Performance trumps seniority when it comes to deciding teacher layoffs — if the goal is to minimize negative effects on student learning.
A new study by Brown University professor Matthew Kraft finds that laying off teachers based on how long they have been in the classroom rather than how effective they are hurts student achievement in math. It’s the first evidence based on actual layoffs that the policy known as LIFO, or “last-in-first-out,” has harmful consequences for students.
“Performance measures clearly dominate seniority as a measure of the impact of laying off a teacher on students’ academic achievement,” Kraft told The Seventy Four. His paper is available here but is not set to formally publish online until next month in the academic journal “Education Finance and Policy.” It will appear in print in October.
Kraft is a former public high school teacher in Berkeley, Calif. who served on the executive board of his local teachers union. The LIFO issue has been intensely debated between reformers and union proponents, and is currently the subject of litigation throughout the country, including in California and New York.
Campbell Brown, The Seventy Four’s co-founder, has been involved in supporting litigation in New York challenging the constitutionality of LIFO, as well as the teacher dismissal process.
Kraft finds that layoffs caused small decreases in student achievement on average.1 But the averages mask large variation depending on what type of teachers were let go. In cases where a low-performing teacher was laid off, test scores actually went up in that grade, relative to others in the same school. But when a good teacher was dismissed,2 achievement dropped significantly.3 The differences were fairly large — between two and fourth months of learning, according to a rough approximation.
Kraft shows that teacher seniority, on the other hand, was not a significant predictor of how layoffs affected students.
Kraft’s study examines Charlotte–Mecklenburg Schools, a large district in North Carolina, which laid off teachers in the midst of the 2009–10 economic downturn. Unlike some districts, Charlotte–Mecklenburg had discretion to take job performance into consideration — as measured by principal evaluations — when making layoffs.
The district also looked at a variety of other factors, including seniority, whether and what staff members were licensed to teach, and subjects taught.
This created a natural experiment of sorts, meaning Kraft could zero in on the causal effects of different types of layoffs. To do so, the study compares student gains in math test scores between grade levels where a teacher was laid off versus grades at the same school where no teacher was dismissed.
Kraft simulates the impact of different layoff policies, again using math test scores to measure. He finds that basing decisions on a combination of principal evaluations and estimates of teachers’ impact on student test score gains maximized student achievement relative to alternatives. What this means is that using both measures of teacher performance — student test scores4 and principal evaluations — was preferable to using either by itself or simply using seniority.
Not having seniority be the cardinal rule means fewer teachers lose their jobs since more experienced teachers tend to be paid more. Older, more highly paid teachers are protected by LIFO when budget cuts have to be made. Proponents argue it also prevents favoritism from playing a central role in firing decisions.
“Laying off teachers based on their seniority in the district, rather than their performance in the classroom, results in greater job losses and exacerbates the negative effects of layoffs on student achievement,” Kraft writes. “Layoff policies that do not incorporate increasingly available measures of teacher effectiveness fail to consider among the best information available when making high-stakes decisions.”
In the paper, Kraft focuses on math test scores, but when he runs the numbers using reading scores he finds that only a teacher’s impact on test scores was a significant predictor of the effects of layoffs — seniority and how teachers scored on principal observations weren’t.
Kraft’s findings are not surprising in many ways; they largely mirror past simulation-based research that estimated effects of different layoff policies. They also mirror the common sense argument that ignoring quality when making personnel decisions is unwise. What makes this paper unique is that it looks at the actual — not just the simulated — effects of teacher layoffs.
LIFO supporters may point out that Kraft measures the effectiveness of layoff policies based on student test scores, which some argue are not good indicators of student learning. Kraft agrees that divining teacher effectiveness from test score gains is far from perfect, but “ignoring it when this decision [of which teachers to layoff] has to be made, is misguided.”
Research also shows that student test scores are tied to adult outcomes like income and college attendance.
This study only looks at LIFO’s short-run effects on student achievement, but that might understate its broader harm. For example, Kraft says, LIFO may be “a real disincentive to new teaching candidates to enter the profession, where regardless of their performance, they could be dismissed — so efforts by districts to invest in recruiting high-quality teachers can be really undercut by this.”
Interestingly, in his past research, Kraft found that teachers do get better with more time on the job — that is, experience matters. Yet his work here shows that using teacher performance is a much better way to minimize the negative effects of layoffs on students than teacher seniority.
For Kraft, that means schools should care about experience, but inflexible LIFO policies are “ultimately harmful for kids.”
1. Kraft notes that this may be a conservative estimate of the negative effects of layoffs because it does not take into account school- or district-wide harms of layoffs. (Return to story)
2. Technically, there’s a difference between” layoffs” (usually meaning teachers lose their jobs and are not replaced) and “dismissals” (usually referring to teachers who lose their jobs for performance reasons and are replaced). Here, the two terms are used synonymously to refer to teachers who lost their jobs for budgetary reasons and were not replaced. (Return to story)
3. Kraft did a number of statistical tests to confirm that there were no systematic differences across grades with a teacher laid off versus with no teacher laid off. He found that there wasn’t, meaning any differences in achievement between the grades could likely be attributed to the layoffs. (Return to story)
4. Notably, however, that these test scores measures, often known as “ value-added” only exist for a fraction of teachers, usually those who teach language arts and/or math in grades 4–8. (Return to story)