary grades teachers specialized; by 2016, 30% of upper elementary grades teachers specialized. These patterns are consistent with the value of self-contained classrooms in early grades and a desire to better prepare students for having multiple teachers as middle school approaches. Likewise, these patterns are consistent with work from rural districts in the southeast United States, showing that early elementary grades teachers are more likely than upper elementary grades teachers to find relative enjoyment in teaching all subjects (Wilkins, [<reflink idref="bib49" id="ref85">49</reflink>]). Data in Figure A1 show that specialization is particularly concentrated in fourth and fifth grades. In 2016, 32% of fourth-grade teachers and 46% of fifth-grade teachers were subject-area specialists.</p> <p>Graph: Figure 1. Panel A displays the percentage of teachers in grades K–2 and in grades 3–5 who are subject-area specialists during the 2011–2012 through 2015–2016 school years. Panel B displays the percentage of upper elementary grades teachers (3–5) who are single- and two-subject specialists during the 2011–2012 through 2015–2016 school years.</p> <p>To better understand the increase in upper elementary grades specialization, Figure 1B presents trends in the percentage of single-subject versus two-subject specialists in grades 3–5. These data indicate that the percentage of single-subject specialists has fluctuated but remained around 8%–10% during the study period. As such, the majority of growth in subject-area specialization has come from two-subject specialists. In 2012, two-subject specialists made up 11% of all teachers in grades 3–5. By 2016, two-subject specialists comprised 21% of upper elementary grades teachers.</p> <p>Finally, to examine these assignments at a more granular level, Figure 2 presents the percentage of upper elementary grades teachers specializing in certain subject areas. For ease of viewing, we display these data for select subject-area categories; Table A1 presents the percentages for all subject-area categories. Figure 2 shows substantial increases in the percentage of two-subject specialists teaching (<reflink idref="bib1" id="ref86">1</reflink>) reading and social studies and (<reflink idref="bib2" id="ref87">2</reflink>) math and science. For example, the percentage of grade 3–5 teachers specializing in reading and social studies increased from 3.8% to 8.8% over our study period. Similarly, the percentage of grade 3–5 teachers specializing in math and science increased from 2.9% to 7.1%. There were also increases in the percentage of upper elementary grades teachers teaching math only and in those specializing in science and social studies.</p> <p>Graph: Figure 2. This figure displays the percentage of upper elementary grades teachers (3–5) specializing in select subject-area combinations over the 2011–2012 through 2015–2016 school years.</p> <hd id="AN0174564218-17">Does Assignment to Specialization Predict an Increased Likelihood of Returning to Teach?</hd> <p>Table 3 presents results from linear probability models assessing whether subject-area specialists are more likely to return to the same school in the following year. These models include our preferred teacher fixed-effect analyses, which compare within teachers over time, and school fixed-effect analyses, which compare generalists and specialists working in the same school. To contextualize these results, we note that during our study period approximately 82% of K–5 teachers returned to the same school in the following year. Across grades K–5, we find positive and significant retention results in our teacher fixed-effect models. Teachers are 1.2 percentage points more likely to return to their school after becoming a subject-area specialist than when they were generalists. These teacher fixed-effect results are comparable in magnitude when we separately examine one- and two-subject specialists. Estimates from our K–5 school fixed-effect models are smaller in magnitude and statistically insignificant.[<reflink idref="bib12" id="ref88">12</reflink>]</p> <p>Table 3. Does Subject-Area Specialization Predict Returning to the Same School?</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–5</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–2</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades 3–5</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Any specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.012*</td><td char="." valign="bottom" rowspan="1" colspan="1">.003</td><td char="." valign="bottom" rowspan="1" colspan="1">.009</td><td char="." valign="bottom" rowspan="1" colspan="1">.013</td><td char="." valign="bottom" rowspan="1" colspan="1">.014<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.005</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.006)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.003)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.014)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.008)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td></tr><tr><td rowspan="2" valign="top" colspan="1">One-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.013<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.003</td><td char="." valign="bottom" rowspan="1" colspan="1">.009</td><td char="." valign="bottom" rowspan="1" colspan="1">.016</td><td char="." valign="bottom" rowspan="1" colspan="1">.014</td><td char="." valign="bottom" rowspan="1" colspan="1">.000</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.019)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.010)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.010)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.006)</td></tr><tr><td rowspan="2" valign="top" colspan="1">Two-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.012<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.003</td><td char="." valign="bottom" rowspan="1" colspan="1">.008</td><td char="." valign="bottom" rowspan="1" colspan="1">.007</td><td char="." valign="bottom" rowspan="1" colspan="1">.013<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.008<sup>+</sup></td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.019)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.013)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Observation count</td><td valign="bottom" rowspan="1" colspan="1">161,356</td><td valign="bottom" rowspan="1" colspan="1">161,356</td><td valign="bottom" rowspan="1" colspan="1">85,885</td><td valign="bottom" rowspan="1" colspan="1">85,885</td><td valign="bottom" rowspan="1" colspan="1">80,416</td><td valign="bottom" rowspan="1" colspan="1">80,416</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>3 Note. This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models assessing whether subject-area specialization predicts staying in the same school. All analyses control for a set of teacher, classroom, and school covariates. <sups>+</sups> and * indicate statistical significance at the.10 and.05 levels, respectively.</p> <p>Turning to our separate analyses by grade band, we find that specialization is not significantly associated with retention in early elementary grades (K–2). Specialists are no more likely to return to the same school than when they were generalists and no more likely to return than their peers who are generalists. In upper elementary grades (3–5), we find that specialists are 1.4 percentage points more likely to return to the same school than when they were generalists. Coefficients for one- and two-subject specialists are similar in magnitude. As shown in Table A3, results for returning to the same school are comparable (with those in Table 3) when we omit the multigrade and new-to-grade covariates from analyses. This suggests that the associations between specialization and retention are not influenced by other assignment practices that may be related to specialization.</p> <p>Because specialization is particularly concentrated in fourth and fifth grades (see Fig. A1), we estimated retention models for those grade levels only. Results in Table A4 show that specialists in grades 4–5 are 2.1 percentage points more likely to return to the same school than when they were generalists. These results are consistent for any specialization and for one- or two-subject specialization. Furthermore, two-subject specialists in grades 4–5 are 1.2 percentage points more likely to return to the same school than their generalist peers (school fixed effect). Overall, these results suggest that the impact of specialization on teacher retention may be strongest at the grade levels in which specialization is most concentrated.</p> <p>Finally, Table A5 displays estimates from teacher and school fixed-effect models assessing whether specialists are more likely to return to NCPS. These results are largely insignificant, suggesting that exit rates from the state's public school teaching workforce are approximately the same whether teachers have specialist or generalist teaching assignments.</p> <hd id="AN0174564218-18">Do Teacher Retention Results Vary by School or Teacher Characteristics?</hd> <p>We view subject-area specialization as a nonpecuniary job benefit that encourages teachers to return to the same school. However, specialization's ability to induce teacher retention may depend on the context of the school or the characteristics of the teacher. As such, we assess whether specialization is a differentially effective retention tool for certain types of schools and teachers.</p> <p>Table 4 presents specialization results for select school characteristics. These estimates come from separate models where the sample only includes teachers working in schools with the specified characteristic. We find positive and significant specialization results in urban/suburban schools. Teachers are 2.7 percentage points more likely to return to their urban/suburban school after becoming a specialist than when they were generalists. Furthermore, the specialization estimate for urban/suburban schools significantly differs from the estimate for rural/town schools in both the teacher and school fixed-effect analyses. These findings are noteworthy because specialization is more heavily concentrated in rural environments (see Table 1). The remaining estimates in Table 4 suggest that specialization may encourage retention in non-high-need schools. Teachers are 1.5 percentage points more likely to return to their non-high-poverty school after becoming a subject-area specialist and are 1.3 percentage points more likely to return to their non-low-performing school after becoming a specialist. Although these estimates differ from zero, postestimation tests indicate that the estimates for high-poverty versus non-high-poverty schools and low-performing versus non-low-performing schools do not differ from each other at statistically significant levels.</p> <p>Table 4. Do Subject-Area Specialization Results Differ by School Characteristics?</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–5</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Specialization: urban/suburb</td><td char="." valign="top" rowspan="1" colspan="1">.027**</td><td char="." valign="top" rowspan="1" colspan="1">.011<sup>+</sup></td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.009)</td><td char="." valign="top" rowspan="1" colspan="1">(.005)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">75,589</td><td valign="top" rowspan="1" colspan="1">75,589</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: rural/town</td><td char="." valign="top" rowspan="1" colspan="1">.005</td><td char="." valign="top" rowspan="1" colspan="1">−.002</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.007)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">85,767</td><td valign="top" rowspan="1" colspan="1">85,767</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: high-poverty</td><td char="." valign="top" rowspan="1" colspan="1">.003</td><td char="." valign="top" rowspan="1" colspan="1">−.008</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.013)</td><td char="." valign="top" rowspan="1" colspan="1">(.007)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">48,695</td><td valign="top" rowspan="1" colspan="1">48,695</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: non–high poverty</td><td char="." valign="top" rowspan="1" colspan="1">.015*</td><td char="." valign="top" rowspan="1" colspan="1">.006</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.007)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">112,661</td><td valign="top" rowspan="1" colspan="1">112,661</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: high-minority</td><td char="." valign="top" rowspan="1" colspan="1">.018</td><td char="." valign="top" rowspan="1" colspan="1">.004</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.015)</td><td char="." valign="top" rowspan="1" colspan="1">(.008)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">37,965</td><td valign="top" rowspan="1" colspan="1">37,965</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: non–high minority</td><td char="." valign="top" rowspan="1" colspan="1">.010</td><td char="." valign="top" rowspan="1" colspan="1">.003</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.006)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">123,391</td><td valign="top" rowspan="1" colspan="1">123,391</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: low performing</td><td char="." valign="top" rowspan="1" colspan="1">.004</td><td char="." valign="top" rowspan="1" colspan="1">.001</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.022)</td><td char="." valign="top" rowspan="1" colspan="1">(.008)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">37,305</td><td valign="top" rowspan="1" colspan="1">37,305</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: non–low performing</td><td char="." valign="top" rowspan="1" colspan="1">.013*</td><td char="." valign="top" rowspan="1" colspan="1">.003</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.007)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">120,816</td><td valign="top" rowspan="1" colspan="1">120,816</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>4 Note. This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models estimating the associations between subject-area specialization and teacher retention in the same school. All analyses control for a set of teacher, classroom, and school covariates. <sups>+</sups>, *, and ** indicate statistical significance at the.10,.05, and.01 levels, respectively. Cells shaded in gray indicate that the coefficient significantly differs from the paired subgroup coefficient.</p> <p>Table 5 presents specialization results for select teacher characteristics. Once again, these estimates come from separate models where the sample only includes teachers with the specified characteristic. We find that specialization strongly encourages the retention of Black elementary grades teachers. Black teachers are 5.3 percentage points more likely to return to their school as a specialist than when they were generalists. This estimate is significantly different from zero and statistically significantly different from the specialization estimate for White teachers (which does not differ from zero). These retention results are particularly important given the benefits of a diverse teacher workforce and recent work showing that Black teachers in North Carolina have lower retention rates (Sun, [<reflink idref="bib48" id="ref89">48</reflink>]). The remaining estimates in Table 5 indicate that veteran teachers and teachers rated below distinguished are more likely to return to their school after becoming a subject-area specialist. Specialization does not predict that the highest-rated teachers return to their schools. The estimates for veteran teachers and those rated below distinguished do not significantly differ from the estimates for the paired subgroups—that is, early career teachers and teachers rated as distinguished.</p> <p>Table 5. Do Subject-Area Specialization Results Differ by Teacher Characteristics?</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–5</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Specialization: Black teachers</td><td char="." valign="top" rowspan="1" colspan="1">.053**</td><td char="." valign="top" rowspan="1" colspan="1">.012</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.019)</td><td char="." valign="top" rowspan="1" colspan="1">(.012)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">17,750</td><td valign="top" rowspan="1" colspan="1">17,750</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: White teachers</td><td char="." valign="top" rowspan="1" colspan="1">.007</td><td char="." valign="top" rowspan="1" colspan="1">.002</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.006)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">136,160</td><td valign="top" rowspan="1" colspan="1">136,160</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: early career teachers</td><td char="." valign="top" rowspan="1" colspan="1">.017</td><td char="." valign="top" rowspan="1" colspan="1">.000</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.016)</td><td char="." valign="top" rowspan="1" colspan="1">(.008)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">43,378</td><td valign="top" rowspan="1" colspan="1">43,378</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: veteran teachers</td><td char="." valign="top" rowspan="1" colspan="1">.013*</td><td char="." valign="top" rowspan="1" colspan="1">.004</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.006)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">117,978</td><td valign="top" rowspan="1" colspan="1">117,978</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: teachers rated as distinguished</td><td char="." valign="top" rowspan="1" colspan="1">.001</td><td char="." valign="top" rowspan="1" colspan="1">.002</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.014)</td><td char="." valign="top" rowspan="1" colspan="1">(.009)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">18,878</td><td valign="top" rowspan="1" colspan="1">18,878</td></tr><tr><td rowspan="2" valign="top" colspan="1">Specialization: teachers rated below distinguished</td><td char="." valign="top" rowspan="1" colspan="1">.012<sup>+</sup></td><td char="." valign="top" rowspan="1" colspan="1">.001</td></tr><tr><td char="." valign="top" rowspan="1" colspan="1">(.006)</td><td char="." valign="top" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="top" rowspan="1" colspan="1">Observation count</td><td valign="top" rowspan="1" colspan="1">133,992</td><td valign="top" rowspan="1" colspan="1">133,992</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>5 Note. This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models estimating the associations between subject-area specialization and teacher retention in the same school. All analyses control for a set of teacher, classroom, and school covariates. <sups>+</sups>, *, and ** indicate statistical significance at the.10,.05, and.01 levels, respectively. Cells shaded in gray indicate that the coefficient significantly differs from the paired subgroup coefficient.</p> <hd id="AN0174564218-19">Discussion</hd> <p>School leaders need effective and affordable approaches to retain their teacher workforce. With this motivation, we investigated a promising, low-cost option for school leaders to encourage teacher retention: subject-area specialization in elementary grades. Specifically, we tracked the incidence of subject-area specialization in North Carolina elementary schools, assessed whether teaching in a specialist role promotes retention, and examined whether subject-area specialization is an effective retention strategy for certain schools and teachers. Our analyses make three contributions to the literature on the strategic assignment of teachers.</p> <p>First, we show that subject-area specialization is already a common practice in upper elementary grades in North Carolina. This bolsters the practical significance of our teacher retention results. During our study period, the percentage of upper elementary grades teachers specializing increased by 50%—from 20% of grades 3–5 teachers in 2012 to 30% in 2016. Specialization is even more highly concentrated in fourth and fifth grades, where 38% of teachers were subject-area specialists in 2016. Much of the growth in this strategic assignment practice has come through two-subject specialization, especially with upper elementary grades teachers assigned to teach (<reflink idref="bib1" id="ref90">1</reflink>) reading and social studies and (<reflink idref="bib2" id="ref91">2</reflink>) mathematics and science. These data suggest that school leaders are pairing upper elementary grades teachers together to cover all four academic subjects.</p> <p>Second, in our preferred teacher fixed-effect models, we find that elementary grades teachers are more likely to return to the same school after becoming a subject-area specialist. These retention estimates are consistent for one- and two-subject specialists and are more robust in models limited to grades 4–5, where specialization is more highly concentrated. Furthermore, results suggest that specialization is unrelated to retention in public schools statewide. This supports our hypothesis that specialization influences teacher satisfaction and self-efficacy at a particular school site. By estimating models with a teacher fixed effect and a rich set of time-varying teacher, classroom, and school covariates, we contend that we have isolated the relationship between specialization and retention. Our specialization results fit with prior research showing that teacher retention is related to teaching assignments (Donaldson & Johnson, [<reflink idref="bib10" id="ref92">10</reflink>]; Ost & Schiman, [<reflink idref="bib38" id="ref93">38</reflink>]).</p> <p>Finally, we find that specialization promotes retention for some key constituencies but not for others. Considering school characteristics, evidence suggests that specialization is an effective retention tool in urban/suburban environments. It is possible that differences in results between urban and rural schools are due to intentionality—that is, urban schools may do a better job of using specialization in ways that enhance teachers' job satisfaction and self-efficacy. The benefits of specialization may also be more salient to teachers in urban/suburban environments, where there may be more employment choices within a reasonable commuting distance. Estimates show that elementary grades teachers are more likely to return to non-high-poverty and non-low-performing schools after becoming a specialist. Specialization results are not statistically significant for high-poverty and low-performing schools. Although this suggests that specialization may not be a powerful enough inducement to keep teachers in high-need schools, it is important to note that the estimates for (<reflink idref="bib1" id="ref94">1</reflink>) high-poverty versus non-high-poverty and (<reflink idref="bib2" id="ref95">2</reflink>) low-performing versus non-low-performing schools did not significantly differ from each other. Considering teacher characteristics, we find strong evidence that Black teachers are more likely to return to the same school after becoming a subject-area specialist. Furthermore, the relationship between specialization and retention is significantly stronger for Black teachers than White teachers. These results are especially salient given prior research showing the benefits of diverse teachers to same-race students and suggest that teacher assignments may be an effective strategy to retain more Black teachers (Egalite & Kisida, [<reflink idref="bib11" id="ref96">11</reflink>]; Egalite et al., [<reflink idref="bib12" id="ref97">12</reflink>]; Gershenson et al., [<reflink idref="bib16" id="ref98">16</reflink>]).</p> <p>Before considering the implications of our findings, it is useful to review the potential limitations of our work—generalizability, methodological approach, assumptions, and intentionality. Regarding generalizability, our data come from a large and diverse state, but our results may not be representative of other locations. Continued research should replicate analyses on subject-area specialization in other states and districts. Regarding methodology, we contend that, outside of a random assignment study, teacher fixed-effect models are the best approach to isolate the relationships between specialization and retention. These models let us test our hypothesis and adjust for time-invariant teacher characteristics that may be related to specialization and retention. Nonetheless, the teacher fixed-effect approach relies on switchers to estimate differences, meaning those who are always generalists or specialists do not contribute to retention estimates. To the extent that switchers differ from other teachers, our estimates may not generalize. Furthermore, it is possible that unmeasured, time-varying teacher characteristics may bias our estimates. Regarding assumptions, we do not have data on and cannot test whether teachers perceived specialist roles to be less burdensome or whether specialist roles increased teacher self-efficacy and job satisfaction. This may be an area for additional analyses to better understand how teachers perceive specialization, its impact on their teaching, and its impacts on these outcomes. Last, regarding intentionality, we do not know the motivations for why schools assign teachers to specialist roles—that is, was it a strategic and considered approach that came with teacher buy-in, was it a reactionary measure intended to address staffing concerns, or was the assignment made at the request and urging of a teacher who desired such a role. Classroom assignments are likely negotiated decisions between school leaders and teachers, and the intentionality behind specialization may matter to its efficacy.</p> <p>Moving forward, a key question is what school leaders should think about specialization and its ability to improve elementary school outcomes. One branch of research shows that teachers are less effective after becoming subject-area specialists (Bastian & Fortner, [<reflink idref="bib4" id="ref99">4</reflink>]; Fryer, [<reflink idref="bib14" id="ref100">14</reflink>]; Hwang & Kisida, [<reflink idref="bib25" id="ref101">25</reflink>]). However, it is important to note that school-level achievement (Bastian & Fortner, [<reflink idref="bib4" id="ref102">4</reflink>]; Hwang & Kisida, [<reflink idref="bib25" id="ref103">25</reflink>]) and school-level attendance and disciplinary incidents (Hwang & Kisida, [<reflink idref="bib25" id="ref104">25</reflink>]) are not influenced by the greater use of specialization. In North Carolina, this may be because school leaders assign more effective teachers, on average, to be subject-area specialists (Bastian & Fortner, [<reflink idref="bib4" id="ref105">4</reflink>]). The present study shows that elementary grades teachers are more likely to return to the same school after becoming a specialist. This retention can benefit school culture, school achievement, and district-level finances. Given that approximately 18% of elementary grades teachers leave their schools during our study period, a 1.5 percentage point reduction in attrition (from our models for grades 3–5 teachers) represents a decrease in school turnover of about 8%. This percentage is even larger for Black teachers, for whom a 5 percentage point reduction in attrition represents a decrease in turnover of about 25%. In financial terms, elementary grades specialization could save districts thousands of dollars each year. For instance, the average district in North Carolina employs approximately 300 elementary grades teachers. Assuming (<reflink idref="bib1" id="ref106">1</reflink>) specialization reduced school-level attrition by 1.5 percentage points and (<reflink idref="bib2" id="ref107">2</reflink>) attrition cost $7,000 per teacher, the average district would save between $30,000 and $40,000 annually.[<reflink idref="bib13" id="ref108">13</reflink>] Across North Carolina's 115 school districts, this is approximately $4 million in savings each year.</p> <p>On balance, scholarship from North Carolina suggests that specialization may be a net positive for elementary schools—promoting retention, especially in urban environments and for Black teachers, without harming school-level student achievement. It is worth noting, however, that teachers become less effective (as measured by value added to student achievement) after switching from generalist to specialist. Continued research is needed to more fully assess specialization and its mechanisms, to confirm whether it is an effective strategy for reducing teacher turnover, and to further consider the range of student outcomes—for example, achievement, social-emotional well-being, relationships with teachers—that may be influenced by subject-area specialization.</p> <hd id="AN0174564218-20">Appendix.</hd> <p>Graph: Figure A1. This figure displays the percentage of teachers in each elementary grade level (K–5) who are subject-area specialists during the 2011–2012 through 2015–2016 school years.</p> <p>Table A1. The Percentage of Teachers Specializing in Subject-Area Combinations (Grades 3–5)</p> <p> <ephtml> <table><thead><tr><th align="left" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1">Year</th><th align="center" style="border-bottom: solid thin black" colspan="4" valign="bottom" scope="colgroup" rowspan="1">Single-Subject Specialists</th><th align="center" style="border-bottom: solid thin black" colspan="6" valign="bottom" scope="colgroup" rowspan="1">Two-Subject Specialists</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Reading</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Math</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Science</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Social Studies</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Reading and Math</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Reading and Science</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Reading and Social Studies</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Math and Science</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Math and Social Studies</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Science and Social Studies</th></tr></thead><tbody><tr><td valign="bottom" rowspan="1" colspan="1">2012</td><td char="." valign="bottom" rowspan="1" colspan="1">4.95</td><td char="." valign="bottom" rowspan="1" colspan="1">2.48</td><td char="." valign="bottom" rowspan="1" colspan="1">.69</td><td char="." valign="bottom" rowspan="1" colspan="1">.47</td><td char="." valign="bottom" rowspan="1" colspan="1">1.69</td><td char="." valign="bottom" rowspan="1" colspan="1">.49</td><td char="." valign="bottom" rowspan="1" colspan="1">3.79</td><td char="." valign="bottom" rowspan="1" colspan="1">2.86</td><td char="." valign="bottom" rowspan="1" colspan="1">.93</td><td char="." valign="bottom" rowspan="1" colspan="1">1.20</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">2013</td><td char="." valign="bottom" rowspan="1" colspan="1">4.34</td><td char="." valign="bottom" rowspan="1" colspan="1">2.45</td><td char="." valign="bottom" rowspan="1" colspan="1">.72</td><td char="." valign="bottom" rowspan="1" colspan="1">.38</td><td char="." valign="bottom" rowspan="1" colspan="1">1.54</td><td char="." valign="bottom" rowspan="1" colspan="1">.55</td><td char="." valign="bottom" rowspan="1" colspan="1">4.83</td><td char="." valign="bottom" rowspan="1" colspan="1">4.22</td><td char="." valign="bottom" rowspan="1" colspan="1">.64</td><td char="." valign="bottom" rowspan="1" colspan="1">1.41</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">2014</td><td char="." valign="bottom" rowspan="1" colspan="1">4.57</td><td char="." valign="bottom" rowspan="1" colspan="1">3.85</td><td char="." valign="bottom" rowspan="1" colspan="1">1.68</td><td char="." valign="bottom" rowspan="1" colspan="1">1.20</td><td char="." valign="bottom" rowspan="1" colspan="1">1.36</td><td char="." valign="bottom" rowspan="1" colspan="1">.62</td><td char="." valign="bottom" rowspan="1" colspan="1">6.62</td><td char="." valign="bottom" rowspan="1" colspan="1">5.21</td><td char="." valign="bottom" rowspan="1" colspan="1">.84</td><td char="." valign="bottom" rowspan="1" colspan="1">1.77</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">2015</td><td char="." valign="bottom" rowspan="1" colspan="1">4.07</td><td char="." valign="bottom" rowspan="1" colspan="1">3.06</td><td char="." valign="bottom" rowspan="1" colspan="1">1.10</td><td char="." valign="bottom" rowspan="1" colspan="1">.27</td><td char="." valign="bottom" rowspan="1" colspan="1">1.67</td><td char="." valign="bottom" rowspan="1" colspan="1">.60</td><td char="." valign="bottom" rowspan="1" colspan="1">7.84</td><td char="." valign="bottom" rowspan="1" colspan="1">6.02</td><td char="." valign="bottom" rowspan="1" colspan="1">.82</td><td char="." valign="bottom" rowspan="1" colspan="1">1.94</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">2016</td><td char="." valign="bottom" rowspan="1" colspan="1">4.43</td><td char="." valign="bottom" rowspan="1" colspan="1">3.32</td><td char="." valign="bottom" rowspan="1" colspan="1">1.22</td><td char="." valign="bottom" rowspan="1" colspan="1">.30</td><td char="." valign="bottom" rowspan="1" colspan="1">1.47</td><td char="." valign="bottom" rowspan="1" colspan="1">.67</td><td char="." valign="bottom" rowspan="1" colspan="1">8.80</td><td char="." valign="bottom" rowspan="1" colspan="1">7.06</td><td char="." valign="bottom" rowspan="1" colspan="1">.86</td><td char="." valign="bottom" rowspan="1" colspan="1">2.09</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>Graph</p> <p>6 For the 2011–2012 through 2015–2016 school years, this table displays the percentage of teachers in grades 3–5 who specialized in certain subject-area combinations.</p> <p>Table A2. Are Generalists More Likely to Return to the Same School and North Carolina?</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Return to the Same School</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Return to North Carolina</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Generalists</td><td char="." valign="bottom" rowspan="1" colspan="1">−.012*</td><td char="." valign="bottom" rowspan="1" colspan="1">−.003</td><td char="." valign="bottom" rowspan="1" colspan="1">−.001</td><td char="." valign="bottom" rowspan="1" colspan="1">.005*</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.006)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.003)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.002)</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Observation count</td><td valign="bottom" rowspan="1" colspan="1">161,356</td><td valign="bottom" rowspan="1" colspan="1">161,356</td><td valign="bottom" rowspan="1" colspan="1">161,571</td><td valign="bottom" rowspan="1" colspan="1">161,571</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>7 This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models assessing whether generalist teachers are more likely to return to the same school and to any North Carolina school. All analyses control for a set of teacher, classroom, and school covariates. * indicates statistical significance at the.05 level.</p> <p>Table A3. Does Subject-Area Specialization Predict Returning to the Same School? (Excluding Multigrade and New-to-Grade Covariates)</p> <p> <ephtml> <table><thead><tr><th align="center" valign="bottom" scope="col" rowspan="1" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–5</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–2</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades 3–5</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1" /><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Any specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.010<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">−.000</td><td char="." valign="bottom" rowspan="1" colspan="1">.006</td><td char="." valign="bottom" rowspan="1" colspan="1">.002</td><td char="." valign="bottom" rowspan="1" colspan="1">.013<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.004</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.003)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.014)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.008)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td></tr><tr><td rowspan="2" valign="top" colspan="1">One-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.008</td><td char="." valign="bottom" rowspan="1" colspan="1">−.004</td><td char="." valign="bottom" rowspan="1" colspan="1">.006</td><td char="." valign="bottom" rowspan="1" colspan="1">.001</td><td char="." valign="bottom" rowspan="1" colspan="1">.013</td><td char="." valign="bottom" rowspan="1" colspan="1">−.002</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.019)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.009)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.010)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td></tr><tr><td rowspan="2" valign="top" colspan="1">Two-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.011<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.002</td><td char="." valign="bottom" rowspan="1" colspan="1">.006</td><td char="." valign="bottom" rowspan="1" colspan="1">.003</td><td char="." valign="bottom" rowspan="1" colspan="1">.013<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.008<sup>+</sup></td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.006)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.019)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.013)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.008)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Observation count</td><td valign="bottom" rowspan="1" colspan="1">161,356</td><td valign="bottom" rowspan="1" colspan="1">161,356</td><td valign="bottom" rowspan="1" colspan="1">85,885</td><td valign="bottom" rowspan="1" colspan="1">85,885</td><td valign="bottom" rowspan="1" colspan="1">80,416</td><td valign="bottom" rowspan="1" colspan="1">80,416</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>8 This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models assessing whether subject-area specialization predicts staying in the same school. All analyses control for a set of teacher, classroom, and school covariates. <sups>+</sups> indicates statistical significance at the.10 level.</p> <p>Table A4. Does Subject-Area Specialization Predict Returning to the Same School (Fourth- and Fifth-Grade Teachers Only)?</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades 4–5</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Any specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.021*</td><td char="." valign="bottom" rowspan="1" colspan="1">.008<sup>+</sup></td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.009)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td></tr><tr><td rowspan="2" valign="top" colspan="1">One-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.021<sup>+</sup></td><td char="." valign="bottom" rowspan="1" colspan="1">.003</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.012)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.006)</td></tr><tr><td rowspan="2" valign="top" colspan="1">Two-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.022*</td><td char="." valign="bottom" rowspan="1" colspan="1">.012*</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.009)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Observation count</td><td valign="bottom" rowspan="1" colspan="1">53,572</td><td valign="bottom" rowspan="1" colspan="1">53,572</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>Graph</p> <p>9 This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models assessing whether subject-area specialization predicts staying in the same school. All analyses control for a set of teacher, classroom, and school covariates. <sups>+</sups> and * indicate statistical significance at the.10 and.05 levels, respectively.</p> <p>Table A5. Does Subject-Area Specialization Predict Returning to Teach in North Carolina?</p> <p> <ephtml> <table><thead><tr><th align="center" style="border-bottom: solid thin black" rowspan="2" valign="bottom" scope="col" colspan="1" /><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–5</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades K–2</th><th align="center" style="border-bottom: solid thin black" colspan="2" valign="bottom" scope="colgroup" rowspan="1">Grades 3–5</th></tr><tr><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">Teacher Fixed Effect</th><th align="center" style="border-bottom: solid thin black" valign="bottom" scope="col" rowspan="1" colspan="1">School Fixed Effect</th></tr></thead><tbody><tr><td rowspan="2" valign="top" colspan="1">Any specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.001</td><td char="." valign="bottom" rowspan="1" colspan="1">−.005*</td><td char="." valign="bottom" rowspan="1" colspan="1">−.000</td><td char="." valign="bottom" rowspan="1" colspan="1">−.005</td><td char="." valign="bottom" rowspan="1" colspan="1">.002</td><td char="." valign="bottom" rowspan="1" colspan="1">−.005</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.002)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.010)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.006)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.003)</td></tr><tr><td rowspan="2" valign="top" colspan="1">One-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">.005</td><td char="." valign="bottom" rowspan="1" colspan="1">−.006</td><td char="." valign="bottom" rowspan="1" colspan="1">−.012</td><td char="." valign="bottom" rowspan="1" colspan="1">−.009</td><td char="." valign="bottom" rowspan="1" colspan="1">.010</td><td char="." valign="bottom" rowspan="1" colspan="1">−.007</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.014)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.008)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.007)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td></tr><tr><td rowspan="2" valign="top" colspan="1">Two-subject specialization</td><td char="." valign="bottom" rowspan="1" colspan="1">−.001</td><td char="." valign="bottom" rowspan="1" colspan="1">−.005</td><td char="." valign="bottom" rowspan="1" colspan="1">.012</td><td char="." valign="bottom" rowspan="1" colspan="1">.001</td><td char="." valign="bottom" rowspan="1" colspan="1">−.002</td><td char="." valign="bottom" rowspan="1" colspan="1">.004</td></tr><tr><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.003)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.014)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.010)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.005)</td><td char="." valign="bottom" rowspan="1" colspan="1">(.004)</td></tr><tr><td valign="bottom" rowspan="1" colspan="1">Observation count</td><td valign="bottom" rowspan="1" colspan="1">161,571</td><td valign="bottom" rowspan="1" colspan="1">161,571</td><td valign="bottom" rowspan="1" colspan="1">85,960</td><td valign="bottom" rowspan="1" colspan="1">85,960</td><td valign="bottom" rowspan="1" colspan="1">80,582</td><td valign="bottom" rowspan="1" colspan="1">80,582</td></tr></tbody></table> </ephtml> </p> <p>Graph</p> <p>10 This table displays regression coefficients and cluster-adjusted standard errors (in parentheses) from models assessing whether subject-area specialization predicts returning to North Carolina public schools. All analyses control for a set of teacher, classroom, and school covariates. * indicates statistical significance at the.05 level.</p> <ref id="AN0174564218-21"> <title> Notes </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> We have no conflicts of interest to disclose. Kevin C. Bastian is a research associate professor in the Department of Public Policy at University of North Carolina at Chapel Hill and the director of the Education Policy Initiative at Carolina, whose research interests include educator preparation, employment, on-the-job learning, effectiveness, and retention; C. Kevin Fortner is an associate professor in the College of Education and Human Development at Georgia State University and conducts applied research focused on education and the application of quantitative research methods; Kate Caton is a doctoral student in the College of Education and Human Development at Georgia State University, and their research is focused on rural students and schools, student discipline, and quantitative methods. Kevin Bastian (ORCID: 0000-0003-2734-9133), C. Kevin Fortner (ORCID: 0000-0002-1910-4972), and Kate Caton (ORCID: 0000-0002-4588-5450). Correspondence may be sent to Kate Caton at kseymour4@gsu.edu.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref19" type="bt">2</bibl> <bibtext> In North Carolina, elementary school teachers typically hold a K–6 license that certifies them to teach all four academic content areas. Licenses for middle school and high school teachers apply to a single subject/content area (e.g., mathematics, biology).</bibtext> </blist> <blist> <bibl id="bib3" idref="ref23" type="bt">3</bibl> <bibtext> Teachers can be in both samples if they switch grade levels across years (e.g., from second grade in 2012 to third grade in 2013). Within a school year, teachers can also be in both samples if they teach multiple grades.</bibtext> </blist> <blist> <bibl id="bib4" idref="ref20" type="bt">4</bibl> <bibtext> Course codes are standardized across North Carolina. Please see https://<ulink href="http://www.dpi.nc.gov/educators/home-base/powerschool-sis/nc-sis-resources">www.dpi.nc.gov/educators/home-base/powerschool-sis/nc-sis-resources</ulink></bibtext> </blist> <blist> <bibl id="bib5" idref="ref16" type="bt">5</bibl> <bibtext> By using salary data from 2016–2017, we know the retention outcomes of elementary grades teachers (generalists and specialists) in the 2015–2016 year.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref25" type="bt">6</bibl> <bibtext> For teachers who do not return to NCPS, it is possible that they enter teaching in a private school in North Carolina, that they begin teaching in another state, or that they exit the teaching profession entirely. We do not possess the data to separate these alternatives.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref12" type="bt">7</bibl> <bibtext> We make first-year teachers the reference group and include single-year experience indicators for those with 1–15 years of experience. We include a single indicator for those with more than 15 years' experience. Some of these teacher characteristics (e.g., gender, race/ethnicity, and lateral/alternative preparation) will drop out of our preferred teacher fixed-effect models. We include these time-invariant teacher characteristics in our school fixed-effect models.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref14" type="bt">8</bibl> <bibtext> With a dichotomous outcome we also estimate logistic regression models. These results are similar to those from ordinary least squares models. We prefer the linear probability model because it is easier to interpret coefficients and is a more flexible model specification for the inclusion of fixed effects.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref44" type="bt">9</bibl> <bibtext> Black teachers make up 75% of the teachers of color in our analytical sample.</bibtext> </blist> <blist> <bibtext> By identifying early careers as those with less than 5 years' experience (rather than 3), we also allow more time to observe within-teacher changes in specialization status and retention.</bibtext> </blist> <blist> <bibtext> Teachers in North Carolina can earn evaluation ratings of developing, proficient, accomplished, and distinguished. Approximately 11% of the teachers in our sample earn distinguished ratings.</bibtext> </blist> <blist> <bibtext> It is possible that estimates from our teacher fixed-effect analyses identify teachers sorting to preferential schools, and that this sorting coincides with them becoming specialists. To address this concern, we estimated models in which we examine teachers switching from specialist to generalist. These results are available in Table A2, where the left panel shows that teachers switching from specialist to generalist are less likely to return to their school. This lessens concerns about teacher sorting coinciding with a switch to specialization.</bibtext> </blist> <blist> <bibtext> This cost estimate is in the middle of the low and high cost estimates used in Alliance for Excellent Education ([1]).</bibtext> </blist> </ref> <ref id="AN0174564218-22"> <title> References </title> <blist> <bibtext> Alliance for Excellent Education. (2014). On the path to equity: Improving the effectiveness of beginning teachers. https://all4ed.org/reports-factsheets/path-to-equity</bibtext> </blist> <blist> <bibtext> Anderson, R. C. (1962). The case for teacher specialization in the elementary school. Elementary School Journal, 62(5), 253–260.</bibtext> </blist> <blist> <bibtext> Atteberry, A., Loeb, S., & Wyckoff, J. (2017). Teacher churning: Reassignment rates and implications for student achievement. Educational Evaluation and Policy Analysis, 39(1), 3–30.</bibtext> </blist> <blist> <bibtext> Bastian, K. C., & Fortner, C. K. (2019). Is less more? Subject-area specialization and outcomes in elementary schools. Education Finance and Policy, 15(2), 357–382.</bibtext> </blist> <blist> <bibtext> Bastian, K. C., & Janda, L. (2018). Does quantity effect quality? Teachers' course preparations and effectiveness. Journal of Research on Educational Effectiveness, 11(4), 535–558.</bibtext> </blist> <blist> <bibtext> Blazar, D. (2015). Grade assignments and the teacher pipeline: A low-cost lever to improve student achievement? Educational Researcher, 44(4), 213–227.</bibtext> </blist> <blist> <bibtext> Borman, G. D., & Dowling, N. M. (2008). Teacher attrition and retention: A meta-analytic and narrative review of the research. Review of Educational Research, 78(3), 367–409.</bibtext> </blist> <blist> <bibtext> Boyd, D., Grossman, P., Ing, M., Lankford, H., Loeb, S., & Wyckoff, J. (2011). The influence of school administrators on teacher retention decisions. American Educational Research Journal, 48(2), 303–333.</bibtext> </blist> <blist> <bibtext> Cook, J. B., & Mansfield, R. K. (2016). Task-specific experience and task-specific talent: Decomposing the productivity of high-school teachers. Journal of Public Economics, 140, 51–72.</bibtext> </blist> <blist> <bibtext> Donaldson, M. L., & Johnson, S. M. (2010). The price of misassignment: The role of teaching assignments in Teach for America teachers' exit from low-income schools and the teaching profession. Educational Evaluation and Policy Analysis, 32(2), 299–323.</bibtext> </blist> <blist> <bibtext> Egalite, A. J., & Kisida, B. (2018). The effects of teacher match on students' academic perceptions and attitudes. Educational Evaluation and Policy Analysis, 40(1), 59–81.</bibtext> </blist> <blist> <bibtext> Egalite, A. J., Kisida, B., & Winters, M. A. (2015). Representation in the classroom: The effect of own-race teachers on student achievement. Economics of Education Review, 45, 44–52.</bibtext> </blist> <blist> <bibtext> Fox, L. (2016). Playing to teachers' strengths: Using multiple measures of teacher effectiveness to improve teacher assignments. Education Finance and Policy, 11(1), 70–96.</bibtext> </blist> <blist> <bibtext> Fryer, R. G. (2018). The "pupil" factory: Specialization and the production of human capital in schools. American Economic Review, 108(3), 616–656.</bibtext> </blist> <blist> <bibtext> Fuller, S. C., & Ladd, H. F. (2013). School-based accountability and the distribution of teacher quality across grades in elementary school. Education Finance and Policy, 8(4), 528–559.</bibtext> </blist> <blist> <bibtext> Gershenson, S., Hart, C., Hyman, J., Lindsay, C., & Papageorge, N. W. (2018). The long-run impacts of same-race teachers (Working Paper No. 25254). NBER.</bibtext> </blist> <blist> <bibtext> Gewertz, C. (2014). Platooning on the rise in early grades. Education Week, 33(21), 1.</bibtext> </blist> <blist> <bibtext> Goldhaber, D., Cowan, J., & Walch, J. (2013). Is a good elementary teacher always good? Assessing teacher performance estimates across subjects. Economics of Education Review, 36(C), 216–228.</bibtext> </blist> <blist> <bibtext> Goldring, R., Taie, S., & Riddles, M. (2014). Teacher attrition and mobility: Results from the 2012–13 Teacher Follow-up Survey. National Center for Education Statistics. https://files.eric.ed.gov/fulltext/ED546773.pdf</bibtext> </blist> <blist> <bibtext> Grissom, J. A., & Bartanen, B. (2019). Strategic retention: Principal effectiveness and teacher turnover in multiple-measure teacher evaluation systems. American Educational Research Journal, 56(2), 514–555.</bibtext> </blist> <blist> <bibtext> Hanushek, E. A., Rivkin, S. G., & Schiman, J. C. (2016). Dynamic effects of teacher turnover on the quality of instruction. Economics of Education Review, 55, 132–148.</bibtext> </blist> <blist> <bibtext> Henry, G. T., Bastian, K. C., & Fortner, C. K. (2011). Stayers and leavers: Early-career teacher effectiveness and attrition. Educational Researcher, 40(6), 271–280.</bibtext> </blist> <blist> <bibtext> Hill, H. C., Umland, K., Litke, E., & Kapitula, L. R. (2012). Teacher quality and quality teaching: Examining the relationship of a teacher assessment to practice. American Journal of Education, 118(4), 489–519.</bibtext> </blist> <blist> <bibtext> Holme, J. J., Jabbar, H., Germain, E., & Dinning, J. (2018). Rethinking teacher turnover: Longitudinal measures of instability in schools. Educational Researcher, 47(1), 62–75.</bibtext> </blist> <blist> <bibtext> Hwang, N., & Kisida, B. (2022). Spread too thin: The effect of specialization on teaching effectiveness. Educational Evaluation and Policy Analysis. https://doi.org/10.3102/01623737221084312</bibtext> </blist> <blist> <bibtext> Imazeki, J. (2005). Teacher salaries and teacher attrition. Economics of Education Review, 24(4), 431–449.</bibtext> </blist> <blist> <bibtext> Ingersoll, R. M., Merrill, L., & May, H. (2016). Do accountability policies push teachers out? Educational Leadership, 73(8), 44–49.</bibtext> </blist> <blist> <bibtext> Jacob, B. A., & Lefgren, L. (2008). Can principals identify effective teachers? Evidence on subjective performance evaluation in education. Journal of Labor Economics, 26(1), 101–136.</bibtext> </blist> <blist> <bibtext> Johnson, S. M., Kraft, M. A., & Papay, J. P. (2012). How context matters in high-needs schools: The effects of teachers' working conditions on their professional satisfaction and their students' achievement. Teachers College Record, 114(10), 1–39.</bibtext> </blist> <blist> <bibtext> Kraft, M. A., Marinell, W. H., & Shen-Wei Yee, D. (2016). School organizational contexts, teacher turnover, and student achievement: Evidence from panel data. American Educational Research Journal, 53(5), 1411–1449.</bibtext> </blist> <blist> <bibtext> Ladd, H. F. (2011). Teachers' perceptions of their working conditions: How predictive of planned and actual teacher movement? Educational Evaluation and Policy Analysis, 33(2), 235–261.</bibtext> </blist> <blist> <bibtext> Markworth, K. A., Brobst, J., Ohana, C., & Parker, R. (2016). Elementary content specialization: Models, affordances, and constraints. International Journal of STEM Education, 3, 16. https://doi.org/10.1186/s40594-016-0049-9</bibtext> </blist> <blist> <bibtext> Monk, D. H. (2007). Recruiting and retaining high-quality teachers in rural areas. Future of Children, 17(1), 155–174.</bibtext> </blist> <blist> <bibtext> Myrberg, E., Johansson, S., & Rosen, M. (2019). The relation between teacher specialization and student reading achievement. Scandinavian Journal of Educational Research, 63(5), 744–758.</bibtext> </blist> <blist> <bibtext> National Center for Education Statistics. (n.d.). Concentration of public school students eligible for free or reduced-price lunch. https://nces.ed.gov/programs/coe/pdf/coe_clb.pdf</bibtext> </blist> <blist> <bibtext> Nguyen, T. D., Pham, L. D., Crouch, M., & Springer, M. G. (2020). The correlates of teacher turnover: An updated and expanded meta-analysis of the literature. Educational Research Review, 31, 100355.</bibtext> </blist> <blist> <bibtext> Ost, B. (2014). How do teachers improve? The relative importance of specific and general human capital. American Economic Journal: Applied Economics, 6(2), 127–151.</bibtext> </blist> <blist> <bibtext> Ost, B., & Schiman, J. C. (2015). Grade-specific experience, grade reassignments, and teacher turnover. Economics of Education Review, 46, 112–126.</bibtext> </blist> <blist> <bibtext> Parker, A., Rakes, L., & Arndt, K. (2017, July). Departmentalized, self-contained, or somewhere in between: Understanding elementary grade-level organizational decision-making. Educational Forum, 81(3), 236–255.</bibtext> </blist> <blist> <bibtext> Perrachione, B. A., Rosser, V. J., & Petersen, G. J. (2008). Why do they stay? Elementary teachers' perceptions of job satisfaction and retention. Professional Educator, 32(2), 1–41.</bibtext> </blist> <blist> <bibtext> Ronfeldt, M., Loeb, S., & Wyckoff, J. (2013). How teacher turnover harms student achievement. American Educational Research Journal, 50(1), 4–36.</bibtext> </blist> <blist> <bibtext> Ronfeldt, M., & McQueen, K. (2017). Does new teacher induction really improve retention? Journal of Teacher Education, 68(4), 394–410.</bibtext> </blist> <blist> <bibtext> See, B. H., Morris, R., Gorard, S., Kokotsaki, D., & Abdi, S. (2020). Teacher recruitment and retention: A critical review of international evidence of most promising interventions. Education Sciences, 10(10), 262.</bibtext> </blist> <blist> <bibtext> Simon, N., & Johnson, S. M. (2015). Teacher turnover in high poverty schools: What we know and can do. Teachers College Record, 117(3), 1–36.</bibtext> </blist> <blist> <bibtext> Smith, T. M., & Ingersoll, R. M. (2004). What are the effects of induction and mentoring on beginning teacher turnover? American Educational Research Journal, 41(3), 681–714.</bibtext> </blist> <blist> <bibtext> Sorensen, L. C., & Ladd, H. F. (2018). The hidden costs of teacher turnover (Calder Working Paper No. 203-0918-1). The National Center for the Analysis of Longitudinal Data in Education Research (CALDER) and American Institutes for Research (AIR).</bibtext> </blist> <blist> <bibtext> Stockard, J., & Lehman, M. B. (2004). Influences on the satisfaction and retention of first-year teachers: The importance of effective school management. Educational Administration Quarterly, 40(5), 742–771.</bibtext> </blist> <blist> <bibtext> Sun, M. (2018). Black teachers' retention and transfer patterns in North Carolina: How do patterns vary by teacher effectiveness, subject, and school conditions? AERA Open, 4(3), 1–23.</bibtext> </blist> <blist> <bibtext> Wilkins, J. L. M. (2009). Elementary school teachers' attitudes toward different subjects. Teacher Educator, 45(1), 23–36.</bibtext> </blist> </ref> <aug> <p>By Kevin C. Bastian; C. Kevin Fortner and Kate Caton</p> <p>Reported by Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib21" firstref="ref3"></nolink> <nolink nlid="nl2" bibid="bib41" firstref="ref4"></nolink> <nolink nlid="nl3" bibid="bib46" firstref="ref6"></nolink> <nolink nlid="nl4" bibid="bib24" firstref="ref7"></nolink> <nolink nlid="nl5" bibid="bib19" firstref="ref8"></nolink> <nolink nlid="nl6" bibid="bib48" firstref="ref9"></nolink> <nolink nlid="nl7" bibid="bib42" firstref="ref10"></nolink> <nolink nlid="nl8" bibid="bib45" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib26" firstref="ref13"></nolink> <nolink nlid="nl10" bibid="bib31" firstref="ref15"></nolink> <nolink nlid="nl11" bibid="bib10" firstref="ref17"></nolink> <nolink nlid="nl12" bibid="bib38" firstref="ref18"></nolink> <nolink nlid="nl13" bibid="bib13" firstref="ref21"></nolink> <nolink nlid="nl14" bibid="bib18" firstref="ref22"></nolink> <nolink nlid="nl15" bibid="bib30" firstref="ref28"></nolink> <nolink nlid="nl16" bibid="bib44" firstref="ref30"></nolink> <nolink nlid="nl17" bibid="bib27" firstref="ref32"></nolink> <nolink nlid="nl18" bibid="bib29" firstref="ref33"></nolink> <nolink nlid="nl19" bibid="bib20" firstref="ref37"></nolink> <nolink nlid="nl20" bibid="bib36" firstref="ref38"></nolink> <nolink nlid="nl21" bibid="bib28" firstref="ref42"></nolink> <nolink nlid="nl22" bibid="bib37" firstref="ref45"></nolink> <nolink nlid="nl23" bibid="bib23" firstref="ref47"></nolink> <nolink nlid="nl24" bibid="bib34" firstref="ref48"></nolink> <nolink nlid="nl25" bibid="bib14" firstref="ref51"></nolink> <nolink nlid="nl26" bibid="bib25" firstref="ref54"></nolink> <nolink nlid="nl27" bibid="bib39" firstref="ref56"></nolink> <nolink nlid="nl28" bibid="bib32" firstref="ref60"></nolink> <nolink nlid="nl29" bibid="bib40" firstref="ref61"></nolink> <nolink nlid="nl30" bibid="bib47" firstref="ref62"></nolink> <nolink nlid="nl31" bibid="bib43" firstref="ref63"></nolink> <nolink nlid="nl32" bibid="bib17" firstref="ref64"></nolink> <nolink nlid="nl33" bibid="bib15" firstref="ref65"></nolink> <nolink nlid="nl34" bibid="bib33" firstref="ref67"></nolink> <nolink nlid="nl35" bibid="bib35" firstref="ref77"></nolink> <nolink nlid="nl36" bibid="bib11" firstref="ref79"></nolink> <nolink nlid="nl37" bibid="bib12" firstref="ref80"></nolink> <nolink nlid="nl38" bibid="bib16" firstref="ref81"></nolink> <nolink nlid="nl39" bibid="bib22" firstref="ref82"></nolink> <nolink nlid="nl40" bibid="bib49" firstref="ref85"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1401768
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Subject-Area Specialization and Teacher Retention: An Elementary School Story
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Bastian%2C+Kevin+C%2E%22">Bastian, Kevin C.</searchLink><br /><searchLink fieldCode="AR" term="%22Fortner%2C+C%2E+Kevin%22">Fortner, C. Kevin</searchLink><br /><searchLink fieldCode="AR" term="%22Caton%2C+Kate%22">Caton, Kate</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Elementary+School+Journal%22"><i>Elementary School Journal</i></searchLink>. Dec 2023 124(2):343-366.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: University of Chicago Press. Journals Division, P.O. Box 37005, Chicago, IL 60637. Tel: 877-705-1878; Tel: 773-753-3347; Fax: 877-705-1879; Fax: 773-753-0811; e-mail: subscriptions@press.uchicago.edu; Web site: http://www.press.uchicago.edu
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 24
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2023
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Evaluative
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Elementary+Education%22">Elementary Education</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Elementary+School+Teachers%22">Elementary School Teachers</searchLink><br /><searchLink fieldCode="DE" term="%22Elementary+Schools%22">Elementary Schools</searchLink><br /><searchLink fieldCode="DE" term="%22Intellectual+Disciplines%22">Intellectual Disciplines</searchLink><br /><searchLink fieldCode="DE" term="%22Specialization%22">Specialization</searchLink><br /><searchLink fieldCode="DE" term="%22Teacher+Persistence%22">Teacher Persistence</searchLink><br /><searchLink fieldCode="DE" term="%22Specialists%22">Specialists</searchLink><br /><searchLink fieldCode="DE" term="%22Teacher+Role%22">Teacher Role</searchLink>
– Name: Subject
  Label: Geographic Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22North+Carolina%22">North Carolina</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1086/727503
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0013-5984<br />1554-8279
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: School leaders need effective, affordable approaches to retain their teacher workforce. We investigated a promising, low-cost option for school leaders to encourage teacher retention: subject-area specialization in elementary grades (K-5). Using data on North Carolina elementary grades teachers and schools in the 2011-2012 through 2015-2016 academic years, we track the incidence of subject-area specialization, assess whether teaching in a specialist role promotes retention, and examine whether subject-area specialization is an effective retention strategy for certain schools and teachers. Descriptive analyses show specialization is common in upper elementary grades and has become a more widely used assignment strategy over time. Retention analyses indicate that elementary grades teachers are more likely to return to the same school after becoming a specialist. These results vary by school and teacher characteristics, suggesting that specialization may be a more effective retention strategy in urban schools, in non-high-need schools, and for Black teachers.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2023
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1401768
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1401768
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1086/727503
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 24
        StartPage: 343
    Subjects:
      – SubjectFull: Elementary School Teachers
        Type: general
      – SubjectFull: Elementary Schools
        Type: general
      – SubjectFull: Intellectual Disciplines
        Type: general
      – SubjectFull: Specialization
        Type: general
      – SubjectFull: Teacher Persistence
        Type: general
      – SubjectFull: Specialists
        Type: general
      – SubjectFull: Teacher Role
        Type: general
      – SubjectFull: North Carolina
        Type: general
    Titles:
      – TitleFull: Subject-Area Specialization and Teacher Retention: An Elementary School Story
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Bastian, Kevin C.
      – PersonEntity:
          Name:
            NameFull: Fortner, C. Kevin
      – PersonEntity:
          Name:
            NameFull: Caton, Kate
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 12
              Type: published
              Y: 2023
          Identifiers:
            – Type: issn-print
              Value: 0013-5984
            – Type: issn-electronic
              Value: 1554-8279
          Numbering:
            – Type: volume
              Value: 124
            – Type: issue
              Value: 2
          Titles:
            – TitleFull: Elementary School Journal
              Type: main
ResultId 1