As more and more states push legislation tying teacher evaluations to student achievement—a policy incentivized by the federal Race to the Top program—many are scrambling to put data systems in place that can accurately connect teachers to their students. But in a world of student mobility, teacher re-assignments, co-teaching, and multiple service providers, determining the roster of students to attribute to a teacher is more complicated than it may sound.
At a conference in Washington on Thursday, representatives from the state departments of education in New York and Louisiana presented two very different systems for linking teachers and students for data collection. Together, the proposed systems highlight the challenges involved in this process.
The event was hosted by the Washington-based Data Quality Campaign, a nonprofit organization working to help states collect and use longitudinal education data. (Both DQC and Education Week‘s nonprofit parent company have received funding from the Bill and Melinda Gates Foundation.)
Paige Kowalski, director of State Policy Initiatives for DQC kicked off the panels by explaining that the teacher-student data link is the “linchpin” in teacher-effectiveness policies, on which 30 states have passed laws this year alone. One major aspect of the data link that states and districts are struggling with, however, is how to define a student’s “teacher of record.” A common definition is “the teacher who assigns grades,” but student-teacher relationships are much more complex than that, Kowalski said. And with high-stakes decisions riding on these determinations, the cost of misattribution can be great.
Panelist Molly Horstman, assistant director in the human capital office for the Louisiana Department of Education, said her state will begin to use a roster-verification system for 2012. Teachers will access their class rosters through a secure log-in and add or delete students as appropriate.
In an interview after the panel, Horstman explained that the teacher’s verification of the class list will be done once a year, just before testing, and that the principal will conduct a second round of verifying after the teachers make their changes. Both teachers and administrators will be trained on how to use the system, which was developed with input from the teachers’ unions.
New York, on the other hand, plans to use a much more nuanced process, according to Ken Wagner, assistant commissioner for data systems for the New York State Education Department, who was also on the panel.
“When you say a student was in a classroom for the year, that begs the question: What portion of the year?” Wagner said. He explained that, in his state, a student with a test score “will count in proportion to the time the student was with that teacher.” Teachers will track the duration of a student’s enrollment and attendance, not by days but by minutes.
Wagner showed an intricate flow chart with the details of the daily tracking and process for audits. Roster verification at the end of the school year “will only lead to lawsuits,” he contended. “If you think this is too complicated, put yourselves in the shoes of a teacher. Some say it’s not complicated enough.”
Tracking Back to Teacher-Prep
A separate panel addressed the idea of linking teachers to student data in order to rate teacher-preparation programs. Sandi Jacobs, vice president of the National Council on Teacher Quality, explained that 99 percent of teacher-preparation programs are rated satisfactory, according to states’ reporting to the federal government.
When the New Teacher Project’s The Widget Effect report came out in 2009, finding that nearly all teachers received satisfactory ratings, “that galvanized people,” Jacobs said. “We have a similar situation with teacher preparation. We need to take actions against the weakest programs.” She praised Louisiana, Tennessee, Florida, and North Carolina for their efforts in that realm.
Jane West, vice president of policy, programs, and professional issues for the American Association of Colleges of Teacher Education, stressed that while there’s a need to track the performance of teacher-education graduates, “we have a long way to go” before the data can be considered reliable.
Teachers who leave the state, teach out-of-field, or move to private schools are nearly impossible to track, she said. And teachers in non-tested subjects and grades are out of the mix as well. Last year, the University of Central Florida was only able to get student-achievement data for 12 percent of its graduating class, yet that information was reported publicly. “What’s the threshold?” West asked. “Where’s the check to ensure that’s a valid and reliable measure? It needs to be more than 12 percent.”
In all, the Data Quality Campaign’s conference was tightly managed and left little opportunity for audience participation, offering attendees a controlled (though still controversial) takeaway: that improved student achievement hinges on improving the teacher-student data link.