When Frederick J. Kelly invented the multiple-choice test in 1914, he was addressing a national crisis. The ranks of students attending secondary school had swollen from 200,000 in 1890 to more than 1.5 millionas immigrants streamed onto American shores, and as new laws made two years of high school compulsory for everyone and not simply a desirable option for the college bound. World War I added to the problem, creating a teacher shortage with men fighting abroad and women working in factories at home.
The country needed to process students quickly and efficiently. If Henry Ford could turn out Model Ts “for the great multitude,” surely there was an equivalent way, Kelly wrote in his dissertation at Kansas State Teachers College, to streamline schooling. What he came up with was the Kansas Silent Reading Test, sometimes called the “item-response” or “bubble” test.
Today, American public school students are still taking versions of Kelly’s test. End-of-grade exams, required under the No Child Left Behind law, are modeled after his idea: Fill in the circles. There is only one right answer. Stop when time is called.
The Obama administration is currently undertaking an important overhaul of key parts of No Child Left Behind, the George W. Bush administration’s signature education law. Perhaps that reform needs to go even further. Antiquated standardized tests are still serving as the backbone for measuring achievement. Our students can’t escape Kelly’s century-old invention.
We often forget that there is nothing fixed or natural about “school” but that, like all institutions, it evolves in response to historical circumstances. In the case of standardized testing, the multiple-choice exam has had an impact far beyond the crisis that inspired it, and a reach and application far beyond what its inventor intended. From his papers at the University of Idaho, it is clear that Kelly didn’t mean for standardized testing to become so widespread. Although he argued for uniform ways of judging achievement, he also indicated that his Kansas Silent Reading Test was intended to measure “lower-order thinking” among the masses (which were then called the “lower orders”). But this form of testing is, of course, now the gold standard for just about everything, from No Child Left Behind tests to college entrance exams to tests for graduate and professional schools.
Once World War I was over, Kelly himself began to ardently champion a different direction for educational reform, a model of liberal, integrated, problem-based learning. In his inaugural address as University of Idaho president in 1928, he argued for expansive changes almost diametrically opposite to his early advocacy of standardized testing. “College is a place to learn how to educate oneself rather than a place in which to be educated,” he insisted.
Too late. By then, the College Entrance Examination Board had adopted Kelly’s test as the basis for its Scholastic Aptitude Test (SAT). Business schools and schools of education were using item-response testing as the new metric for measuring success. Kelly’s faculty was furious that the inventor of the bubble test now advocated a different course, and he was asked to step down as president barely two years later.
Kelly couldn’t get rid of the test he created, but we should be able to. Institutions of education should be preparing our kids for their future — not for our past. In the Internet age, we are saddled with an educational system that was designed for the industrial age, modeled on mass production and designed for efficiency, not for high standards.
We know that bubble tests address only a quarter of the kinds of knowledge students master in schools. For low-income kids, who have limited resources for college costs and thus little reason to think that their test scores matter to their future, the exams can seem irrelevant. For them, low scores can denote not just a possible lack of knowledge but also a possible lack of motivation to concentrate on the exam. Affluent kids, if they pay enough and take enough test-prep courses, can get higher scores.
We are not teaching and testing our students for responsible participation in the interactive digital age. Even at IBM, that industrial age behemoth, the developer of the time clock, 40 percent of employees now work at least partly at home in a system called “endeavor-based work.” School bell? Timed tests? Right answer chosen from four preselected ones? What does that old form of education have to do with our children’s future — or, for that matter, their present?
A standard test question for young students today, dating back to Kelly’s original exam, is which of four animals is a farm animal (a cow, a tiger, a rat or a wolf). The same child who flubs this question can go home, Google “farm animals” and get 14.8 million results. What part of our educational system prepares a student to sort through and evaluate all those Web sites? Multiple-choice exams do not equip kids for either the information avalanche or the fine print that they encounter online every day.
In a decade of researching digital education, I have never heard an educator, parent or student say that the tests work well as teaching tools. Beyond the flaws of these rigid exams — which do not measure complex, connected, interactive skills — there is little room in the current curriculum or even in the current division of disciplines (reading, writing, math, natural sciences and social studies) for lessons about key questions that affect students’ daily lives. Teaching kids about credibility, security, privacy, intellectual property and other bases of their online lives, after all, would take time away from the tested subjects that lead to merits or demerits, funding or reduced funding, depending on exam scores. Yet every school curriculum should include interactive lessons in practical, creative and cautious participation in the World Wide Web. Ideally, students should be learning these lessons even as they learn the basics of code.
We’re facing a crisis in education today, much like Kelly faced in 1914. The U.S. high school completion rate is dropping slightly in real terms, and dramatically relative to other industrialized nations. But even more serious is the rate at which teachers are leaving the profession. Teacher attrition has increased by more than 50 percent in the past decade, and it is often the best teachers who leave first, many citing the demoralizing testing requirements of No Child Left Behind.
Right now, we have teachers, out of self-preservation and to protect their schools and their students, teaching to a test that was designed in the era of the Model T. We are 15 years into the information age. Now is the time to begin to rethink how we assess learning for the challenges of the digital world that lie ahead. It’s not as simple as filling in the bubbles.
Cathy N. Davidson is the author, most recently, of “Now You See It: How the Brain Science of Attention Will Transform the Way We Live, Work, and Learn,” from which parts of this essay were adapted. She teaches at Duke University.