Teacher Turnover Debate: Linda Darling-Hammond, Colleagues Respond to Critiques of Their Latest Study
In August, the Learning Policy Institute (LPI) issued a new report on teacher turnover, offering an in-depth analysis of how often teachers leave their schools and why. We focused on turnover because teacher attrition from the profession comprises about 90 percent of annual teacher demand, and drives many of the shortages we see today, particularly in high-need schools across the country where turnover can reach 30 percent or more annually.
One of our goals was to point out that solving these shortages cannot rest on recruitment alone, since high rates of turnover quickly undo schools’ efforts to bring in new hires. Another goal was to analyze the conditions associated with turnover to inform policy strategies to reduce the “leaky bucket” plaguing many districts.
In a recent blog in The 74, Chad Aldeman questioned our findings and the severity of the problem, suggesting that this emphasis might generate “too many … generic efforts to boost teacher retention, like districtwide pay increases.”
A week later, in an opinion piece in The 74, Mike Antonucci argued that teacher turnover is not a problem because it is, on average, lower than quit rates in some other occupations. He, too, protested what he called “generic solutions” and suggested that focusing on turnover rates is a “frightening” scare tactic.
Antonucci also pointed to the National Council on Teaching Quality’s (NCTQ) efforts to rebut teacher shortages. NCTQ claims the shortages must be a fiction, largely because the size of the teaching force is growing: “How could we have shortages if we’re hiring more teachers?” NCTQ’s Kate Walsh argues. Never mind that more than 100,000 vacancies have been filled by substitutes and untrained individuals because qualified teachers cannot be found to work where they are needed. Indeed, in our teacher shortage report last year, we accurately estimated this growth in the teaching force, which was why we predicted growing shortages, given the declines in entry rates to the profession.
We find these arguments puzzling at best. Our research on teacher shortages and the turnover that contributes to them has always emphasized that these conditions vary across teaching fields, types of schools, and locations. We document how they are more problematic in some regions, states, and districts than others; more widespread in particular subjects; and most pronounced in schools that serve students of color and those from low-income families.
We have highlighted research-based policy strategies that can address different contexts and needs, based on findings comparing recruitment and retention outcomes across places where turnover rates differ. By definition, this means that neither the problems nor the need for policy action are the same across settings. The goal is to help policymakers who care about the plight of all students to make better, more cost-effective investments where they are needed.
It seems our focus on teacher turnover and shortages disturbs those who worry that solutions to problems might become “generic” (i.e., widespread) — that states and districts might get carried away and make unwarranted investments in teachers’ training, compensation, or working conditions to create a more qualified and equitably distributed teaching force.
Of course, there are some highly stable places with strong salaries and working conditions that don’t need reform. But overall, teaching conditions here compare poorly with those of other nations. Unlike teachers in Organization for Economic Co-operation and Development (OECD) countries whose salaries are, on average, comparable to those of other college graduates, U.S. teachers are paid 30 percent less on average than other college graduates.
U.S. teachers’ wages have declined relative to those of other college-educated workers since the early 1990s, when they were at their most competitive. While salaries vary significantly across and within states, a recent report showed that, in more than 30 states, the average teacher heading a family of four would qualify for several forms of government assistance.
Furthermore, U.S. teachers teach the greatest number of hours per week among countries in a recent OECD survey and have nearly the lowest number of hours for planning. They also have above-average class sizes and teach more low-income students than teachers in the other higher-achieving OECD countries.
In this response, we address three areas of concern: 1) whether and how teacher turnover matters, 2) whether turnover varies in ways that are amenable to policies; and 3) what kinds of policy strategies may make a difference.
How teacher turnover matters
Aldeman agrees that teacher turnover matters, noting that “churn, in general, is a problem, particularly for disadvantaged students.” At the same time, he argues that turnover rates may not be all that high, are lower in public schools than private schools, and may not be related to the policies we highlight — arguments we address in the sections below.
High turnover undermines student achievement
Aldeman recognizes, as do we, that the most important reason that teacher turnover matters is its impact on students. Research shows that high teacher turnover rates in schools negatively impact student achievement for all the students in a school, not just those in a new teacher’s classroom.
These rates are highest in schools serving low-income students and students of color. Constant churn exacerbates staffing difficulties that lead to shortages. Thus, students in these hard-to-staff schools disproportionately suffer the consequences of both turnover and shortages: substitute teachers, canceled classes, and inexperienced, underprepared teachers. Office for Civil Rights data show that districts serving children of color are about four times as likely as other districts to be assigned uncertified teachers.
Turnover imposes significant financial costs
Turnover also extracts a significant financial cost. Research shows that teacher replacement costs, including expenses related to separation, recruitment, hiring, and training, can range from an average of $9,000 per teacher in rural districts to more than $20,000 in urban districts. A 2007 study estimated a national cost of over $7 billion a year, a price tag that would exceed $8 billion today. In high-turnover settings, it’s important to consider what else these dollars could buy — including teacher mentoring and learning opportunities to increase effectiveness — if they weren’t being spent on filling the leaky bucket each year.
Turnover undermines quality especially when there is an inadequate supply
If the supply of highly qualified teachers were plentiful, we might feel no need to worry about turnover, even if it is high and costly. That is not the case currently. In most states, headlines and data about the effects of shortages — especially in the fields of math and science, special education, and teachers of English learners — are commonplace. Data also show that shortages reach to other fields in the highest-need districts.
In our 2016 report on teacher shortages, which projected supply and demand trends based on several national databases, we estimated that U.S. schools could be short 100,000 qualified teachers by 2017 if policy trends did not change significantly. This estimate appears close to today’s reality. A recent LPI review of state teacher workforce reports reveals that, in the 36 states that reported such data in 2016 or 2017, at least 87,000 positions were not filled by a fully certified teacher. Assuming the same rates of shortages in the remaining states, the national total of uncertified teachers would be approximately 109,000. This number understates total shortages because some states only report uncertified teachers in core academic areas, and because districts also address shortages by canceling courses, increasing class sizes, or hiring substitute teachers. For example, Florida does not report uncertified teacher counts, but reported 6,628 unfilled vacancies at the start of 2017–18.
Solving shortages is much more difficult if turnover rates remain high, as roughly 9 of 10 teachers hired each year are replacing colleagues who left teaching, two-thirds of them having quit before retirement. When shortages exist, recruiting costs increase — as districts try to recruit from other states and countries and high-need teachers turn down many job offers to take the best one, usually from the most well-heeled district. In addition, the negative effects on student learning climb, as the Band-Aids to address vacant classrooms — substitutes, untrained staff, canceled classes, larger class sizes — undermine learning.
Small changes in turnover rates can have big effects on the adequacy of supply
But does the increase in turnover rates matter? Aldeman notes that attrition rates increased from 5.6 percent to 7.7 percent between 1988 and 2012. He characterizes this as an increase of only 2.1 percentage points, suggesting that such an increase is tiny. However, in a workforce of 3.8 million, this seemingly small amount adds 79,000 to the number of teachers needing to be hired each year, leaving a total of about 292,000 teachers to be hired annually.
In fact, the swing in attrition rates has been even greater than that. This National Center for Education Statistics report shows that, at its lowest point, in 1992, attrition was 5.1 percent annually. At its highest point, in 2008, it was 8.4 percent — an increase of about 60 percent. Although this represents “only” 3.3 percentage points, a return to this lower level of attrition would allow us to hire 125,000 fewer teachers — virtually ending today’s shortage conditions, with the exception of targeted needs in particular content areas. Small changes in the national attrition rate have a large impact on teacher labor markets.
While multiple factors influence attrition rates, a recent analysis of wage trends shows that the lower attrition rates in the late 1980s and 1990s occurred when the gaps between teachers’ wages and those of other college graduates were at their smallest. The gaps have grown significantly since then.
Teacher turnover rates vary substantially and can be addressed by policy
Aldeman suggests that policy may not be a cause or cure for turnover because turnover merely tracks the demographics of the teaching force, and anyway, private school leaving rates are higher than public school rates. But he does not address the policy implications of these observations.
Demography is not destiny
Aldeman downplays the significance of increased attrition rates, arguing that they are mainly due to demographic patterns — the natural “greening” of the profession as newer teachers enter and veteran teachers retire. In 1988, the most common teacher was a veteran with 15 years’ experience; by 2012 — the most recent year for which national data are available — the most common teacher was an early-career teacher in her fifth year.
This trend Aldeman points to is important, but he stops short of examining the implications of this situation. First, retirement represents only about one-third of overall annual attrition, and less than one-fourth in 30 states. Stemming the attrition of early- and mid-career teachers can go a long way to reducing turnover and mitigating teacher shortages.
Second, it is not inevitable that beginners leave at high rates: Policies influence turnover. Beginning teacher attrition is higher for teachers who are underprepared and unmentored. Schools and Staffing Surveys (SASS) data indicate that teachers with little or no training leave at two to three times the rates of teachers with comprehensive preparation. Another survey — the Beginning Teacher Longitudinal Study — found that among teachers who entered in 2007–08, only 14.6 percent of those who were certified to teach left over the first five years, compared with 30.2 percent of those who were uncertified when they entered.
Similarly, teachers who lack induction supports leave at about twice the rate of those who receive the highest-quality induction. However, the proportions of teachers receiving both sets of supports have been declining in recent years. In 2012, only two-thirds of entering teachers were prepared before they entered, and only 59 percent received support from a mentor and their principal in their initial year on the job, down from 75 percent in 2008. The problem feeds on itself. One reason some states and districts have seen heightened turnover in the teaching force is that they have not had policies in place that could retain beginners and stabilize the profession over time.
Private schools have policies too
Aldeman also suggests that higher attrition at private schools “throws a wrench” into the idea that policies can influence turnover. However, evidence suggests that private and public school teachers leave for similar reasons. An NCES analysis indicates that private school teachers who left their schools reported less administrative support, lower satisfaction with salary, less control over classroom policies, and less input into school policies — the same issues raised by public school teachers.
Private school teachers also earn significantly lower salaries, on average, than do public school teachers. Furthermore, high leaving rates are most pronounced for beginners, who — like public school teachers — leave more quickly when they are untrained. Our analysis of SASS data found that only about 40 percent of private school teachers in 2007 had completed the preparation and requirements for full state teaching certifications, compared with about 90 percent of public school teachers. Thus, it is not surprising that attrition rates were higher.
But if one wanted to lower these rates, similar policy strategies would apply within organizations in that sector: improve administrative supports and/or salaries where they are inadequate, and support training and mentoring that allow teachers to become more competent and efficacious, which reduces attrition.
Variable attrition indicates how policy matters
Aldeman’s analyses focus on changes in average turnover rates, but these averages mask enormous variations, which are themselves the result of differing policies. Currently, approximately 8 percent of teachers leave the profession annually and another 8 percent change schools, creating additional turnover at the school level. But annual teacher turnover rates range from 8 percent in Utah (where only 3 percent of teachers leave the profession entirely) to 24 percent in Arizona (where fully 15 percent of teachers leave the profession). Attrition is generally highest in the South and Southwest, and lowest in the Northeast, where several states have leaving rates of under 5 percent. It is generally higher in shortage fields and in schools serving more students of color and low-income students.
These variations are not just a function of chance or demographics. In a regression analysis that controlled for many student, teacher, and school factors — including teaching field and school composition — we found that teacher turnover was significantly higher for those less prepared to teach, in districts with lower salary schedules, and among those who reported low levels of administrative support.
(The SASS 2012 data also listed pressures from testing and accountability policies as a reason 25 percent of teachers said they left. We reported this along with the other SASS findings, but did not include it in our recommendations, because we could not examine the issue further. Aldeman criticized our mention of this fact from the survey, incorrectly suggesting we had flagged it as a policy issue, and linked to a report which he said found otherwise. Although the report he cited found no increase in voluntary turnover associated with the accountability policies, it actually did find an increase in involuntary attrition associated with these policies — probably due to reconstitution and closure of schools — which is consistent with teachers’ reports that accountability policies were a reason that they left. Other research on test-based accountability also found that teachers left low-rated schools at disproportionately high rates. We mention this not to advocate for a policy related to this issue, but to point out the inaccurate treatment of our report and other research.)
Another argument Aldeman uses to try to dismiss teacher attrition concerns is that teaching has turnover rates like those of nurses and social workers — two other feminized professions that have long suffered from low wages — and argues that “if policymakers are worried about teacher turnover and teacher shortages, they should be just as concerned with nursing turnover and nursing shortages.”
Indeed, such concerns are widespread among policymakers and have produced legislation focused on funding to build the nurse pipeline and support retention through service scholarships and funding for preparation programs, as well as improving working conditions.
Policy solutions are possible
High teacher turnover is not inevitable: U.S. rates were once much lower when salaries were more competitive and working conditions were stronger. Teacher attrition rates are also much lower in high-achieving jurisdictions like Finland, Singapore, and Ontario, Canada, where attrition averages 3 percent to 4 percent per year, and where studies show that teachers are well trained, supported, and compensated. In Toronto, Ontario, where a well-designed teacher induction program extends through the first four years of a candidate’s tenure, retention rates for beginning teachers are 98 percent. Not long ago, before these conditions were in place, Ontario had attrition rates similar to those in the United States. Their efforts to improve succeeded.
Aldeman objects to this comparison to other countries, arguing that there are major cultural differences that would make these kinds of attrition rates impossible to achieve in the United States. He asks, “What if American teachers have higher turnover rates not just because of our education policies, but also as a result of being American?” Several American states, including Massachusetts and other New England states, have attrition rates well below 5 percent. It is not surprising that, like the countries we noted, these states also have policies and conditions associated with increased retention, such as competitive compensation, high-quality training, few entrants through low-quality backdoor routes, and positive working conditions.
In short, lower attrition rates are possible. We have seen them abroad, we have seen them in the United States in past years, and we see them in some states today.
Reducing teacher turnover where it is high is not a pipe dream; it’s a policy question. In the many states, districts, and schools where turnover rates are double or triple the national average, progress is possible. If we cut overall attrition from the current levels of 7.7 percent even to 5 percent annually, U.S. hiring needs would decrease by more than 100,000 teachers annually, cutting demand by more than one-third immediately. If we did this with high-retention pathways into teaching — like residencies with mentoring — the demand would drop further each year as these teachers stayed rather than cycling through. This reduction would virtually eliminate shortages, with the exception of some fields and locations that may require specific incentives.
There is no denying that there is a teacher turnover problem in many communities and that it is harming hundreds of thousands of students daily. Policymakers and practitioners can work together to improve the key factors associated with teacher turnover through stronger teacher preparation and support, competitive and equitable compensation, and supportive teaching conditions.
It is hard to understand how building a well-prepared educator workforce that is well supported and adequately compensated for their work is a “frightening” idea. If we value high-quality schools, we must attend to how we can recruit and retain talented, dedicated people who want to stay and do the vital work of educating our children to become informed and productive citizens. If that is indeed a shared goal, it’s time to turn from denying the problem to working on the solutions.Submit a Letter to the Editor