While teachers may have a good sense of what's happening in their own classrooms, their individual experiences cannot paint the broader patterns of what occurs across classrooms in thousands of schools and districts—and certainly not over time. That's why we look to statewide data systems to help us gain perspective. Much like a pointillist painting, the lens a state system can provide allows us all to take a step back and view the bigger picture.
Pursuant to that broader perspective, since 2006 the U.S. Department of Education has awarded a total of $826 million to states and territories to help them develop longitudinal data systems, tracking public school students from early learning through the workforce, including some key information on their teachers. The average award stands around $7.6 million per round over seven rounds of grants, ranging from $3 million to Wyoming and Alabama, to $35 million to Texas.
The Department of Education recently released the results of its survey on where state longitudinal data systems currently stand or whether states might have other data systems in place that serve similar purposes. Forty-seven states responded to the survey; New Mexico (which has not yet received a grant to develop its system), New York, Ohio, and Virginia did not.
While most states have made some progress over the last 15 years, key data connections that make these systems useful are missing in many states, some of which were the focus of the largest grants awarded over a decade ago.
Almost all of the states that responded to the survey do collect information on student characteristics such as demographics, homelessness status, attendance, grade level, course enrollment, student dropout, and student transfers. Also, most states also collect information on teachers, such as certification path, salary, and years of experience.
Much less prevalent, however, are states that connect these students with teachers. While two-thirds of states claim to be able to connect some statewide teacher data with student data, most still lack the essential student-teacher connections needed to inform decision making.
For example, while about 60% of states have the data links that allow them to know which teachers are teaching which students (via course assignment, roster verification, or unique teacher identifiers linked to student identifiers), for most it is not yet possible to connect basic teacher attributes with student data, such as a teacher's certification path or whether she's teaching out of field in connection with, for example, student outcomes.
Less than 40% of states indicate that they can connect student data to their teachers' preparation program information, and therefore use student test data to establish whether programs are generally producing more or less effective teachers. This kind of data is much sought after from school districts.
Only 14 states are able to report having statewide systems that directly link their teachers' evaluation ratings to the students they teach, even though 34 states mandate that teacher evaluations employ some measure of student growth. In attempting to square these competing claims, it could be that in the 20 states that lack the systems for direct linkage, school districts are using more aggregate measures of student growth (e.g., school-level growth rather than classroom-level growth) in teacher evaluation or that districts are complying with the law but not reporting their data to the state, which would indicate an abnormally high degree of latitude not typically granted by states. Regardless, this lack of data poses limitations on a state's ability to analyze their evaluation systems or design reforms in connection with student achievement.