Smart Use of Benchmark Tests is One Way to Mend, but Not End School Accountability

RCEd Commentary

One inescapable reality from the era of school accountability, whose roots go back at least to the 1980s, is that students were improving - they began doing better on tests in core subjects like reading and math. Long-term trend data on the National Assessment of Educational Progress shows that African American and Hispanic nine-year olds gained the equivalent of one-and-a-half to two grade levels in reading from 1999 to 2008.

Another reality is that parents and educators have pushed back against "too much testing." Part of the response is against benchmark testing. Schools administer them monthly, quarterly or otherwise to see if students are on track to meet standards set by their district or state. The exams are usually shorter than state achievement tests, often created by districts and sometimes cover material not already taught in class. The latter element frustrates students and parents.

Interestingly, some administrators and principals are now trying to deal with both realities. They are using tools to improve accountability, make it less burdensome and yet be serious about results. They are entrepreneurial in their strategies, even using incentives for educators to show their instruction is tied to the state's standards. In short, they are being smart about better use of tests and strong teaching.

Let me offer three examples from Texas, one of the handful of states that pioneered school accountability.

In Houston, Superintendent Terry Grier has stopped requiring benchmark tests for campuses ranked in the top 20 percent of Houston's elementary, middle or high schools. The high-performers still can use such exams, but they also can use the time for more instruction.

Low-performing schools, however, must give benchmark exams. Yet they are no longer given quarterly. Grier found quarterly tests spread too far apart to lead to quick interventions.

Instead, the benchmarks are administered every two to four weeks. Schools use the results to intervene more immediately with students, including through one-on-one tutoring.

Grier sees benchmark tests as a beginning, not an end. Schools can use them to set out a map of activities that will help students learn their material.

There is wisdom in this management strategy. Grier essentially is saying we can have less testing where instructors are teaching well. First, though, show me results. He wants to know if students are on track, and testing is a valuable way to do that.

Indeed, ample research shows the right kind of testing can help students understand course content. Art Graesser, a University of Memphis psychology professor, studies how sustained engagement in a subject can lead to deeper learning. (Graesser, founding co-director of the Institute for Intelligent Systems at his university, has served as a Bush Institute education consultant.)

Graesser and a team have found that testing can help learning, especially when aligned with key content. Regularly administered tests also help more than a single exam with long-term retention. 

In effect, the Dallas school district is using that latter finding in spacing out its benchmark tests. The district administers its exams twice annually.

Campuses use data from the mid-year exam to adjust instruction. The district uses end-of-year scores to adjust larger strategies, such as improving writing performance. "Other than these tests," Superintendent Mike Miles told me in an interview, "we don't dictate others."

You can find innovative strategies at the campus level, too. Don Snook, a Texas principal I spoke with recently, said he has dropped quarterly district tests that his suburban Dallas high school was supposed to take. He decided to stay on top of his school's progress another way.

In return for dropping the benchmarks, he regularly checks in on his 22 teachers. Last year, he made 150 classroom visits to observe instruction. He also meets weekly with teachers, reviewing lesson plans and discussing strategies. Similarly, he looks at "bell-ringer" tests teachers give when class begins to see if students grasp the latest content.

Snook intervenes if a teacher's instruction does not keep up with what Texas expects students to learn. If a class is behind in, say, writing, he works with the teacher on better strategies for incorporating writing.

None of this really matters if academic results don't flow from these strategies. We still don't know how these efforts will play out, although Snook's school scored 90 percent or better on Texas' core tests this year.

Yet these examples do stand out for this reason: Creative leadership is finding a way to significantly reduce testing and yet more strategically use exams. This is about maintaining accountability and at the same time adjusting strategies.

One major strategy is focusing intensely on classroom instruction. Grier and Miles especially are known in Texas school circles for spending their political capital on improving instruction. If critics really want fewer tests, insisting on better instruction is one way to get there. Students still will need some tests, for their own good. But this is one way to mend but not end accountability.