Last month, when New York Gov. Andrew Cuomo proposed that all public school kindergartners sit for "entry assessments" starting in the 2014-15 school year, the education blogosphere erupted. Historian-turned-testing-apostate Diane Ravitch took to Twitter to decry the move, while the blogger Teacher Tom accused Cuomo of jumping on a "crazy train bandwagon" in which "even our three and four-year olds will be subjected to the dull tedium of a test-prep curriculum." What could be crazier, after all, than the prospect of a kindergartner gripping a No. 2 pencil in his tiny fist, looking down in confusion at a Scantron sheet he couldn't hope to decode, and bursting into tears?
Cuomo's proposal comes in response to the Early Learning Challenge, the latest iteration of Race to the Top, the Obama administration's 2-year-old school-reform grant competition. Critics of Race to the Top were complaining about its emphasis on standardized testing even before this latest program was announced back in May, but it seems poised to feed their fears: Starting next month, $500 million will be awarded to states that agree to rate preschools and day-care centers according to how well their "graduates" perform on "kindergarten entry assessments." A few states have already developed and deployed such assessments in classrooms, and last month, 35 states, Washington, D.C., and Puerto Rico turned in applications to vie for the funding.
Most 5-year-olds can't read, are just learning to write their names, and have trouble sitting still. Which raises some questions: Can kindergarteners even take tests? Where is all this going?
Race to the Top leaves it up to individual states to decide exactly what their kindergarten assessments will look like, and chances are, most will ape those states, like Maryland and Ohio, that are already assessing kindergarteners—albeit in very different ways. So what does kindergarten assessment actually look like in these states, and how reliable is it? First off, the nightmare scenario—the chubby fingers grasping a No. 2 pencil—bears little resemblance to what are currently considered the best practices in early-childhood assessment.
Maryland, Georgia, and Arkansas already require schools to use a highly-regarded kindergarten assessment called the Work Sampling System, which has almost nothing in common with a test as the term is popularly understood. Work Sampling was developed by Samuel Meisels, a leading researcher on early childhood development and the president of the Erikson Institute, a graduate school that trains educators to work with young children. The system simply formalizes what a good kindergarten teacher should already be doing: observing each child's behavior and skills over time, not on any specific test day.
While standardized tests of older children tend to focus exclusively on their academic knowledge and skills, the Work Sampling System looks at the whole child across seven "domains of learning": her social development, language and literacy skills, mathematical thinking, scientific thinking, social studies awareness, interest in the arts, and physical/motor skills. To perform the assessment, teachers in Maryland take notes on whether their students display key skills such as playing cooperatively with other children; drawing; singing; recognizing letters and numbers; writing their names; describing the properties of a physical object; showing awareness of basic geography, such as the name of their town, state, and country; and performing basic "self-care" tasks, such as putting on a coat, fastening Velcro shoes, and using a toilet. The teacher then gives each student a rating in each of the seven categories.
The goal of a kindergarten assessment like Maryland's is to help a child's current and future teachers tailor instruction to her strengths and weaknesses. Because the assessment is holistic and the ratings broad, however, it is difficult to build a school rating system—something that Race to the Top requires of states—around Work Sampling. That's why some states may instead choose to follow the example of Ohio, which has adopted Meisels' core principles, but attached them—controversially—to something that looks much more like a traditional test.
Within the first few weeks of school, every Ohio kindergartener sits down with his teacher for a 10- to 15-minute scored literacy assessment. The child is asked to perform six tasks: answer a question about chronology (What did Sam do first, wake up or eat breakfast?); repeat a sentence spoken by the teacher; recognize two words that rhyme (Which pair rhymes: cat and hat, or door and sky?); choose a word that rhymes with another (Can you tell me a word that rhymes with cat?); recognize and name a printed letter; and sound out the beginning of a printed word.
New York has already announced that, like Ohio, it will assess kindergartners on a specific date and at a specific time. But there are some difficulties in testing easily distracted 5-year-olds in this manner, as Ohio's guide for teachers performing the assessment hints. "Having to locate materials during the course of the assessment will create moments of downtime, during which children could lose focus," it warns. "A child who may be hesitant to participate may become even more reserved if encountering long delays, and a child who appears restless may become more so as a result of delays."
Of course, anyone who has cared for a kindergartner knows that immediate physical and emotional needs, such as having to use the bathroom, needing a nap, or being hungry or thirsty take precedence over more cerebral concerns such as learning to read and write. Some children will undoubtedly get up and walk away from an assessment setup. And as educational psychologist David Berliner wrote for the Washington Post over the summer, many a 5-year-old is unable to follow multistep directions, let alone weigh two options in the sophisticated manner that would lead to a "correct" answer on Ohio's literacy test—even when that child actually has acquired the skills the assessment calls for, such as the ability to rhyme.
The risk here is that many states will adopt assessments like Ohio's, which may not be based on the best developmental science, but which have the benefit of providing each child with a single "score" that can easily be fed into a computer database to create the school "quality rating and improvement systems" the Obama administration supports. So here's hoping that the Race to the Top judges favor states that plan to assess the "whole child" over many weeks and months, as Maryland does, not in one stressful 15-minute period, as Ohio does.
Just by promoting the idea of "assessing" 5-year-olds, the Obama administration has picked sides in one of the nastiest battles of the education reform war: the perennial debate over whether it's fair to create school accountability systems based on student test scores, or whether reliance on testing distorts teaching and learning, processes that critics maintain are not so easily measured. There is some irony here, though. If we get these new early childhood assessments right, they could provide an important model for the rest of the education system. Meisels' idea of seven distinct "domains of learning" is a powerful one. Of course, schools must teach children to read, write, and measure, but we also want to develop kind, empathetic, responsible young adults, people who get along with their peers, serve their communities, and ultimately become good citizens, not just effective students or workers. In other words, it is a vision that many testing critics could not just tolerate, but embrace.