What Standardized Tests Do Not Measure

By Peter Murrell

by Peter Murrell, J r.

Performance on standardized tests has become the bedrock upon which most discussions about excellence in education have been built. The technology of standardized tests — including standardized achievement, aptitude and intelligence tests — is so widely accepted by educators and psychologists that the professional literature on academic excellence is saturated with references to test scores. Yet few educators have adequately considered the limitations inherent in these tests or the dangers of framing excellence in terms of standardized testing.

Much recent research on intelligence and human cognitive development suggests that standardized tests are not valid as measures of excellence or scholastic aptitude. Schools which over-emphasize standardized tests and “test-wise” training are immersing students in modes of thinking that not only trivialize knowledge, but are also largely irrelevant to real-life problem solving. Failure to acknowledge the aptitudes not measured by standardized tests can deprive children of important opportunities for intellectual growth. Moreover, over-reliance on the standardized testing model limits the development of alternate ways of understanding and assessing intelligence.

In a review of the three major theoretical perspectives on intelligence, Robert Sternberg of Yale University notes that the standardized testing approach provides the fewest opportunities for improved educa – tional practice. The approach of standardized testing — called the psychometric approach — does not provide the information about learning processes necessary for designing instruction to overcome learning problems. For example, suppose that a child is given the Stanford Diagnostic Reading Test — a standardized achievement test for reading skills — and scores low on the ability to blend letter sounds. Even though this test provides specific diagnostic information, it does not provide the information necessary to help the child master the learning processes which have gone awry in her failure to do blending. Without an understanding of the specific processes that underlie this reading skill, the appropriate teaching and learning practices cannot be prescribed. If tests are to be of any use,

they should inform us not only what went

wrong, but also why.

According to Sternberg, the other two major perspectives on testing intelligence — the Piagetian approach and the information processing approach — are much more useful for determining how to help the child with learning problems. Sternberg argues that the kinds of skills and  knowledge called for on standardized tests are removed from the contexts in which students typically learn and then apply their knowledge. For example, tests which assess grammar skills, such as subject-verb agreement, in isolation, do not tap all the skills needed to write an essay.

Many of the skills that contribute to success on standardized tests are actually skills that we would not dream of teaching to our children. Take problem-solving skills for example. We want children to be able to recognize and define a problem, gather relevant information, and continue working on it until they are satisfied that they have found the best solution. In stark contrast, standardized tests give students “ready-made” problems having little or nothing to do with their real-life experience and ask them to pick the one best answer and proceed quickly on without knowing whether the solution is correct.

In the critical area of reading, overreliance on standardized testing can limit our ability to assess students. The Report of the Commission on Reading, published by the National Academy of Education, notes that standardized tests “do not measure everything required to understand and appreciate a novel’, learn from a science book, or find items in a catalogue.” Most standardized tests do not permit the reader to make use of the contextual and situational cues normally provided by the author. The report suggests that assessment should include observations of reading fluency, and inspection of children’s independent reading and writing. Additionally, performance on standardized reading tests are frequently based upon strategies not used in normal  reading. For example, “test-wise” students develop strategies used only when answering multiple choice questions. The subsequent test results make these students appear to be better readers than they actually are.

Just as important as their failure to fully tap the cognitive aspects of learning achievement, is the failure of standardized tests to incorporate the motivational and situational factors which are critical in human learning in real-world situations. When giftedness and learning achievement are examined in real life settings, what emerges is a picture of excellence more complex than that portrayed by an assessment of intellectual abilities. Mihalyi Csikzentmihalyi, professor of psychology at the University of Chicago, studied how gifted people approach daily problem solving and other mental tasks. He found that motivational factors are most important to outstanding learning achievement. Qualities of temperament^— persistence, determination and tenacity — are more clearly associated with high achievement than intellect.

The claim here is not that testing is evil  or useless. Testing does have a legitimate use in assessing student achievement. But we should move beyond the narrow definition of scholastic aptitude inherent in standardized testing. The dominance of traditional testing in our schools has made  it harder for us to help children become more aware of talents then can develop in areas in which they are likely to be satisfied and competent. 

A growing number of educators and psychologists are using alternative conceptions of aptitude and intelligence to predict and develop a wide range of talents among young people. A central theme in these new approaches is that intelligence is not a single property of the human mind. Rather, there are many intelligences which may be relatively independent of one another and which do not easily lend them – selves to measurement by short-answer, pencil-and-paper tests. For example, social perceptiveness is one of several types of intelligence researched by Howard Gardner of Harvard University and David Feldman of Tufts University in the Spectrum research project One of many methods they developed is the Classroom Game, a dollhouse version of a preschool classroom in which the identity of each child is represented by one of the dolls. A child’s understanding and awareness of other childrens’ play preferences, social inter – actions, and relationships is determined by  how the child places the dolls representing members of the class. For example, a child might be asked to “put each child in the room where he or she most likes to play” or to “put each child with the children he or she most likes to play with”. Children with an aptitude in social perceptiveness can  correctly identify the other childrens’ play preferences and social interactions based on the doll placements.

Will we allow our schools to nourish and develop a wide range of aptitudes, such as social perceptiveness, in children who have them? Or will we continue to discourage children from developing important aptitudes because they are unacknowledged or devalued by, our primary testing instruments. We would not consciously choose to ignore valuable information about a child’s total range of aptitudes (e.g., social intelligence) in the design and management of successful schools. Yet we tacitly make this choice as long as we continue uncritical acceptance of standardized testing  as a measuring stick for academic excellence. We need to regard children’s full ranges of aptitudes and their full participation.in learning as no less important than their test scores. Improving our schools will require close scrutiny of the purposes for test administration and constant vigilance for misapplications of test results.

Decisions about children which take into account their social development as well their cognitive development come closer to the notion of excellence than does merely demanding that children liv up to arbitrary numerical standards. Instead of learning the limits of childrens’ achievement through analysis of standardized tests, let us learn the limits of standardized tests in analyzing childrens’ achievement.

Peter Murrell, Jr. is a psychology instructor at MATC.