Tag: teacher wellbeing

By Clay Shumate

Teacher burnout in special education runs on a set of demands that does not appear on the general list at all. A caseload is a negotiated number, not a class roster. The paperwork carries statutory deadlines. The second adult in the room is a relationship to manage rather than a resource to use. The research on all three is consistent in direction and thinner in quality than you would hope.

This page does not restate the general list. The general demand ledger every teacher is working from already covers behavior, hours and out-of-pocket spending, and all of it applies here too. For the word itself, start with the three dimensions the word is actually made of. What follows is narrower: the demands belonging to this job specifically, in grades 6–12, and how good the evidence behind them really is.

Key takeaways

• Three demands separate this job from the general one: caseload as a negotiated number, compliance work with statutory deadlines, and co-teaching as a relationship rather than a resource. None gets smaller by managing your time better.
• Caseload is the most-named reason in the retention research: 33 percent of leavers in one urban district’s exit survey, more than any other factor — though that is 99 people in one district, not a national rate.
• Paperwork was third in the same survey, at 24 percent, and lands differently because a missed deadline is a legal event rather than a late grade. The system behind it is large and mostly never touches individual teachers: 29,490 due process complaints filed nationally in 2021–22, 2,645 fully adjudicated.
• In the co-teaching research the special educator is repeatedly observed in a subordinate role, most of all at secondary level. Across 32 studies, teachers reported roughly 45 minutes a week of joint planning and said they needed about three times that.
• Special education is the hardest area in the country to staff: in August 2024, 66 percent of high schools reported difficulty filling it, and only 77 percent of those vacancies went to a fully certified teacher.
• The evidence is directionally consistent and methodologically weak — of 30 studies in the major review, only two used designs supporting causal claims. The familiar “75 percent leave within five years” figure is one state’s data from 1987 to 2001, not a current rate.
• None of these numbers is a statement about you, and none is a diagnosis. Burnout is an occupational condition; nothing here is a screening instrument and nothing returns a score.

What makes special education a different burnout problem?

Three demands a general education teacher never meets, none of which you control the shape of. A caseload is set by somebody else, a compliance calendar by federal statute, a co-teaching partnership by a master schedule you did not build. Time-management advice is aimed at demands whose shape you can change, which is why it does not land here — and being told to work smarter about something you do not control is a reliable way to feel worse about it.

One thing to settle before the numbers start. Special educators report heavier loads and leave at higher rates, and it is easy to read that as a statement about the people who do the job. It is not. Every figure below is an average across thousands of teachers in very differently resourced buildings, and an average says nothing about any individual in it. Reading it as a verdict on yourself is statistically wrong, and it is the self-blame that makes a hard year harder.

How many students is a special education caseload?

Nobody can tell you, and that is the problem. A caseload is not a class size. It is a number produced by state rule, district practice, how many positions are filled, and what got decided in a meeting in July. Two teachers with the same title in the same state can be carrying very different loads.

It is also what special educators name first. In a survey of teachers who left or moved out of one large urban district, 33 percent named large caseloads as an important reason — more than any other single factor. A teacher quoted in the same research put it plainly: “I currently see well over 75 students a day with little support.”

Label that number before you carry it anywhere. It comes from 99 special educators, one district, three years. It is not national, and it does not tell you what a reasonable caseload is — no study does. It tells you what people say on the way out, and they have been saying it for a long time.

The useful consequence: because a caseload is negotiated rather than given, it is one of the few demands here that somebody with a title can change in a scheduling meeting. That is not easy, but the conversation has a right address.

Why does special education paperwork land differently?

Because a missed deadline is a legal event rather than a late grade. Every teacher has paperwork. Very little of it sits inside a federal statute with a dispute-resolution process attached to the far end.

Paperwork was third in that same district survey, at 24 percent. One teacher described it in terms any teacher will recognise: the IEPs ate up much of the time previously spent on students. The complaint is not that the documents are pointless, but that the documents and the students compete for the same hours.

The system behind those deadlines is genuinely large. In 2021–22, states and outlying areas reported 29,490 due process complaints filed nationally, 6,160 written state complaints and 10,806 requests for mediation — up from 17,118 due process filings a decade earlier, with the share still pending at year end rising from 12 percent to 43 percent.

Two qualifications, because this is the kind of number used to frighten people. Most of that volume never reaches a hearing — 2,645 were fully adjudicated — and the overwhelming majority of special education teachers finish a career without being personally involved in any of it. The point is not that you are about to be sued. It is that the deadlines are not negotiable the way a grading deadline is, which makes the pressure attached to them rational rather than something to talk yourself out of.

What does co-teaching do to the job?

In the research it repeatedly turns the special educator into the junior adult in somebody else’s room, and it does this most at secondary level. That is not a complaint about colleagues; it is what falls out of a structure where one adult owns the content and the other is assigned to the students.

The largest qualitative synthesis of co-teaching covered 32 studies, 454 co-teachers, 42 administrators, 142 students, 26 parents and 5 support staff. Its central finding: the special education teacher was often observed to play a subordinate role, typically responsible for modifying instruction and managing behavior while the general educator controlled the content. One special educator described the job as being an aide sometimes, an interpreter sometimes, and sometimes a teacher.

The pattern sharpened at secondary level, where specialised content knowledge decides who leads and teachers reported less training and more negative attitudes about co-teaching than elementary counterparts. If you teach grades 6–12, that is your setting, not a footnote. The six recognised co-teaching structures divide that work very differently, and it is worth knowing what the research has actually measured for each one before a partnership is assigned to you.

The pressure point is planning time. Teachers reported about 45 minutes a week of joint planning and said they needed close to three times that — and the authors note it was discussed almost entirely as a question of administrative support. That is the right framing, because planning time is a scheduling decision somebody signs off on, not a personal organisational failing.

This is the one item on the page that genuinely responds to coaching. The subordinate pattern is produced by defaults — who stands where, whose name is on the gradebook, who the class looks at when directions are given — and an explicit division of roles, said out loud in front of students in the first week, changes what the room assumes for the year. If you are the general education teacher in that room, most of that fix is yours. Though a coaching cycle offered instead of staffing relief will be read as exactly what it is.

Does the staffing shortage make all three worse?

Yes, and it is the mechanism connecting them. Every unfilled special education position is a caseload that gets distributed, a compliance calendar absorbed, and a co-teaching assignment covered by whoever is available.

The federal survey data is unambiguous about where the shortage bites. In a national panel of 1,392 public schools surveyed in August 2024, special education was the hardest teaching area in the country to staff: 66 percent of high schools and 74 percent of middle and combined schools reported difficulty filling those vacancies, and only 77 percent of the ones filled went to a fully certified teacher — eleventh of thirteen subject areas, against 89 percent for social studies.

Read that last figure the right way. It is no judgement on people hired under emergency credentials, who are usually doing a hard job with less preparation than they deserved. It is a statement about what the building asks of everyone else, because an under-prepared colleague needs support that comes out of somebody’s week — usually the experienced special educator already carrying the largest caseload. A first-year special educator is short of both at once: short of the stock every new teacher lacks, and short of specialised knowledge the role needs on day one.

What the data does not tell us is where inside a district those gaps land. The major review notes that few studies examined school poverty or student demographics at all. It is reasonable to suspect the shortfall concentrates in the schools least able to absorb it — but that has not been established, and a district that wants to know should look at its own vacancy and caseload numbers by school rather than assume either way.

How good is the evidence behind any of this?

Consistent in direction, weak in design, and worth knowing before you quote it at anybody. This literature has a myth problem, and a number that falls apart under scrutiny costs more than it gains.

The major review of special educator attrition covers 30 studies from 2002 to 2017, and its accounting of itself is blunt. Only two of the 30 used methods permitting causal inference. Only 13 examined teachers actually leaving; the rest measured intent to leave or administrator perceptions, which are related but not the same. Mentoring, which teachers called valuable in the qualitative studies, showed insignificant direct effects on retention in the quantitative ones.

The most recent burnout synthesis, covering 29 studies from 2013 to 2023, describes its own base as predominantly cross-sectional analyses of convenience samples — a snapshot of whoever responded, not a study over time.

The co-teaching synthesis carries a limitation its authors state outright: in nearly a third of the studies, participants were selected as outstanding examples of co-teaching, which they say likely presents a more favourable picture than exists generally. That cuts in an unusual direction — it suggests the real picture is worse, not better — and that is a reason to hold the finding loosely, not a licence to assume the worse version.

The number to stop repeating. “Seventy-five percent of special educators leave within five years” is a real finding — Illinois administrative data on teachers who entered between 1987 and 2001, 42 percent leaving teaching and 34 percent moving schools. One state, and a cohort that started work before most teachers now in classrooms were born. It is not a current national rate. The nearest national figure in the same review is older than it sounds too: 17.1 percent of special educators left their schools in 2012–13.

And one finding that cuts against the simple story. A study of novice teachers in their first three years found, as expected, that special educators rated their workloads significantly less manageable than general educators did, and that workload manageability predicted emotional exhaustion, which predicted whether they meant to keep teaching. But the effects were stronger for the novice general educators. One reading is that special educators arrive expecting an unmanageable workload and are less shaken when they get one, which is not comforting. Another is that the measure misses what is heavy about this job. That study sampled elementary and middle school novices in two states, so it is not a high school result either.

The honest summary: the direction is well supported and the fixes are not. Special educators consistently report heavier, less manageable workloads, name caseload and paperwork first when they leave, and are harder to replace than anyone in the building. What almost nothing here establishes is that any particular intervention fixes it.

What can you actually check, and who do you take it to?

Four things, all facts about the job rather than judgements about you. The point is to turn a feeling of being underwater into four numbers somebody with a title can act on. Short on time, which is the whole premise? Do items one and three only — five minutes, and the pair most likely to produce a change.

1. The caseload number, and where it came from. How many students today, how many in August, who decided. If nobody in the building can name the rule it came from, that is itself the finding.
2. Compliance events per week for the next 60 days. A count by week — meetings, evaluations, reviews. Not to panic about; to see whether the load is spread or stacked into two weeks that were always going to break.
3. Co-planning time, scheduled versus actual. What the master schedule says you get with each co-teacher, and what you got last month. The gap is the request.
4. Unfilled positions, and whose work landed where. What is vacant, what is covered provisionally, which pieces came to you.

Three rules for keeping it. Write counts and dates, never names or initials — item two can identify a student with no name on it at all, because a dated meeting is traceable, which is why it is a count per week rather than a list. Keep it somewhere that is yours; a district device is not a private place to write that you are not coping. And keep it separate from compliance itself: this is a record of workload, not a log of deadlines you expect to be missed. If you believe a legal deadline is about to be blown, that goes to your coordinator today, on its own, in writing.

It goes to the special education coordinator or department chair first, then whoever builds the master schedule. Open it as a question rather than a case — “here is what I am carrying, can we look at whether this is the intended distribution” lands differently from a document handed over in silence. Ask for the criterion rather than the favour: the conditions worth asking to have changed work better when the rule for who qualifies is written down first. A load lightened by private arrangement creates a fairness problem down the hall.

If you are the only special educator in your building — common in small secondary schools — there is no department to take this to. The chain then runs to the district’s special education director, and sideways to a counterpart at another school or your state’s technical assistance center. Do not skip the step because the usual person does not exist.

One thing this sheet is not. It is not a test, it produces no score, and it cannot tell you whether you are burned out. There is no validated cutoff that turns numbers into a finding about a person. If what you are carrying has stopped being about the job — if it follows you home, into the weekend, into how you feel about yourself — that is a conversation with your doctor or a counsellor, not with a worksheet. A short weekly record you can actually keep beats any one-off audit, because it shows a direction instead of a verdict.

Are the students the reason?

No. This needs saying directly, because “special education is exhausting” slides easily into “these students are exhausting,” and that is both untrue and corrosive.

A caseload is a staffing decision. A compliance calendar is a federal statute. An unfilled position is a labour market. A co-teaching structure is a master schedule. Every load on this page came from an adult decision, and a student with an IEP is who those decisions exist to serve. Where the research finds this job heavy, it is heavy because of how it was resourced.

And if you have an IEP and you have noticed your teacher is stretched: none of it is yours to carry. The services on your plan are yours by law regardless of how the building is staffed, asking for them is not making trouble, and a staffing shortage is never a reason to ask for less than you are owed.

This matters practically as well as decently. A teacher who quietly concludes the students are the problem starts lowering the bar, and that is the one response that makes everything worse for everyone. Holding the bar where it belongs while the supports do their job is the whole design.

What to do next

Fill in the audit this week, or just items one and three. Most of it is already in your email and the master schedule. Then take the two biggest gaps to the person who can change them, in writing, as a question about the criterion rather than about you.

If the answer is no, you have still done something useful: a written record of a load and a dated request is what makes a pattern visible when somebody finally asks why the position turned over again. And if the honest answer is that this job cannot be made survivable where you are, that is a real answer, not a failure of nerve. Nobody gets credit for absorbing a staffing shortfall until it breaks them.

Frequently asked questions

Sources

• Billingsley, B., & Bettini, E. (2019). Special education teacher attrition and retention: A review of the literature. Review of Educational Research, 89(5), 697–744. A systematic review of 30 studies published between 2002 and 2017. Source of the 33 percent caseload and 24 percent paperwork figures (from Billingsley, 2007 — 99 special educators who left or moved from one large urban district over three years), the 17.1 percent national attrition figure for 2012–13 (Goldring, Taie & Riddles, 2014), and the “75 percent within five years” figure (DeAngelis & Pressley, 2011 — Illinois administrative data, 1987–2001). Limitations stated by the authors: only two of the 30 studies used methods permitting causal inference, only 13 examined teachers actually leaving rather than intent to leave, and few studies examined school poverty or student demographics. journals.sagepub.com
• Brunsting, N. C., Morin, L. E., Gómez, L. R., Jones, B., Bettini, E., Cumming, M. M., Garwood, J. D., & Ruble, L. A. (2026). Burnout and occupational wellbeing of special education teachers: Recent research synthesized. Review of Educational Research. A synthesis of 29 studies published between 2013 and 2023, extending the 2014 Brunsting, Sreckovic & Lane synthesis. The authors describe the evidence base as predominantly cross-sectional analyses of convenience samples and call for longitudinal and intervention designs. journals.sagepub.com
• Scruggs, T. E., Mastropieri, M. A., & McDuffie, K. A. (2007). Co-teaching in inclusive classrooms: A metasynthesis of qualitative research. Exceptional Children, 73(4), 392–416. 32 qualitative reports covering 454 co-teachers, 42 administrators, 142 students, 26 parents and 5 support personnel. Source of the subordinate-role finding, the secondary-level intensification, and the 45.5 minutes of weekly joint planning against a stated need of roughly three times that. Limitation stated by the authors: in nearly a third of the investigations participants were selected as outstanding examples of co-teaching, which they say likely presents a more favourable picture of co-teaching than exists in general. ncipp.education.ufl.edu (PDF)
• Bettini, E., Jones, N., Brownell, M., Conroy, M., Park, Y., Leite, W., Crockett, J., & Benedict, A. (2017). Workload manageability among novice special and general educators: Relationships with emotional exhaustion and career intentions. Remedial and Special Education, 38(4), 246–256. Novice teachers in their first three years, drawn from the Michigan–Indiana Early Career Teacher Study. Special educators rated workload manageability significantly lower than general educators; workload manageability predicted emotional exhaustion, which mediated the relationship with intentions to keep teaching. Effects were stronger for novice general educators, and the sample was elementary and middle school teachers in two states — not a high school result. journals.sagepub.com
• National Center for Education Statistics (2024). School Pulse Panel, hiring and vacancies collection. 1,392 participating public K–12 schools, surveyed 13–27 August 2024. Special education was the hardest teaching area to staff: 74 percent of elementary and of middle and combined schools and 66 percent of high schools reported difficulty filling those vacancies, and only 77 percent of filled special education vacancies went to a fully certified teacher — eleventh of thirteen subject areas, against 89 percent for social studies. ies.ed.gov
• CADRE (Center for Appropriate Dispute Resolution in Special Education). IDEA dispute resolution data summary for U.S. and outlying areas. Data for the 2021–22 school year, reported by states under IDEA Section 618: 29,490 due process complaints filed, 2,645 fully adjudicated, 6,160 written state complaints filed and 10,806 mediation requests; due process filings rose from 17,118 in 2011–12, and the share still pending at year end rose from 12 percent to 43 percent over the same decade. A more recent summary covering 2023–24 exists but could not be opened from this session, so only the 2021–22 figures are used here. cadreworks.org (PDF)

Every source above was opened and read for this article. Where a figure could not be confirmed in the original, it is not on this page.

---

About Clay Shumate

Clay Shumate is a certified secondary Social Studies teacher in the public schools of West Alabama, with seven years of classroom experience, a B.A. in History, and an M.Ed. in Secondary Education. He writes about project-based learning, student responsibility, respect, and practical ways to hold young people to a higher standard while giving them room to learn from mistakes.

By Clay Shumate

Teacher burnout shows up in students’ experience of the classroom long before it shows up in their test scores. The largest review to date links it clearly to weaker emotional support, weaker classroom management and lower student motivation — and almost not at all to achievement. The larger measured cost to students is what happens after a teacher leaves.

This page is about outcomes measured on students, and it is the one article in this series written mostly for administrators and families rather than for the teacher carrying the load. For the word itself, start with the definition and the three dimensions behind every number on this page; for what a principal or district can change, the conditions an administrator can actually change and the measured size of each one covers it and this page does not restate it. What follows is narrower: when a teacher is burned out, what actually happens to the kids in the room, and how strong that evidence really is.

Key takeaways

• The effect lands on the classroom experience, not the score. In a 2026 meta-analysis of 86 studies and 38,457 teachers, burnout correlated with weaker emotional support, classroom management and instructional support at r = −.13 to −.30.
• Student motivation is the clearest student-side link, at r = −.23 across ten samples and about 6,100 students. Achievement is the weakest link in the field: r = −.05, with a confidence interval that crosses zero.
• There is little evidence linking teacher burnout to students’ depressive symptoms or emotional distress. The alarming version of this claim is not supported.
• None of those pooled numbers says anything about an individual teacher. A correlation is a statement about tens of thousands of people at once.
• The bigger measured student cost is turnover. In North Carolina middle schools a 10-point rise in turnover tracked about .013 standard deviations lower math scores, and replacement staff were more often inexperienced, lateral-entry or teaching out of field.
• Those effects are largest in the highest-poverty and lowest-performing schools, which makes this an equity question and not only a staffing one.
• Almost none of this is causal. These are correlates — signals about conditions, not proof that one thing produced the other.
• Burnout is an occupational condition, not a diagnosis anyone gives from a web page. Nothing here is a screening instrument or a way to assess an individual teacher.

How does teacher burnout affect students?

Mostly through the quality of the interaction, not the quantity of the learning. When researchers pool the studies, the strongest and most consistent associations are with how the room runs and how supported students feel in it. The weakest association in the whole pile is the one people quote most: test scores.

The best current summary is a 2026 meta-analysis in Review of Educational Research by Gyde Wartenberg, Karen Aldrup, Simon Grund and Uta Klusmann, pooling 86 studies and 90 independent samples covering 38,457 teachers. It reports burnout’s three dimensions separately, which matters, because they do not behave alike.

Depersonalization — the cynical, checked-out dimension — is the one most tied to how students are treated. It correlated with lower emotional support at r = −.27 and with weaker instructional support at r = −.27. Emotional exhaustion tracked with classroom management at r = −.23. Reduced personal accomplishment tracked with instructional support at r = −.30, the largest single correlation in the set.

Those are small to moderate correlations — a real signal, not a large one. A correlation of −.27 does not mean a burned-out teacher stops supporting students; it means the two move together across tens of thousands of people.

One thing has to be said before anybody reads further, particularly any teacher who found this page while having a hard year. None of these numbers is about you. A pooled correlation across 38,457 people describes a tendency in a population. It cannot tell you anything about one classroom, and it is not a verdict on whether you are failing your students. Reading it as one is statistically wrong and it is also the exact self-blame that makes a hard year harder. What follows is an argument about how schools are staffed and scheduled, not evidence against any individual.

Does teacher burnout lower students’ test scores?

The evidence does not support it, and it is the single most repeated claim about teacher burnout and students. Emotional exhaustion and student achievement pooled to r = −.05 across ten independent samples, 95 percent confidence interval −.11 to .01. An interval that crosses zero means the data cannot rule out no relationship at all.

The authors do not treat that as a surprise. They write that “individual student characteristics such as intelligence, prior knowledge, and conscientiousness…determine achievement to a large extent, thereby reducing the potential of teacher characteristics to influence student achievement.” Test scores are a noisy, slow-moving outcome with a lot of other things pushing on them.

The earlier systematic review is a little more generous. Daniel Madigan and Lisa Kim, in the International Journal of Educational Research, went through 14 studies covering 5,311 teachers and 50,616 students — eight primary, six secondary. They found burnout associated with worse achievement “more often than not,” in four of seven studies, at small-to-medium effect sizes. Their verdict on the field is blunt: “It is almost cliché to bemoan the lack of longitudinal and rigorously designed studies.”

So: four of seven studies found something, and the pooled estimate across ten samples is roughly nothing. Anyone telling you burned-out teachers demonstrably lower test scores is reporting half of that.

What changes in the room that a student would actually notice?

Autonomy, patience, and the sense that the teacher is present. This is where the secondary-specific evidence lives.

Bo Shen and colleagues, in the British Journal of Educational Psychology, surveyed 1,302 high school students and their 33 physical education teachers across 20 schools in two demographically similar Midwestern districts, at the start and end of a semester. Teachers’ emotional exhaustion was negatively related to how much autonomy support students perceived. Depersonalization was inversely related to students’ autonomous motivation even after controlling for that perceived autonomy support — so the cynical dimension had a link to motivation running past the obvious channel.

Two labels belong on that: physical education rather than a typical grades 6–12 academic classroom, and correlational over one semester. But it is a high school sample, which most of this literature is not, and the pattern matches the pooled numbers.

What that looks like in practice is not dramatic. Fewer choices offered, because choices take setup. A shorter fuse on the fourth redirect. Feedback that gets thinner in March than it was in September. Students read all of that accurately, and they usually read it as being about them.

One finding here a coach can genuinely use: the largest single correlation in the meta-analysis was between reduced personal accomplishment — the “none of this is working” dimension — and weaker instructional support. That dimension responds to evidence, because the belief underneath it is a factual claim about whether the work is landing, and claims like that can be checked. Making growth visible is a real intervention. It is also not a substitute for fixing a load that is genuinely too heavy, and a coaching cycle offered in place of staffing relief will be read as exactly what it is.

Does it affect how students feel about school?

Motivation, yes. Mental health, there is no good evidence for it.

Motivation is the clearest student-reported finding in the meta-analysis: emotional exhaustion correlated with student motivation at r = −.23, 95 percent CI −.28 to −.17, across ten studies and about 6,108 students. That interval does not cross zero.

The wellbeing claim is a different story, and it is worth saying out loud to any parent who has read a frightening headline. Madigan and Kim looked and reported “little evidence to suggest an association between teacher burnout and students’ depressive symptoms, emotional distress, or attempted suicide.” One study found exhaustion linked to slightly lower school satisfaction, a small effect. That is the whole finding. A teacher having a hard year is not measurably damaging children’s mental health, and claiming otherwise borrows a weight the research does not carry.

Which is also why a teacher’s apparent state is not a subject for the class group chat. Speculating in public about whether a particular adult is struggling is a guess about somebody’s health, it travels faster than any correction, and it does nothing about the conditions that produced the problem.

What happens to students when a teacher leaves?

This is the largest measured student-side cost in the literature, and it has the strongest designs behind it. Burnout is not the only reason teachers leave, so treat it as a downstream consequence rather than a direct burnout effect.

The most relevant study for grades 6–12 is Lucy Sorensen and Helen Ladd’s analysis in AERA Open, covering North Carolina middle schools — grades 6 through 8 — from 1994–95 through 2015–16. A 10 percentage point rise in teacher turnover was associated with scores about .011 standard deviations lower overall: .013 in math, .007 in reading. Small per unit, and they accumulate in buildings where turnover is chronic.

Be honest about how small that is. A hundredth of a standard deviation is close to nothing for any individual child, and citing this study as proof that turnover is devastating students oversells it. The reason to act is not the effect size. It is the second finding, about what the school is left holding.

Turnover has “marked, and lasting, negative consequences for the quality of the instructional staff.” Schools with more churn ended up with higher proportions of teachers with limited experience, more lateral entrants and provisional licenses, and more teachers not certified in the subject they were assigned. And the effect concentrated: the increase in lateral or provisionally licensed staff was 8.7 percentage points larger in the highest-poverty schools and 14.5 points larger in low-performing schools.

An earlier New York City study points the same way. Across roughly 625,000 observations of fourth and fifth graders, Matthew Ronfeldt and colleagues found math scores six to seven percent of a standard deviation lower in a grade-year with complete turnover than in one with none. The finding that matters most: the damage was not limited to students who got a new teacher. In lower-achieving schools, students whose own teacher stayed also did significantly worse when turnover around them was high; in higher-achieving schools there was no such effect. That is grades 4 and 5, so read the pattern, not the number.

Do teacher absences cost students learning?

The best study on this found a real effect — and it is a fourth-grade result in one district. Raegen Miller, Richard Murnane and John Willett examined 6,166 students and 231 teachers across 73 schools in a large urban district over two school years. The first ten days of teacher absence were associated with about 0.15 of a standard deviation lower fourth-grade math achievement.

That is the biggest number on this page, which is exactly why the caveats travel with it. One district, one grade, mathematics only — English Language Arts effects were smaller and inconsistent — leaning on instrumental variables with 16 to 24 percent of prior achievement values imputed. It is also not a burnout study. People are absent for many reasons unrelated to the job.

It matters here for one reason: the meta-analysis found emotional exhaustion associated with absenteeism at r = .18. If exhaustion raises absence a little and absence costs learning a lot, coverage is one of the few places where the student cost of a depleted staff becomes visible and countable. It sits alongside the student-attendance side of the same problem, tracked closely for the same reason.

One wrong turn is worth naming before somebody takes it. A school that answers a coverage number by leaning on staff not to take sick days has reduced nothing except its own visibility into the problem. Absence is a symptom of a staffing plan with no slack. Suppressing it leaves the plan as it was and adds sick adults to a building full of teenagers.

How much of this is actually cause and effect?

Very little of it, and the researchers say so first. This section exists because the honest version of this topic is weaker than the version that circulates.

The meta-analysis authors are explicit that they “refer to teacher absenteeism, teacher–student interactions, and student motivation and achievement as correlates” and that these “do not convey causal relationships.” Of 31 longitudinal studies they identified, only 11 reported longitudinal correlations between earlier burnout and later outcomes, and that evidence was mixed — it supported the motivation finding better than the achievement finding.

There is also a measurement problem that cuts against the strongest numbers. Correlations were stronger when the classroom-quality rating came from the teacher’s own self-report than from an outside observer. A tired teacher rating both their exhaustion and their own classroom management harshly produces a correlation without anything happening in the room. The external-observer estimates were weaker.

And the arrow may run backwards. A class that is genuinely hard to teach can exhaust a teacher, rather than an exhausted teacher making the class hard. Nothing in a cross-sectional correlation can separate those, and both are probably true at once. The practical reading is the uncomfortable one: you should take these findings seriously enough to fix conditions, and not seriously enough to blame a person.

What can a school check, without turning it into a file on a teacher?

Check the conditions, not the people. Everything above points at building-level facts that are already recorded somewhere. None of it requires assessing an individual teacher’s state of mind, and none of it should ever be written down next to a teacher’s name.

Four things are countable, and all four sit with whoever builds the schedule and the budget rather than with the person in the classroom.

• Continuity. How many sections changed teachers mid-year, and how many positions turned over between June and August? Sorensen and Ladd’s finding is about the staff a school ends up with, so watch who arrived, not only who left. Mostly central office — hiring and retention.
• Assignment fit. How many teachers are teaching outside the field they are certified in? That was one of the specific quality losses the North Carolina data documented, and it is already in the staffing file. Shared — and often a supply problem rather than a choice, which makes it a diagnostic rather than an accusation.
• Coverage load. How often is a class covered by someone other than its teacher, and how is that spread across the building? Absence is not a behavior problem to manage down. High coverage says the staffing plan has no slack in it. Building level — a principal can usually see and change this one.
• What students say about the room. Not about the teacher — about the room. Do they get choices in how they show what they know? Do they know what happens next? Climate surveys already ask this, and answers aggregate by course or grade without naming anybody. Building level.

Three lines stay attached to that list. It measures a building, not a person, so never write a teacher’s name on it and never let it feed an evaluation — a name in a box turns a diagnostic into an informal personnel record with none of the protections a real evaluation carries. It is not a burnout screen: no validated cutoff turns any of these numbers into a finding about an individual, and treating a bad number as evidence that a particular teacher is burned out is unsupported and unfair. And if you act on what you find, write the rule for who qualifies before you apply it to anyone, because a private fix for one teacher creates a fairness problem for everyone else in the hallway.

Is student behavior why teachers burn out?

No. This series has answered that four times now and the answer does not change. Managing behavior is real work and the most commonly named stressor in the national survey data, but a class of teenagers behaving like teenagers is the job, not a malfunction. What turns ordinary classroom demand into burnout is how much of it one person absorbs without help, and that is a scheduling and staffing decision. the demand side, and how good the evidence behind each piece of it is goes through that in detail.

The students who need the most redirecting usually have reasons nobody in the room has been told about, and handling the behavior itself without spending the relationship is a separate skill from carrying an unreasonable load. The two problems get confused constantly, and only one of them is fixable inside the classroom.

If you are a student reading this because you are worried about a teacher: you did not cause it, and it is not your job to fix it. Being decent to them is a good thing to do and it is not the solution to a workload problem.

What to do next

If you are an administrator, pull those four numbers for one grade level and look at them together. They are already in your data. If any of them is bad, the fix is a schedule or budget decision, not a conversation about resilience — which decisions those actually are is the companion page.

If you are a parent, the useful question at a conference is not whether a teacher seems tired. It is whether the course has had the same teacher all year, whether it is taught by someone certified in the subject, and how often it is covered by someone else. Those are answerable, and they are about the school’s plan rather than a person.

If you are a teacher and this page is why you are here: the research says the cost of an under-resourced job lands on your relationships with students before it lands on their scores. what getting back on your feet actually involves is the recovery side. Nobody gets credit for absorbing a staffing problem quietly, and students do not benefit from it either.

Frequently asked questions

Sources

• Wartenberg, G., Aldrup, K., Grund, S., & Klusmann, U. (2026). How strongly is teachers’ burnout related to teacher absenteeism, teacher–student interactions, and student motivation and achievement: A meta-analysis and systematic review. Review of Educational Research. 86 studies, 90 independent samples, 38,457 teachers. Pooled correlations used on this page: instructional support −.30 (reduced personal accomplishment), emotional support and instructional support −.27 (depersonalization), classroom management −.26 (depersonalization) and −.23 (emotional exhaustion), student motivation −.23 (95% CI −.28 to −.17, 10 studies, ~6,108 students), absenteeism .18, student achievement −.05 (95% CI −.11 to .01, 10 samples). Limitations stated by the authors: these are correlates and “do not convey causal relationships”; only 11 of 31 longitudinal studies reported longitudinal correlations; self-reported classroom quality produced stronger correlations than external ratings. Egger’s tests indicated no evidence of publication bias. journals.sagepub.com
• Madigan, D. J., & Kim, L. E. (2021). Does teacher burnout affect students? A systematic review of its association with academic achievement and student-reported outcomes. International Journal of Educational Research, 105. 14 studies, 5,311 teachers and 50,616 students; 8 primary settings and 6 secondary; 8 cross-sectional, 4 longitudinal, 2 experimental. Burnout associated with worse achievement in four of seven studies at small-to-medium effect sizes; “little evidence to suggest an association between teacher burnout and students’ depressive symptoms, emotional distress, or attempted suicide.” Read as the accepted manuscript in the White Rose repository. eprints.whiterose.ac.uk (PDF)
• Shen, B., McCaughtry, N., Martin, J., Garn, A., Kulik, N., & Fahlman, M. (2015). The relationship between teacher burnout and student motivation. British Journal of Educational Psychology, 85(4), 519–532. 1,302 high school students and 33 physical education teachers across 20 schools in two demographically similar Midwestern U.S. districts, surveyed at the start and end of a semester and analysed with hierarchical linear modelling. Emotional exhaustion negatively related to perceived teacher autonomy support; depersonalization inversely related to student autonomous motivation controlling for autonomy support. Read via the abstract in the ERIC record; the full text sits behind Wiley and was not accessible, so nothing beyond the abstract is claimed here. eric.ed.gov
• Sorensen, L. C., & Ladd, H. F. (2020). The hidden costs of teacher turnover. AERA Open, 6(1). North Carolina middle schools, grades 6–8, 1994–95 through 2015–16, math and English Language Arts. A 10 percentage point increase in turnover associated with about .011 standard deviations lower scores overall (.013 math, .007 reading); “marked, and lasting, negative consequences for the quality of the instructional staff”; the rise in lateral or provisionally licensed staff was 8.7 percentage points larger in the highest-poverty schools and 14.5 points larger in low-performing schools. Limitations stated by the authors: no value-added measures of teacher quality, and some student characteristics available only in select years. journals.sagepub.com
• Ronfeldt, M., Lankford, H., Loeb, S., & Wyckoff, J. (2011). How teacher turnover harms student achievement (NBER Working Paper No. 17176). Roughly 625,000 observations of 4th and 5th grade students in New York City, 2000–2002 and 2004–2007. Math scores six to seven percent of a standard deviation lower in grade-years with 100 percent turnover than with none; in lower-achieving schools, students whose own teacher stayed also performed significantly worse when turnover was high, while in higher-achieving schools there was no such effect. Grades 4–5, not a secondary result. Later published as Ronfeldt, Loeb & Wyckoff (2013), American Educational Research Journal, 50(1); the working paper is the version read here. nber.org (PDF)
• Miller, R. T., Murnane, R. J., & Willett, J. B. (2006). The impact of teacher absences on student achievement. Harvard Graduate School of Education / NBER Summer Institute. 6,166 fourth-grade students, 231 teachers, 73 schools in one large urban district, school years 2003 and 2004. The first ten days of teacher absence associated with about 0.15 of a standard deviation lower fourth-grade mathematics achievement. Limitations stated by the authors: one district, one grade, mathematics only (English Language Arts effects smaller and inconsistent), reliance on instrumental variables, and 16–24 percent of prior achievement values imputed. Grade 4, and an absence study rather than a burnout study. Later published in Educational Evaluation and Policy Analysis, 30(2). conference.nber.org (PDF)

Every source above was opened and read for this article. Where a figure could not be confirmed in the original, it is not on this page.

---

About Clay Shumate

Clay Shumate is a certified secondary Social Studies teacher in the public schools of West Alabama, with seven years of classroom experience, a B.A. in History, and an M.Ed. in Secondary Education. He writes about project-based learning, student responsibility, respect, and practical ways to hold young people to a higher standard while giving them room to learn from mistakes.