Test Season

This spring, along with the dogwoods and the pollen, Mecklenburg County has been hit by waves of standardized tests. It's the season of pay for performance, the latest fad in a long line of efforts to change American education from the top down.

State standardized tests already consume a good chunk of the school year, when you count the weeks of focused preparation, the drawn-out process of making sure that every child is tested and possibly retested, and then the time it takes for teachers, students and schools to recover from the frenzy.

And testing affects the entire school, not just the tested grade. Art teachers, music teachers, ESL teachers, EC teachers are pulled from their classes to proctor. Media centers close. Volunteer tutors take the week off. The whole school tiptoes around, in order not to disturb the testing students. It's the most frustrating part of teaching in a public school these days, or sending your child to one.

Now, CMS is ramping up this process, adding dozens of new local tests. At one CMS elementary, of the thirteen school weeks between March 7 and June 10, precisely five will be free of state or district mandated tests. How many times can you weigh a pig?

How did this happen? Let's take it step by step.

In national debates over how to improve American schools, a number of powerful players, most notably private entities such as the Gates and Broad Foundations, have argued that raising "teacher quality" should be the central focus of education policy. The U.S. Department of Education, alas, has gone along with them (for a pithy account of how this came to be, see Diane Ravitch's recent Newsweek column).

How do you raise teacher quality? The Department of Education has decreed, in typically soaring federal prose, that districts should assess teachers "using multiple rating categories that take into account data on student growth as a significant factor."

How do you get "data on student growth?" You measure how well students do on standardized tests from one year to another, or one term to another, and then calculate the effect in a variety of complicated ways. This means you have to give a lot of tests.

Pay for performance, another favorite of the foundation world, takes this idea one step further – making growth scores part of a salary scale. On the surface, the idea of pay for performance doesn't seem so bad. But if every teacher's pay depends, even in part, on "data on student growth," then every student, in every subject, has to take standardized tests.

At this point the educational value of the tests becomes a secondary concern.

It does not matter that standardized tests are more effective at measuring science than art; more suited to high school students than kindergartners.

It does not matter that teachers seem to have considerable influence over math test scores, and very little over English test scores.

It does not matter that testing pulls a first grade teacher out of her classroom for more than 25 hours a year, or forces an ESL teacher to spend a full month administering and proctoring tests rather than teaching his students English.

It does not matter that standardized tests, by nature, can only measure a small part of what parents want their kids to learn at school.

Pay for performance requires that every teacher have a growth score, so there must be a test for every subject at every grade, and all those scores must carry equal weight. As with Procrustes' bed, any piece that doesn't fit the system must be either stretched or chopped off, regardless of the consequences.

In addition, these are high-stakes tests. Because a teacher's rating – and thus salary – depends on student scores, the tests must be given under controlled conditions, and they must be proctored (hence the ESL teacher who has to stop teaching his students so he can watch other students take a test).

Elaborate systems must also be designed to ensure that student scores are credited to the "right" teachers – an especially complicated matter when students are moved from school to school, or from one classroom to another, or in any other situation that doesn't fit the standard one-teacher-to-a-single-set-of-students model.

All this time and expense is justified, we parents are told, because it will get us better teachers. And better teachers will mean higher achievement.

But we parents know better. It's not just that we know research shows that pay for performance isn't fair or effective. We know that a school is not a collection of graphs, statistics and fully aligned, thoughtfully integrated, optimally balanced learning progressions. It's a group of parents, teachers and children exploring a messy world that can't be compressed into a neat set of multiple-choice bubbles, where learning doesn't move along straight, upward lines, and where a test that on paper takes two hours in reality consumes far more time and effort.

We're not opposed to tests per se – there's material our children need to master, and that needs to be monitored. But we're opposed to blowing testing so far out of proportion that real children and real classrooms and real learning fade from view.

We know what happened with the last round of high-stakes testing: a narrowed curriculum, more teaching to the test, and lowered standards overall.

We also know that increasing high-stakes testing is far more likely to lose us our best teachers than to gain us better ones.

As parents, we want our children to have the best teachers possible. We want them to achieve all that they can. But we know that this growth in testing is not the way to reach these goals.


For a research-based account of why tying teacher pay to test scores will not improve either teacher quality or student achievement, please visit the website of Parents Across America: http://parentsacrossamerica.org/performancepay/