Why can’t we just get rid of standardized tests?

At various times, we certainly have wanted to get rid of standardized tests. Large scale assessment has been so beset with problems for so long that simply getting rid of standardized tests has been an attractive idea, even to us.

However, high quality tests used for valid purposes are not a problem. The problem is low quality tests and the misuse of tests.

Are there really decent uses of standardized tests?

Putting aside for a moment the most important use of standardized tests in K-12 schools, there are many other purposes that are not entirely without merit. They are all built either on the desire to sort and/or rank students, and/or the need for a credible/trustworthy measure of test takers.

For example, colleges may wish to determine whether incoming students merit some sort of college credit for classes they took in high school. Advanced Placement (AP) exams are designed for precisely that purpose. Colleges know and trust the AP exams and do not have to determine how trustworthy or credible individual teachers’ or schools’ attestations of students’ learning are. More generally, high quality standardized tests can provide insight into student learning without relying on trusting teachers or schools — something quite valuable, at times.

There also are a wide variety of decisions that education leaders—school leaders, district leaders, elected official, experts in various departments of education—need high quality information in order to make. These decisions are not based on seeing and understanding individual students, but rather rolled up summary data about large groups of students (or other test takers). Standardized assessment is the only way to produce the kind of comparable information that can be rolled up into valid summaries of large groups of students (or test takers).

Sorting and ranking students is obviously much more difficult. We do not believe that standardized tests should be the sole determiner of such decisions. Professional tennis player rankings are based on a whole year of performance, not just one match on one day. Nonetheless, using a highly comparable measure can be valuable when that measure actually assesses what it is purported to measure.

That is, there are no decent uses of bad standardized tests. But when they actually measure what they are supposed to measure, they can provide valuable information.

What is the most important use of standardized tests?

We believe that the most important use of standardized tests in America might be — and certainly, around K-12 schools, is — to inform members of a community about the academic performance of their schools. The moral legitimacy of our public schools stems from the fact that they are overseen by school boards who are elected by the people. Standardized tests ought to offer the public information and insight into this. This is not to say that academic performance is the only important quality of schools, but it certainly is one of the very most important ones.

Our frustration with tests that are not up to this task, with psychometrics that is not up to this task and with score reporting that is not up to this task is one of the main original motivators for this project.

Voters need to be able to evaluate the performance of their local schools — all voters, not just parents. And parents need information about other teachers and classrooms and schools if they are to vote wisely as members of their communities.

When standardized tests are up to this task, they will be up to many other tasks, as well.

So, we need standardized tests?

We need good standardized tests that are used properly. We need good standardized tests that support important purposes, especially when there are not other sufficient sources of information to inform those important decisions.

But we do not need bad standardized tests. Nor do we need decent standardized tests to be misused for purposes that they do not support. In fact, we need all of that to stop.

Does the Rigorous Test Development Project support standardized testing?

We believe that the RTD Project can be used to make among the strongest arguments against bad standardized tests and the misuse of tests because we provide the tools to best criticize the shortcomings of the very building blocks of tests.