The federal “No Child Left Behind” law (NCLB) requires that states test all their third through eighth graders, as well as students in at least one high school grade, annually in reading and mathematics. Testing these students makes sense because of the critical importance of these subjects and because virtually all students receive instruction in them.
In that broad sense, there is a common core curriculum applicable to every school’s mission. States reinforce this common core in several ways, such as specifying statewide curriculum goals for each subject and grade level, selecting certain tests for statewide use, and approving certain textbooks for statewide adoption.
But postsecondary education is different. Colleges and universities have varying missions. Students choose differing academic majors. Even within an institution, faculty members often disagree on what should be taught in their field and how learning should be measured. In short, one size does not fit all.
Hence, the NCLB approach is not applicable to colleges. However, the increasing push for accountability by parents, trustees, and the elected officials has led accrediting agencies and the public to put more pressure on colleges and universities to provide information about educational outcomes. This pressure has led some to suggest looking at how well students perform on admissions tests for graduate and professional schools.
While this approach helps solve the problem of student motivation, it has serious flaws. Different admissions tests measure different abilities. The abilities these tests measure may not be central to a school’s mission, and only a small percentage of graduates take such exams. And those who do are unlikely to be representative of a school’s graduates. Using graduate admissions tests to assess an institution’s effectiveness is therefore analogous to measuring the oral communication skills of a school’s graduates on the basis of the performance of its debate team.
NCLB and graduate school admissions tests rely almost exclusively on multiple-choice questions. This reliance stems from statistical, logistical, and cost considerations. For example, it is much cheaper to process several hundred machinescored answer sheets than it is to grade a student’s essay. Although multiple-choice tests are effective in measuring many things, they are not appropriate for evaluating the higher order abilities we expect of college students. For example, open-ended tests are needed to assess a student’s ability to spot flaws and logical inconsistencies in someone else’s arguments and to integrate information from various sources to effectively form, present, and justify their own conclusions.
Some colleges have responded to accountability pressures by leaving it up to faculty in each department to develop their own indicators of student learning that are tailored to their specific goals. While this strategy has some positive features, it cannot assess whether the improvements that do occur at an institution (say, between freshmen and senior years) are exemplary, adequate, or below par.
To make that type of judgment, we need benchmarks common across institutions. Only then can we assess whether the gains at a school are more or less than what would be expected given the incoming abilities of its students.
These considerations have led the Council for Aid to Education to develop a set of open-ended measures called the Collegiate Learning Assessment (CLA) that are administered and scored online. These measures assess some of the important abilities that are mentioned in almost every college’s mission statement, such as written communication skills, problem solving, critical thinking, and analytic reasoning. CLA results are reported in terms of whether a school’s gains are more or less than what would be expected given (1) its students’ entering abilities (as measured by their SAT and ACT scores) and (2) the gains of comparable students at other schools.
Under NCLB, each state selects its own tests and passing scores; thus, there is no uniformity in standards. In contrast, the CLA uses the same value-added metric for all schools, encourages high standards, and emphasizes improvement. I see this as the more promising option for assessing learning.
