I have argued in the New York Times and on these pages that we sometimes need “More Tests, Not Fewer,” but I also have concerns about “More Tests” and by listing some of the things that can go wrong with too much testing—or poor testing—I’m hopeful it will be possible to come up with better policy concerning the use of tests as tools for selection and more generally in schools.
First, in a nutshell, my interest in expanding testing originated with a desire to enhance the fairness of tests. Specifically, I’m concerned that tests such as the SAT and Wechsler tests of IQ, although excellent in many respects, overlook aspects of key human abilities. These overlooked abilities include:
- creativity (sometimes called “creative intelligence”)
- spatial intelligence
- personal intelligence (concerning reasoning about personality) and emotional intelligence (more specifically about emotions)
- social intelligence (reasoning about social communication and groups)
For me, using “More Tests” is a means of better and more fairly representing a person’s mental capacities and potential.
That said, there are many concerns that naturally go along with using tests more.
The Potential Overuse of Tests in Education and in Assessing Schools
The most frequently raised objection to “More Tests” is that we already test our youth too much, and that tests that are imposed by states and governments (which I am not advocating) over-control what teachers do, because the teachers end up “teaching to the test.”
I see this as a very legitimate concern. Tests that specify certain broad standards of learning end up ensuring that such material is taught, but also exert high degrees of control over what must go on in the classroom. This reduces an instructor’s capacity to choose what is important to teach, which may be influenced by the interests of her or his students, the teacher’s specific interests and abilities in the classroom, geographical opportunities, and other matters. The problem with over-controlling curricula through the use of tests is that it reduces class flexibility, creativity, and potentially, the initiative of teachers. I believe teachers deserve all the support they can get.
The issue here is not so much overusing tests, but rather who designs the test: the state or the teacher, how well the tests are designed, and how the tests are used—for example to assess the students, to grade the teacher, or to evaluate the school—and whether such uses are fair and reasonable.
I am not a policy expert in education but my impression is that there are very legitimate areas of concern regarding the ways in which tests are used for these purposes. Because I am not a policy expert in this area, I’ll refrain from speaking more to this point, but I acknowledge there are many difficult issues involved.
Test Bias Intrinsic to Tests
Another concern over “More Tests” is that tests are biased. I think this is less of a problem than was the case a few decades ago, but I acknowledge and agree that the test-taking process can be biased, and that economic and social factors can impact students’ test scores.
There are many kinds of test bias. Years ago, a meaningful number of questions on tests appeared sexist or posed questions that only students of middle-class or more privileged backgrounds could be expected to know. That kind of test bias is, thankfully, mostly gone today. That said, there exist other ways in which tests can reflect bias.
Bias Intrinsic to the Testing Process
Social prejudice as an influence on test-takers. Although test makers today have done a generally conscientious job of purging their tests of internal biases, there are external factors that can affect test performance. Among these are stereotype threat. If a member of a group has internalized stereotypes that she or he is inferior in thinking in a certain realm, and that stereotype is salient when taking a test, the test-taker’s performance will drop.
For example, if a European student believes that Asian students are better than Europeans at mathematics, and recalls that belief just before taking a math test, the European’s performance will be lower than otherwise. Stereotype threat isn’t a part of the test, but it becomes part of the testing process in such instances and can cause erroneously-biased measurement.
Bias Intrinsic to Family, Economic, and Related Factors
Economic disadvantage as an influence on test-takers
Economic disadvantages can impair test scores as well. If a student needs to hold a part-time job during high school or college in order to make ends meet, and that job is more than a few hours a week, the time spent getting to work, working, and returning from work is likely to interfere with the student’s ability to study over the long-term and to impair the student’s academic performance as a consequence. That will show up as poorer test-taking performance. In this case as I see it, the test is performing as it should, but the student’s education has been impaired.
Family disadvantage as an influence on test-takers
If a student comes from a family that is stressed, disorganized, or under intense economic pressures or pressures due to medical, disability, or other issues, a student may become so distracted that his or her ability to study will be impacted. If the student is anxious, distracted, or must take on new obligations to support siblings or parents, the student’s time for studying (and ability to concentrate on studies) will once again become impaired. Once again, the test is performing as it should, but the student’s education has been impaired. The test does, however, reflect the student’s poorer performance without specifying from where it has come.
The Potential Use of Bad Tests
A great deal of thought has gone into distinguishing “bad” or invalid tests—tests that don’t measure what they should—from good tests. The rules by which good tests are constructed and validated is a matter that is too broad to take up here, but I did want to raise the point that many tests in use today do not measure what they claim to. I believe that most educational and psycho-educational tests are designed by experts and do generally measure what they are supposed to but there can be exceptions.
Over-Controlling Our Children
Finally, I worry that by testing too much, and over-emphasizing tests, we are in danger of over-controlling our children in unhealthy ways. Children need some control over what they want to learn and like to learn. They need (and will take) the opportunity to decide that some subjects aren’t of interest to them at a particular point in time. Fewer tests can give students the opportunity to enjoy a subject for a while without facing that they might not be good at it right away, and that can be a pleasant experience.
More certainly, by testing often, parents and educators place students in factory-line like conditions of learning and testing, learning and testing. Although that might lead to their learning a great deal, it may come at the cost of self-motivated intellectual exploration and creativity.
These, then, are some of the concerns I have about the overuse of tests. As I try to create a balance between the desirability of testing more areas of intelligence, on the one hand, and over-testing, on the other, I’m trying to keep in mind some of the issues above.