American Tests Fail the Test

One of the top two reasons we don't have the schools we need is the way we test kids (the other is work rules for adults, but that's subject for another day).

There are two big problems with our tests--we test the wrong thing in the wrong way.

State tests reflect state standards--we have 50 of them and they are all different. But they do have this in common--they are all too broad, vague, and in some respects they expect to little. Fortunately, there's strong momentum for national standards--fewer, clearer, higher standards. It will be easier to share curriculum and compare results across state lines and it will encourage investment in innovative content like learning games and smart online tests that quickly zero in on learning levels.

While we're adopting national standards and locking in higher graduation requirements, it would be a great time to rethink college entrance requirements especially in math where the historical path to Calculus reigns supreme. All young people should learning algebraic problem solving but I'd gladly swap statistics and probability for factoring polynomials. I use statistics daily but haven't factored a polynomial (other than helping my daughters slog through high school) in 30 years.

In addition to testing the wrong stuff, we test the wrong way. We use one cheap end of year test to diagnosis individual problems, to certify student progress, and to hold schools accountable. It's asking way too much of an outdated testing system.

We need measurement to drive improvement, but we could be much smarter about how we measure learning. Computer games are a great example of background assessment with instant feedback. EdSector reports that, "Other fields, such as military training and medical education, are already using technology-enabled assessment to enhance teaching and learning." Adaptive games and online diagnostics provide fast, reliable, and inexpensive measurement that should replace the bubble sheet paper and pencil exams. Broadband and computer access are no longer an excuse not to make the switch.

I'd like to see kids write more--at least one rubric-scored research paper, an op-ed, and new media presentation each semester of high school. Writing prompts that encourage science and social studies teachers to participate would encourage more cross-disciplinary work (and spread the assessment work load). Computer essay scoring continues to improve and will also help avoid the "how do I grade 150 exams this weekend?" problem.

Students should have the opportunity to go deep and become an expert. They should have the opportunity to present what they've learned. How about a Science Fair every other year from 4th to 12thgrade? We know how to reliably score project-based work but we're too lazy and too cheap to do it. If you want to see good student portfolios, visit place where kids show what they know.

The options for fitting all of this together get kind of wonky. The point is that schools, districts, and charter networks need the opportunity to show how new assessments can improve learning, reliably allow students to show what they know, and verify that they're doing their job. The battle to reauthorize the Elementary and Secondary Education Act (what Bush called No Child Left Behind) is an opportunity to encourage innovation in measurement. Like Virginia, states should move most of their testing online with opportunities for innovative local pilot programs.

We need standards and measures but they don't need to be as intrusive as they are today. We need to teach students to read, write, compute, and solve problems, but we can embed it or complement it with opportunities to discover, to go deep, and to apply learning. With ubiquitous broadband, cheap netbooks, and $5 billion in federal incentives, it's time to make progress on testing.