Connecting teachers to their students—and vice versa—remains one of the thorniest problems for state longitudinal data systems, both technically and politically, and more states and districts seem to be trying to get teachers invested in the process.
According to the Data Quality Campaign's most recent analysis, 17 states are unable to match individual teachers to students. In part, this is because while all states provide a unique identification to each teacher, many either cannot legally or practically link those identification numbers to student identifiers.
The DQC also found that states and districts often use educator codes inconsistently, sometimes assigning educator codes to non-instructional staff, and more than half of states cannot link more than one teacher to a student for a given course, making co-teaching and integrated services for English-language learners and special education students harder to track.
Patricia Sullivan, who chairs the National Education Statistics Agenda Committee, announced at last week's National Center for Education Statistics data forum that the committee would launch a new working group dedicated to the issue.
"The forum must call for deeper study of linking teacher-student data," she said.
Building the understanding of people using the teacher-student links will be just as important as building technical systems, according to Hella Bel Hadj Amor, director of teacher effectiveness research and evaluation for the District of Columbia public schools. She said the teacher-student data link has been one of the most challenging parts of implementing the district's new teacher evaluation system, which calculates individual students' growth to gauge whether a teacher qualifies for bonuses or needs professional development.
The district has been conducting data training for schools and teachers on how to ensure teachers get credit for each student, particularly in older grades when a child may have several instructors during the course of the day.
"You have to take these technical concepts and translate them to a layperson's terms. It's very important to get this right; we have found ourselves accused of lying when we meant to simplify," she said.
Officials also hope that by conveying the complexity of the system to teachers, they can prevent adult attempts to cheat in order to reap bonuses. "We try to say, by trying to game the system, you may end up hurting yourself. Are you sure you know what a low-growth student is?" in the context of the merit pay system, Amor explained.
"It's very important to let [schools] know what the information is for and what the stakes are," Amor said. "A mistake can cost a teacher their job; you can't get more high-stakes than that."
In her own state of Texas, Sullivan noted that the University of Texas at Austin has developed a teacher effectiveness model, but state law does not allow teachers to be connected individually to their classrooms; all data must be aggregated. Many districts do not track when a student changes teachers mid-year, leaving only the last teacher as recorded, even if he or she taught the child relatively briefly. The state has now developed a tool to allow teachers to see and analyze their students' longitudinal data privately, though it will not be used by the public.
Kentucky has taken a different tack. Barbara Q. Shoemaker, K-12 liaison for the University of Kentucky's Partnership for Math and Science Education, launched a pilot program of nine Appalachian universities and surrounding school districts to train teachers to use their own students' data to plan instruction and professional development.
"The teachers were looking at the reports on the data, but they were just looking at changes to the data; they weren't really delving into the data," she explained. "It was easy for them to just look at the deficiencies."
Instead, the program taught teachers to look for trends in performance problems across students' academic careers and change instruction to catch problems earlier. For example, elementary and middle school teachers noticed that about a third of high school science students earning "D" grades in high school had been at the same struggling level in middle school, and they have started to look for ways to identify and motivate these students in lower grades.
"Teachers became focused on what was happening in the overall math and science program, not just their own grades and scores," she said.