Print

Tests Aren't The Problem — And New One Is No Solution

By JENNIFER DOLAN

July 20, 2012

State Education Commissioner Stefan Pryor said that a new standardized test given on computers will adjust its questions according to a student's answers. If a student gets a question incorrect, "the system will provide other questions to try to get at what the underlying deficiency is," he said.

Amazing.

This computer-based test, which replaces the Connecticut Mastery Test, will narrow down what lies behind a student's incorrect answer. As a former elementary school teacher with years of mastery test experience, I question this new test's ability to determine whether children had breakfast, are proficient in English or are familiar with standardized testing and computer use, among the myriad factors that affect scores or create "deficiencies."

New is not necessarily better. Rather than ask whether a standardized test score accurately reflects what students know, or are able to do, or how standardized tests may reflect students' access to opportunity rather than their achievement; our answer appears to be simply creating a "smarter" test.

Under this new computer-adaptive testing, students who answer a question correctly will get a more difficult question. When an answer is incorrect, the computer reduces the degree of difficulty. Eventually, you are able to determine a person's ability (imagine a see-saw that leans to one side or the other until it is balanced). The issue is, however, that you start with an assumption that the questions are valid.

A sample question reported by The Courant is a perfect example of an imperfect question. The first sentence of the question is "You will learn about young people who, because of their actions, are considered to be wonders." I asked my husband to read the question. He said, "Wonders? What are wonders?" Exactly.

A CPA with two master's degrees was immediately confused about the vocabulary and stopped reading the directions. At the bottom of the question, it tells students to look up wonders in a dictionary and then define it in their own words. As a teacher, I know that many students will try to answer the question without reading all of the directions. In addition, there are eight steps a student needs to complete for this question, after defining a "wonder."

The convoluted language, difficult directions and multiple steps all confuse students rather than prompt student thought. The vocabulary in the sample was difficult enough for my husband, but factor in cultural and socioeconomic differences, as well as students with limited English, and the tests limit access to achievement.

Compounding the overall issue of test complexity is the added pressure of the timed format of standardized tests. The ticking clock may cause students to rush through directions, leave work unfinished or give less in-depth responses. These responses do not accurately show what students know or can do, but what they can do in 45 minutes?

Also, some types of questions, such as those requiring essays or the new performance tasks, need to be hand-scored. Differences among scorers, their biases or fatigue could change a student's score without reflecting their ability. If students get a question wrong (or loses points) will the computer attribute it to time management, scorer bias, vocabulary, reading fluency, unfamiliarity with technology or attention issues? How can you determine students' ability or challenges simply in the way they answer a question?

As we move to this new testing, there is an additional unmentioned elephant in the room. This system requires that all students complete at least a portion of the tests on a computer. Schools will need to be equipped. When school systems such as New Britain are cutting budgets (and apparently 50 teaching positions), where will they find the money to not only provide the technology, but the time and the instructors necessary to prepare students?

Connecticut has the largest opportunity gap in the country with regard to education. Implementing new standardized tests that not only have similar issues to the old tests, but also depend on technology does nothing to reverse this trend.

Is testing the problem? No, but testing that narrows curriculum and reduces content to the point that we determine whether or not a student is "achieving" by one variable is. To truly help our students achieve, we need to move away from our current mindset of "test pedagogy" in schools and think about how to create environments that draw from students' strengths.

Jennifer Dolan of Storrs is a doctoral candidate at the NEAG School of Education Department of Curriculum & Instruction University of Connecticut. She previously taught fifth grade at the Charter Oak International Academy in West Hartford.