Friday, September 5, 2008

Where Advocating Accountability Falls Short

Calling for greater accountability continues to be a theme in American education policy. Recently, Senator Barack Obama made this proposition: “I’ll recruit an army of new teachers, and pay them higher salaries, and give them more support. And in exchange, I’ll ask for higher standards and more accountability” (August 28, 2008).

Although the details of policy positions are not generally provided in political speeches, this one is worth pulling apart to see what might be the issues in implementing such a policy.

First, what is the accountability that there will be more of? In this case we may presume that, since accountability is linked specifically to teacher salaries, teachers will be held accountable. Is this appropriate—or even possible—as federal policy? The educational enterprise can be held accountable at many levels. While teachers have face-to-face contact with students who may do well or poorly, the team of teachers working at a grade level or in a small school could be collectively accountable. Moving up a level, a principal could be held accountable for the school’s results. And the district superintendent and the state schools chief can also be accountable for results in their jurisdictions. From purely an accountability point of view, teachers are not necessarily the best focus for federal policy. Certainly, recruiting and incentive efforts can be federally funded, but it seems at best awkward to legislate sanctions for individual teachers based on holding them accountable for their individual performance in raising their students’ scores.

It is currently possible technically to hold teachers accountable. Database and statistical technologies are now available to link teacher identities to student records. District data systems routinely provide links between teachers and their rosters of students and, in many cases, these are extended longitudinally. Many state data systems have also begun providing unique teacher IDs so that linkages to the achievement of individual students can be tracked. And drawing on these longitudinal linkages, “value-added” analyses are being used to quantify the contribution over time of individual teachers to students in their classes.

However, as an approach to federal policy, taking a top-down tactic—making superintendents and principals accountable—may be better than promoting technologies that attempt to measure individual teachers. (The technical controversies about the statistics used in some versions of value-added analysis are a noteworthy topic that we’ll save for another day.) A more productive approach may be to focus on the disincentives for teachers to collaborate or help one another when accountability is at the individual level. We find that many teachers report teaching students other than those officially registered in their classes. Frequently these are informal arrangements that can increase aggregate achievement for the students involved but muddy the district or state records for individual teachers. A school may be a more appropriate unit of accountability and, in that local context, data on individual teachers can more accurately be evaluated. The school principals will know both their schools’ standing among other schools in the district and will have school-level data to help in making staffing decisions. The central office staff, in turn, will have a broader view of the progress of the individual schools on which to base decisions about allocation of resources.

For any achievement-based accountability approach—whether at the district, school, or teacher level—it is important to understand achievement in relation to challenges related to, for example, the economic status of the district or neighborhood or the prior preparation of the students in the classroom. We must also consider the growth of the students, not just their proficiency status at the end of the year. These considerations require statistical calculations, not just counting up percentages of proficient students. And once we begin looking at analyses such as trajectories of some schools compared to others facing similar challenges, we can take the next step and begin tracking the success of interventions, professional development programs, and other local policies aimed at addressing areas of weakness or supporting teachers who are not helping their students make the kind of progress the school is looking for. The data systems and the analytic tools needed to track a teacher’s or a school’s progress over time can also be turned to guiding resource allocations and interventions and, as a next logical step, providing the capability of tracking whether the additional interventions, support, or professional development are having the desired impact.

Given the capacities of current data systems, how might a policy involving greater support and greater accountability for teachers be implemented? Here is one example. Federal funding to districts for principal leadership training could be tied to district-level labor contracts giving the building leaders greater control over personnel decisions. The leadership training will include the interpretation and use of longitudinal data, constituting tools for comparing the principal’s school to others facing similar challenges. Professional learning communities for the principals can be part of this leadership program, assisting district teams to work through ideas for interventions. While the achievement-based accountability measures of teacher performance can be used as one of the factors in building-level decisions, the leadership training would include how to use the data systems for tracking both teacher job performance and the impact of support and training on that performance. —DN