Thursday night's negotiations didn't lead to a deal in the Chicago teachers' strike but it appears that progress is being made. According to the Chicago Tribune's Bridget Doyle, Jennifer Delgado and Joel Hood, the union and the school district are moving on from discussions of teacher evaluations and pay increases, though there remain differences on those matters, but "recall rights" remain a sticking point. So what are recall rights, and why do they matter?

Recalls occur when a school district that has laid off teachers needs to hire again and, under the terms of the union contract, is obligated to fill the new opening with recently laid off teachers before looking elsewhere. Like the initial layoffs, recalls are usually based on seniority. Less experienced teachers are laid off first, then more experienced teachers, and recalls work in reverse: more senior teachers get rehired first, then less senior ones thereafter. This policy is frequently known as "last in, first out." The intent is to maintain job security for teachers in the face of staffing changes that are beyond their control.

As a general policy, "last in, first out" (or LIFO) has generally negative effects on student achievement. A study by Dan Goldhaber and Roddy Theobald at the University of Washington found that under Washington State's LIFO rule, laid off teachers were only slightly worse than average, whereas if an effectiveness measure were used instead, the layoffs would have been concentrated on poor teachers:

Source: Dan Goldhaber and Roddy Theobald

An Urban Institute study (pdf) by Donald J. Boyd, Hamilton Lankford, Susanna Loeb, and James H. Wyckoff found similar results. That paper estimated that the gains due to using effectiveness-based, rather than seniority-based, layoffs improved teachers' performance by the same amount as is gained when one replaces a teacher with one year of experience with a teacher with five years. So even though seniority is, all else being equal, correlated with teacher quality, using effectiveness measure is a better way to get good teachers.

A lot of this effect is due to the fact that effectiveness-based layoffs lead to fewer layoffs than seniority-based ones. Seniority-based layoffs hit inexperienced, poorly paid teachers more, and thus you need to lay off more teachers to save the same amount of money. Because effectiveness-based layoffs also affect experienced teachers who make more money, districts can save the same amount while firing fewer teachers. This is especially true in poor and minority districts, where Cristina Sepe and Marguerite Roza estimate more teachers are fired, and thus class sizes increase more, due to seniority-based layoff policies than in districts that serve whiter, more affluent populations.

But what about recall policies specifically? I took a look at the National Council for Teacher Quality's excellent TR3 database of the nation's 100 biggest school districts to find out. The median school district has mandatory teacher recalls after layoffs, with the median district allowing recalls up to two years after the initial layoffs. Some, like Miami-Dade County, Pittsburgh, Dayton, and Toledo, mandate recalls indefinitely, while others, like North Carolina school districts, the Atlanta suburbs and, interestingly, Chicago, do not have recalls at all. Unfortunately, we only have good NAEP score data for certain school districts, so a rigorous evaluation of how recall policies affect performance is difficult, but here's how math scores differ based on the length of recall periods (I represented indefinite periods as 10 years):

And here's how reading scores differ:

The charts suggest that generous recall policies hurt achievement, but the evidence is extremely weak. It's a small sample size, and there are many other factors, like poverty, racial composition class sizes, other hiring/firing policies and so forth, that are more important factors in determining student achievement, and which recall policies a district adopts may depend on those factors. Most importantly, the effects aren't statistically significant. The estimate effects are still negative once you control for the percentage of students getting school lunches, teacher salaries, and the percentage of black or Latino students, each additional year where recalls are mandatory is associated with a 0.1832 point decrease in 8th grade reading scores and a 0.03751 point decrease in 8th grade math scores. But the policies actually increase 4th grade math and reading scores once you control for these other factors. So there's no reason to believe, from the NAEP data, that recall policies have any effect whatsoever on student achievement.

The latest Chicago Public Schools proposal would strike a compromise on the issue. It would create two pools, one for high-achieving teachers and the other for low-achieving teachers, and would automatically rehire high-achieving teachers who have been laid off if openings arise, whereas low-achieving teachers would have to reapply for their positions. The research on last-in, first-out suggest that this will probably be better for student learning than the alternative. But the data on recall policies themselves is decidedly ambiguous.