/

Japan should treat test scores with discretion

by Walt Gardner

Special To The Japan Times

The education ministry’s decision to allow municipal boards of education to make public the results of achievement tests for individual schools is defended as an indispensable part of the accountability process. The policy has great intuitive appeal to those who are frustrated by what they perceive as a lowering of standards.

But there’s another side to the story. Although standardized tests provide one piece of information in determining how schools are performing, they are far too unstable to be considered reliable or fair, according to a briefing paper by the Economic Policy Institute. That’s because factors beyond the control of teachers play an inordinate role. These include such things as parents’ education, the home literacy environment and the influence of neighborhood peers.

Schools’ test scores largely reflect the backgrounds of the students who enroll. Data certainly matter, but it’s how the data are used that is the problem.

Contrary to popular opinion, Japan’s proposed new policy places it somewhere in the middle of the educational pack internationally.

Finland, which is widely acknowledged to have the world’s best schools, uses standardized test scores strictly for diagnostic purposes, and never makes the results public. It selects about 100 schools each year for testing in order to determine systemic weaknesses that national policy leaders should address and municipalities can consider for whatever reasons they choose.

Finland does not use test scores for naming and shaming schools, or worse for naming and shaming teachers. Comparisons between schools are frowned on as a counterproductive strategy that undermines morale at a time when teachers are already beleaguered.

The United States is moving in the opposite direction. Many districts not only publish test scores for schools but also for individual teachers.

Los Angeles, home of the nation’s second-largest school district, began releasing performance reports after the Los Angeles Times in August 2010 published a database of some 6,000 third-through fifth-grade teachers ranked in part on their students’ test scores.

Then in February 2012, New York City was cleared to do the same for thousands of teachers after a state court declined to hear a final appeal from the teachers’ union to keep the information private. The reports covered about 12,500 teachers of math or English in fourth through eighth grade.

It’s inevitable that the greater the emphasis on using test scores as the final factor in evaluating schools, the greater the chance that wrongdoing will occur. Atlanta was the scene of the biggest cheating scandal in American history involving half its elementary and middle schools. At least 178 educators — principals, teachers and other staff members — took part in widespread test-tampering. That’s not surprising because of Campbell’s Law. The more any quantitative indicator is used for decision-making, the more it will be subject to corruption, and the more it will corrupt the process it is intended to monitor.

Relying on test scores as the overwhelming factor will also eventually lead to school closures. That happened in Chicago, when 50 schools were shuttered at the beginning of the present school year, despite parents’ protests.

If school test scores become headline news in Japan, parents could be tempted to begin a movement to opt out of nationwide standardized testing. Many parents in Colorado, Connecticut and New York are already pulling their children from participating, in the belief that the obsession with testing is destroying educational quality.

Japan has the advantage of being able to learn from the experiences in Finland and the U.S. It doesn’t have to commit itself to one extreme or the other. Assessment is an indispensable part of the learning process. Without it, all stakeholders are shortchanged.

In the final analysis, the challenge is to design better tests and use the results properly.

Walt Gardner writes the Reality Check blog for Education Week in the U.S.