MENU
May 20, 2010

Taking Attendance with Teachers

New research shows just how much teacher absences hurt student achievement.

Like everyone else, teachers get sick, take personal days and must sometimes be away from work for extended medical leave. But until recently, even as schools face continued scrutiny from parents, taxpayers and legislators over performance, there was little evidence on how much teacher absences affect student achievement.

Professor Jonah Rockoff, whose special agreement with New York City schools gives him access to rich data that he uses to tease out insights on teacher performance, efficacy and student achievement, worked with doctoral student Mariesa Herrmann of Columbia University’s Department of Economics to fill that gap. The researchers examined detailed human resources files for full-time teachers in New York City public schools spanning all of the system’s third- through eighth-grade classrooms over eight school years to assess if and how teacher absences harm student achievement. They also had access to student attendance and suspension records and test scores.

The combined data afforded a twofold opportunity to look at the impact of teacher absences in schools while considering a universal question in labor economics: how costly is it — that is, how much productivity is lost — when a firm must bring in short-term substitutes when workers experience health or other personal shocks?

“In teaching, the person with whom an absent teacher is replaced is clearly worse in a substantial way,” Rockoff says. “Substitute teachers are less likely to be highly skilled, since otherwise the chances are she would already have found a full-time teaching job. Even if a substitute is highly skilled, there is a start-up cost: just because someone has a degree in math doesn’t mean she can hop in and be a great sixth-grade math teacher.”

The researchers weren’t surprised to find negative effects, Rockoff says. “But we did not expect to find that the effects were as large as they are. When a teacher takes an extended medical leave, it causes a drop in math and English test scores on par with putting a rookie teacher in the place of a teacher with four years of experience for an entire school year.”

What’s more, it’s clear that teacher absences rather than other variables are responsible for these effects. In addition to controlling for past student achievement and other student characteristics, including absences and suspensions, the researchers were able to compare the same teacher in different years and found that students do worse in years when their teacher takes more time off. Their only remaining concern, says Rockoff, was that “Mr. Smith may be more motivated in some years than in others, and in years when he took more time off he was less motivated.” To address that issue, they took advantage of the fact that student exams are taken in early spring, and compared absences occurring before exams (which should hurt exam scores) with those occurring after (which should not). Absences taken before exams always impaired test scores, while, in almost every case, absences taken after exams were unrelated to exam performance.

There was only one exception: absences taken after the exam due to serious medical illness had a small but statistically significant negative relationship with math scores. How could an absence taken after an exam affect student performance? The researchers believe these absences point to teacher health issues in prior months. “A teacher who had to see a doctor for a serious illness in the month after the exam might already have had a serious medical problem before the exam that hampered his teaching,” Rockoff says, “even if it didn’t cause him to take additional days off.”

Somewhat surprisingly, shorter absences have far more detrimental effects on a per-day basis than longer absences: 10 single-day absences spread out over a year will be significantly worse for student achievement than a single two-week medical leave.

Rockoff points out that short-term absences are typically last-minute and as such don’t provide an opportunity for preparation, whereas longer absences are often foreseeable. “Principals can compensate by taking care to identify a higher-performing long-term substitute teacher,” he suggests. “And, if the teacher cares about what happens when he comes back, he might make sure his temporary replacement is well-prepared with lesson plans.” But the researchers still found large detrimental effects for long-term absences.

These findings reinforce the necessity of sending additional resources into an affected sector when skilled workers are expected to be absent for long periods. More specifically, for a principal faced with the long-term absence of a teacher, that means thinking about how to inject additional resources into that classroom to compensate for the fact that students get a less effective instructor for a period of time. “It might be worth investing more resources in maintaining a staff of highly skilled teachers trained to deal with the specific challenges of substitute teaching,” Rockoff suggests, “or implementing a common curriculum across schools that would make it easier for well-trained substitutes to teach effectively.”

One commonly proposed solution is to give staff incentives not to miss work, but Rockoff insists it is not clear that this is a good policy response, especially in the case of schools. “Presenteeism — where I’m there but I’m not really there — is not a solution for absenteeism. Do we really want someone teaching when she is sick and not giving it her all?” he asks. “Can we improve worker health so teachers take fewer days off and are more productive? Or can we provide incentives for teachers to be more fully productive when they are at work?

“Principals and school districts should make sure policies reduce the root causes of absences.”

Jonah Rockoff is the Sidney Taurel Associate Professor of Business in the Finance and Economics Division at Columbia Business School.

Read the Research

Mariesa Herrmann, Jonah Rockoff

"Worker Absence and Productivity: Evidence from Teaching"

Download PDF 
View abstract/citation