Friday, December 11, 2009

Tests Hamper Teaching Students to Think Like Einstein

by Cathy N. Davidson, Duke University

Reblogged from the Durham Herald Sun, December 11, 2009

I often lecture or blog about grading, arguing that the way we now assign grades is an antiquated system that may have worked well for the Industrial Age but that undercuts what is valuable, exciting, or potentially useful for interactive thinking in the Digital Age. I’m often critical of so called "standards-based education" such as No Child Left Behind, with its reduction of evaluation and assessment to standardized testing.

But I’m actually criticizing here a much broader way of thinking that reduces the process of thinking to “a result,” even to “the best result chosen from among a select number of choices” (i.e., multiple choice exams).

That concept of grading seems the exact opposite of critical and daring thinking, and inconducive to the kind of integrative, creative, innovative thinking our era demands, in all fields from the arts to the theoretical sciences and engineering.

Whenever I talk about new ways of evaluating, someone in the audience inevitably retorts, “Well, that’s subjective. It may be fine for humanists — but it would never work for science. We need rigorous, standardized testing to produce the highly specialized scientists necessary for our world.”


But within a minute, I can get this same person pontificating in a different direction simply by switching the topic a little, lamenting, “and isn’t it terrible that America today undervalues science and produces so few scientists?” No argument there!

But now let’s put those two arguments together. What if it turned out that our “rigorous” standardized, multiple choice form of testing — in all fields, including science and math — selected out those who do well on standardized tests but who lack precisely the forms of inquisitive, inductive, hypothetical reasoning and willingness to tirelessly test out a hypothesis that is the basis of the experimental method and exactly what science demands?

Our entire practice of testing is based on a theory of knowledge that is out of date. It used to be thought that brains and neural connectors grew in the same way feet do, tiny at birth, growing until maturity.

We now know that infants have an overabundance of neurons and that, if neural development proceeds on course, they will shear off about 40 percent of their neurons on their way to an adult understanding of the world, working on streamlining neural pathways by repetition and experience, using the scaffolding of one experience (and that of their culture) on which to build ever-more reflexive ways of reacting on which to then build more nuanced, interactive, reflective ways of thinking later.

Much of our standardized testing is still based on an outmoded filling-station view of neural development and of knowledge. Heads don’t fill up with knowledge. New kinds of knowledge build upon older knowledge and often replace that knowledge. Everything works in that process of selection, adaptation, revision, selection.

Memorizing correct answers to questions has some function, but it is not at all clear to anyone what that function is or how useful it is in an era of search and browse.

Socrates had it right. If you want to model higher level thinking, you don’t lecture about your insights achieved as the result (“the answers”) of such thinking. You certainly don’t have students take a multiple choice test to ensure that they remember your conclusions. If you want to encourage the love of thinking and the skill of critical thinking, you question them, you hear their ideas, you debate them, you give them feedback, you lead and mislead them, you intellectually thrust and parry, you joust, and you have them reach conclusions by learning which intellectual moves are fruitful and which lead to dead ends.

That Socratic method is used in law schools today, but I’m suggesting should be true for all fields — including the sciences.

It is a profoundly humanistic method and, to make great scientists, it is that profoundly humanistic method that is required, the ability to think through an idea, to revise an idea in light of other ideas, to test and question, to think critically, to analyze data, to respond to the arguments or hypotheses of others, and on and on.

It may not yield the highest test scores on SAT’s, but it may well be what sorts out the kind of process-oriented mental habits of those who are most likely, someday, to think like Einstein.

Einstein, of course, grew up loving to make little mechanical devices. And he had, as a very young man, two favorite books: Euclid’s “Elements” and Kant’s “Critique of Pure Reason.”

He was one of the world’s most famous dyslexics, but he was also someone who, throughout his life, understood the contuities between mechanism, geometry, number theory, a priori concepts and experience.

How do you answer a multiple choice test for pure reason? I fear that some of the standardized assessment aspects of No Child Left Behind may well be constructed to leave behind exactly those non-linear, inductive, intuitive, critical, curious, humanistic, and scientific thinkers who, if nurtured, might well grow up wanting to Be Like Einstein.


  1. Elegant!

    Thanks. I'm in the process of designing immersive, interactive learning environments. This is a great foundation and inspiration for a conceptual approach.

    Marv Lyons

  2. It sounds like teaching is being confused with testing. It is the fault of the system that teachers teach to the test and only the test. I see no conflict between a child being able to answer factual multiple choice questions and critical thinking. I'm sure Einstein could tell you which element had how many electrons, or what force applied in this or that situation. He had to understand gravity and all the calculations to know and verify that they were indeed wrong. He set up the experiments according to traditional methods to prove it.

    Evaluation on a mass and standardized scale has its benefits. Instead, it is the system and method of teaching that fails to incorporate and accomodate both. Innovative approaches can do both.

  3. Exactly. No argument. These are not mutually exclusive by any means, or should not be, but I am suggesting that NCLB (and others too, such as some college admissions offices) place too much emphasis on one form of standardized testing. I'm not against testing per se but only the belief that one kind of test reveals all there is to reveal about knowledge. I agree entirely that innovative approaches should be multiple in emphasis and in assessment. We've gone way too much in one direction and it may be that encouraging right answers postulated in a particular form does not encourage process-oriented scientific thinking and the experimental method (including trial and error and other forms of iterative thinking. I am suggesting that, if we have a crisis in STEM interest, we should think about what we are doing to model and reward and, most important, inspire true scientific thinking.