IN THIS “RESEARCH INSIGHTS” EDITION

In this edition of our “Research Insights” series, we examine what four different NNERPP members have found when investigating teacher retention and teacher evaluation. We first highlight two studies on teacher retention, including one that examines how working conditions are related to teacher retention in Richmond, VA, and the other examining which teacher supports are associated with teacher retention in Michigan. We then highlight two studies on teacher evaluation, which take a close look at potential inequities in how teachers are evaluated in their home states of Tennessee and Wisconsin by exploring which teachers tend to receive higher or lower evaluations. Let’s dive in!

WHY THIS ARTICLE

Teacher shortages in the U.S. are leading to serious problems in ensuring all students have access to qualified teachers, negatively affecting student achievement. Understanding how to retain and support high quality teachers is therefore critical. Generally speaking, teacher evaluations are intended to provide evidence supporting pedagogical improvement; improving teacher quality is then assumed to also improve student learning. Evaluations have the potential to encourage teachers’ professional growth by providing feedback on their strengths and weaknesses, but may also function as a mechanism to remove teachers who are performing poorly. Evaluations can thus have serious consequences for teachers, as they can lead to dismissal, placement, and compensation decisions. As much as teacher evaluations are intended to improve the quality of teachers and therefore the quality of student learning, they only work as such if they are robust, fair, and objective – and if they are not perceived as such by teachers, teachers may leave their profession.

Especially in the current environment where the challenges teachers experience on a day-to-day basis are particularly salient and leading to serious retention problems due to pandemic-induced burnout and turnover, we turn to research produced by RPPs on the topic of teacher retention and teacher evaluation to see what they’ve found.

OVERVIEW

Let’s first take a quick look at the four artifacts included in this article. In Table 1 below, you’ll find the partnership name as well as a brief description of the partnership in column 1 and the title and link to each research artifact in column 2.

Table 1. List of RPPs + Artifacts Included in This Article

RESEARCH SNAPSHOTS: TEACHER RETENTION

We begin by introducing the two studies examining teacher retention. We encourage you to check out each individual study to dive deeper into the intricacies of the data and findings.

(I) RICHMOND, VA AREA

Research Questions

This study examined how teachers in the MERC region (seven Richmond-area school districts) reported on their working conditions and how reports varied by teacher characteristics, and whether teacher working conditions predicted teachers’ intentions to stay or leave their current school building.

Context and Research Methods

The study is part of the larger MERC Teacher Retention study, which identifies patterns of teacher retention in the MERC region and determines the school and system-level factors driving them. With many existing efforts to address teacher shortages being based on improving teacher working conditions, this study presented the findings from an analysis of the 2019 Virginia Department of Education Working Conditions Survey and examined which working conditions predicted a teacher’s intent to stay or leave their current school.

In the spring of 2019, the Virginia Department of Education (VDOE), in partnership with the University of Virginia, administered the working conditions survey to any teacher or school staff holding a license from the VDOE. All schools were required to participate, but individual teacher participation was voluntary. The survey operationalized working conditions within four broad constructs: (1) professionalism; (2) teaching, instruction, and student services; (3) school and community supports; and (4) safety. Each of the four main constructs was measured using a subset of scales, each composed of a set of items. Statewide, 1,678 (93%) schools participated in the survey with a total of 54,207 responses from teachers leading to a 67% response rate across all participating schools.

What Does the Research Show?

Compared to teachers across Virginia (the full sample), teachers in the MERC districts were slightly less likely to indicate an intention to stay in their current school (77.8% versus 82.1%). Middle school teachers were slightly less likely to indicate an intent to stay (75.9%) than any other group. Black teachers were less likely to indicate an intent to stay (70.6%) than any other group. Of all of the working conditions scales, teachers were most satisfied with collegiality and least satisfied with demands on their time. Strong school leadership emerged as the most important working condition for teachers in terms of predicting an indication to stay at their current school. Teacher autonomy also emerged as an important predictor to stay. On the other hand, teachers were least likely to indicate an intention to stay if they viewed the school as having insufficient support and insufficiently clear policies for managing student behavior.

How Was the Work Used in Practice?

We asked the RPP teams how the research studies were used by their practice-side partners. Here’s what the MERC team shared: In addition to the report, there are several additional dissemination efforts connected to this work designed to inform district and school level practices. This includes a podcast with report authors and local school leaders, and presentation of the findings at local professional conferences. The research from this project is also being developed into professional learning modules for school leaders that encourage professional reflection on the research in relationship to leadership practices. The MERC team is currently working in collaboration with district professional learning offices to pilot the modules.

(II) MICHIGAN

Research Questions

To inform efforts by the Michigan Department of Education to improve the retention of effective teachers, especially in high-poverty, low-performing local education agencies, this study examined which teacher supports implemented by local education agencies (traditional school districts and charter schools) in Michigan were associated with teacher retention.

Context and Research Methods

The study examined the average annual retention rate for teachers who taught in Michigan public schools between 2013/14 and 2018/19 and used data from a survey administered by the Michigan Department of Education between September 18 and October 12, 2020 to a sample of certified teachers who had been teaching for three to five years in their local education agency. The survey responses indicated teachers’ awareness and perceptions of the supports provided by their local education agency. The survey included 30 policies, practices, or programs. The final survey sample included 539 teacher respondents from 305 local education agencies for a response rate of 12.2% of teachers who were estimated to be eligible (note that due to the low response rate, findings from the study’s analyses of survey data should not be viewed as representative). The sample for the teacher retention rate data included 114,283 teachers.

What Does the Research Show?

The supports that were associated with teacher retention varied by the type of local education agency and the percentage of students who were economically disadvantaged. Generally, teachers were more likely to stay in local education agencies that had regular supportive communication between new teachers and school leaders, those that implemented mentoring programs for new teachers and provided new teachers with an orientation to their school, those that gave teachers opportunities to set goals in their evaluations, those that provided teachers with sufficient instructional resources, and those that provided teachers with release time to attend professional development and training opportunities. 

Teacher retention was higher in local education agencies that served lower percentages of students who were economically disadvantaged, higher percentages of students who were White, and higher percentages of students proficient in English language arts.

How Was the Work Used in Practice?

Here’s what the REL Midwest team shared about how their practice-side partners are using the research findings: 

The research findings were summarized in a 15-page report, a 4-page brief, and a 1-page brief that were shared with partners interested in learning about how they can improve their district’s teacher retention efforts. In addition, REL Midwest developed a blog post and a companion infographic, which were designed to reach key practice partners to highlight the key takeaways and provide recommendations for what education leaders can do to improve teacher retention in their district. REL Midwest also conducted a training project building on the report to create practitioner-focused materials to help district leaders improve their teacher recruitment and retention strategies. In this training project, REL Midwest worked with Michigan Department of Education staff to translate research findings and recommendations into practitioner-focused resources that school districts can use to improve their teacher workforce. In particular, this work considered the needs of small districts that do not always have the local resources needed to tackle their recruitment and retention needs. The resources developed for this project include (a) a handout featuring definitions and examples related to educator total compensation plans, (b) an interactive practitioner’s guide to the supports associated with teacher retention in Michigan, and (c) a resource guide for Michigan districts on nontraditional teacher career pathways to expand teacher recruitment and retention.

RESEARCH SNAPSHOTS: TEACHER EVALUATION

Next, let’s take a look at the two studies on inequities in teacher evaluation. For more detailed information on the data and findings, please explore each individual study.

(I) TENNESSEE

Research Questions

This study examined Tennessee teachers’ observation scores – which make up a large percentage of teacher’s overall evaluation scores in Tennessee – from 2011/12 to 2018/19 by teacher characteristics, paying special attention to teacher gender and race.

Context and Research Methods

In Tennessee, teacher observation scores comprise 50–70% of a teacher’s final overall evaluation rating. The most commonly used observation rubric is the TEAM rubric, which has four domains – instruction, planning, environment, and professionalism – with multiple indicators each, and uses a scale of 1 (“significantly below expectations”) to 5 (“significantly above expectations”) for each indicator. Note that this study examined teacher observation scores only, and not the overall evaluation scores. Approximately 460,000 teacher-by-year observations from 2011/12 to 2018/19 were included in the sample.

What Does the Research Show?

Black teachers consistently received lower classroom observation scores than their White counterparts in each of the examined years, across every observation rubric, and at every school level. This also held true when they had similar qualifications and their students achieved similar test scores and other outcomes. The racial gap in observation scores was largest in schools with few black teachers and grew smaller in schools with more racially diverse faculties. The gap disappeared in schools where a little more than half of the teachers were black. This study also looked at the role that the race of the teacher observer might play and found it had a very small impact on observation scores. 

In terms of gender, the study found that male teachers received lower classroom observation scores than their female counterparts. This was the case in each of the examined years, across every observation rubric, and at every school level, even when they had similar qualifications and their students achieved similar test scores and other outcomes. There was no evidence that the gender of the teacher observer impacted the score.

How Was the Work Used in Practice?

TERA’s findings raise concerns that the observation score gap reflects some form of systemic bias—that is, that Black and White (or male and female) teachers receive systematically different observation scores even when they have similar student achievement growth scores. Importantly, bias in this sense does not require individual observers to be biased against particular groups of teachers. Nonrandom sorting of students within schools, which the research documents, could be a source of bias, if teachers tend to receive lower observation scores when they teach students who bring some challenges to the classroom—such as a history of disciplinary infractions, which may require a greater emphasis on classroom management—that other students do not bring. Another source of bias could be the observation rubrics themselves, which may give higher marks to teaching practices that some teachers are more likely to employ, even when other practices, employed by other teachers, are similarly effective.

Our findings will inform several follow-up studies that will seek to understand more about the source of the systemic differences we have uncovered, and identify possible solutions to mitigating these biases from the evaluation system in the future. Additionally, state-level advocacy groups are currently using TERA’s findings to lead important conversations about race and diversity in Tennessee’s teacher workforce.

(II) WISCONSIN

Research Questions

The study examined statewide effectiveness ratings data of Wisconsin educators from 2014/15 to 2019/20 for racialized and gendered differences.

Context and Research Methods

Wisconsin uses the Wisconsin Educator Effectiveness (EE) System, which requires that schools provide ongoing, standards-based feedback to educators about their professional practices. Districts may use the Danielson Framework for Teaching (FfT) or the Stronge Teacher Effectiveness Performance Standards (EP) as rating rubrics or can apply to use another equivalent rubric. The Danielson Framework for Teaching is the most commonly used rubric. It has four domains – Planning & Preparation, Classroom Environment, Instruction, and Professional Responsibilities – and a total of 22 components and uses a one-to-four scale from Unsatisfactory (1) to Distinguished (4). The study included 93,299 ratings representing 55,963 educators who received effectiveness ratings at least once between the 2014/15 school year (the first year of statewide implementation) and the 2019/20 school year.

What Does the Research Show?

Looking at the intersection of gender and race, the study found that Black and Asian male educators were rated as least effective while White female educators were rated as most effective: 89% and 78% of White female educators were rated as more effective than the average Black and Latinx male educator, respectively. For educators matched by experience and credentials within the same school, the gap between ratings for White female educators and ratings for other educator groups grew smaller but persisted, and was still largest between White female and Black male educators.

How Was the Work Used in Practice?

The Wisconsin Educator Effectiveness Research Partnership is planning a number of follow-up studies to better explore the issues of bias and their sources as well as potential strategies, for example through identifying schools that have demonstrated less bias in their teacher evaluations to see what might be learned from these schools. The partnership hopes that taken together, findings from these studies can inform guidance to the Wisconsin Department of Public Instruction for how to make the Wisconsin Educator Effectiveness process fairer, better support educators and students of color, and reduce opportunity gaps.

IMPLICATIONS AND RECOMMENDATIONS

Teacher retention: The two studies identified some teacher working conditions (Richmond, VA) and some teacher supports (Michigan) that were more predictive of teacher retention than others, helping to pinpoint approaches that can help bolster teacher retention going forward. In the Richmond-area school districts, continuing to build and support strong school leadership and giving teachers autonomy emerged as the most promising approaches. The finding that Black teachers were less likely to indicate an intent to stay than any other group is both troubling and harder to address, particularly since further investigations into the relationship between race, working conditions, and teacher retention were beyond the scope of this specific study. It does point to the need to investigate and address the barriers that teachers of color face – similar to what the findings of the two studies focused on teacher evaluation suggest. In Michigan, the following recommendations emerged: Prioritizing supports for new teachers, prioritizing adjustments to teacher evaluation systems to provide opportunities for teachers to set their own goals (note the connection between teacher retention and teacher evaluation here), and prioritizing professional development and instructional resources for teachers. 

Teacher evaluation: Both studies found that teacher ratings were not consistently assigned. Teachers of color and male teachers were more likely to receive low ratings than their White and female counterparts even when other variables were controlled for. This may point to the need to improve the design, implementation, and monitoring of teacher evaluation systems in order for them to be a reliable and unbiased means of improving teacher quality and student learning. It may also be that bias runs much deeper: Ratings might reflect actual teacher performance in the classroom, but teachers of color might be consistently placed into schools or classrooms where it is more difficult to succeed. The studies make these observations when looking at drivers for observed gaps: The Tennessee study found that Black teachers tended to work in schools that served more students experiencing poverty, so the differing characteristics of students who are assigned to Black and White teachers might also play a role in observation scores – and paying more attention to student placement might help guard against such differences. The Tennessee study also examines the role that the racial composition of a school’s faculty played. Additionally, both studies point to the need to further investigate potential sources of bias and drivers of the observed inequities, including disentangling interpersonal and systemic bias (Wisconsin). The Tennessee study puts forth a recommendation of regular training for school leaders on potential sources of bias in teacher evaluation.

CONCLUSION

Findings from these four RPPs provide insight into the challenges teachers face, from difficult working conditions that can lead to retention problems to evaluation systems that offer a context for discrimination on the basis of teacher race and gender. Teachers must be supported in targeted ways and evaluation systems must be set up to guard against systemic and interpersonal bias if we hope to address teacher shortages and set up all teachers –and their students– for success.   

Suggested citation: Spitzley, N. (2022). Examining Teacher Retention and Teacher Evaluation With NNERPP Members. NNERPP Extra, 4(2), 2-8.