Building a Better Assessment Program

• 12/01/05

Now is the time to utilize the technology we have, to create a system that encourages students to learn—rather than just pass—a test.

HAVEN’T WE ALL WANTED to be governor for a day with unlimited powers? I have. But only for a day. I just want to institute two simple policies in the state of Washington that could apply to any state. To wit:

1. The state’s assessment program will live online and consist of a variety of material, including performance measures, open-ended items, and essay questions. There will be no multiple- choice problems. Said program will be accessible 24/7 to everyone on the Web, and its purpose will be to help teachers, students, and parents understand the extent to which individual students and aggregations of students are learning.
2. The program will have no name and cannot be referred to as an aspect separate from any other component of the teaching/learning process.

As with many policies, each of the two has a genesis and each requires some explanation. Let’s start with the second policy, which grew out of a front-page article I read in October’s Seattle Post-Intelligencer, titled, “Prepping for the WASL: Seattle puts new teaching tools in place for students, educators.” (The WASL is the Washington Assessment of Student Learning, the state testing program.)

I confess that I sat slack-jawed with coffee dripping down my chin as I read the article. The story was all about the policies, tools, and support the Seattle Public School District was implementing…to help kids pass the test. Passing the state test has become a proxy for actual learning. This isn’t a problem unique to Seattle. I am not so naïve to think that the mantra of “All we do is teach to the test” isn’t chanted daily in schools across the country. It’s even worse in states where students must pass a test before they can graduate. I did not think, however, that the emphasis on the test, as opposed to student learning, was as bad as portrayed in this article. Below are sample quotes from the article.

• “Recognizing that math is the weakest WASL subject for many students, the district is in the process of reviewing instructional materials and establishing a standardized curriculum.”
• “The WASL requires students to both reach the correct answer and explain how they got there, necessitating what [a school official] said is a major shift from how teachers used to teach.”

I must say that the district is taking some excellent steps toward improving teaching and learning. Ensuring students have mastered certain skills and knowledge before they move on; reviewing instructional materials; establishing a standardized curriculum; and helping teachers change how they teach so students not only get the correct answer, but also understand the concept the answer represents and how they got to the answer—all these activities, and many more the district is working on, demonstrate good policy. But it took the threat of a test to get the district to do this? As my parents used to say in looking at my grades: I am disappointed.

This brings me to my first policy, which is predicated on two assumptions: 1) Teachers need to know as precisely as possible, as often as they want, how each of their students is doing every day and at key benchmark times; and 2) an assessment program should meet the needs of its users, rather than force the users to fit the needs of the assessment program.

The policy actually sprung from a conversation I had over dinner with Booth Gardner, former governor of Washington state. Gardner is tired of the complaining he hears over the WASL, but he understands the legitimacy of some of the grievances. He has assembled an advisory rump group to come up with some ways to make the program better. When large assessment programs started 25 years ago, if we wanted results back within a reasonable time, the only real option we had was a machine-readable multiple-choice test. What is different today is the capability of the technology.

We have unlimited storage capacity to maintain all kinds of different examples of student work. We can store written essays; audio files of music performances; and video files of dance, exercise, and presentations. In short, we can store almost anything a student does. Likewise, we have the technology to grade electronically open-ended questions, such as fill-in-the-blank, short answer, and essay. And the technology can grade just as consistently and reliably as a group of humans can, in a much shorter time.

To carry out my policies, we need more technology in the classroom and more bandwidth, but that can be handled relatively easily with more money. The hard part is convincing policymakers that we can fairly evaluate students’ knowledge and skills with measures other than multiple-choice tests. We need to develop a trust in technology’s ability to be as accurate and fair as humans. I know the devil is in the details, but if we don’t implement my policies, or something like them, we will never get past focusing on the test to put our attention on learning.

Geoffrey H. Fletcher is editor-at-large of T.H.E Journal and executive director of T.H.E. Institute.