Millions of kids are competing for seats in gifted programs and private schools. Admissions officers say it’s an art: new science says they’re wrong, 73% of the time.
Picture the little child, just turned five, being escorted into a stranger’s office. Her mother helps her get comfortable for a few minutes, then departs.
Mommy might have told the girl that the stranger will help determine which school she goes to next year. Ideally, the word “test” will never be spoken; if the girl asks, “Am I going to take a test?” she will be told, “There will be puzzles and drawing and blocks and some questions to answer. Most kids think these activities are really fun.”
The child is directed to a seat at a table. The test examiner sits across from her. If she grows restless after a while, they might move to the floor. (If there is a significant problem, some schools may allow a retest, but most won’t allow her to return for a year or two.)
They begin each test with a series of sample questions, which the examiner demonstrates. Then they start the real test. The examiner begins with a question appropriate for the child’s age. A five-year-old child might start with question no. 4 of the testing book. Each question gets a tiny bit harder, and the child keeps receiving questions until she has made several errors in a row. At this point, the “discontinue rule” is triggered—the little girl has reached the top of her ability, and they move on to a new section.
Vocabulary is tested two ways; at first, the child merely has to name what’s pictured. When it gets harder, the child will be told a word, like “confine,” and be asked what it means. Detailed definitions merit a 2; less detail scores a 1.
The little girl will also have to discern a word from just a few clues. “Can you tell me what I’m thinking of?” the examiner will ask. “This is something you can sit or stand on, and it is something that can be cleaned or made of dirt.”
She has five seconds to answer.
On another section, the child will be shown pictures, then asked to spot what’s missing. “The bear’s leg!” she’ll answer—hopefully within 20 seconds.
Later, the examiner will set some red and white plastic blocks out on the table. The child will be shown a card with a shape or pattern on it, and she’ll be asked to assemble four blocks to mirror the shape. Blocks arranged more than one-quarter inch apart are penalized. The harder questions use blocks with bicolored sides—red and white triangles. Older kids get nine blocks.
She may also see some mazes; no lifting of the pencil is allowed, and points are taken off for going down blind alleys.
Discerning patterns is a component of all tests. For instance, the child will need to recognize that a circle is to an oval as a square is to a rectangle, while a triangle is to a square as a square is to a pentagon. Or, snow is to a snowman as a bag of flour is to a loaf of bread.
If a child is six years old, she might be read four numbers aloud (such as 9, 4, 7, 1) and asked to repeat them. If she gets them right, she’ll move up to five numbers. If she can do seven numbers, she’ll score in the 99th percentile. Then she’ll be asked to repeat a number sequence in reverse order; correctly repeating four numbers backward counts as gifted.
Every winter, tens of thousands of children spend a morning or afternoon this way. Testing sessions like this one are the keys to admittance to elite private schools and to Gifted and Talented programs in public schools. Kids are scored against other children born in the same third of the year. Mostly based on these tests, over three million children—almost 7% of the American public school population—are in a gifted program. Another two million children won entry into private independent schools.
The tests vary by what exactly they examine. Some are forms of a classic intelligence test—for instance, the Wechsler Preschool and Primary Scale of Intelligence, known by its acronym, WPPSI. Other schools opt for an exam that doesn’t strictly measure IQ; they might use a test of reasoning ability, such as the Cognitive Abilities Test, or a hybrid test of intelligence and learning aptitude, such as the Otis-Lennon School Ability Test.
Regardless of what is being tested, or which test is used, they all have one thing in common.
They’re all astonishingly ineffective predictors of a young child’s academic success.
While it’s no surprise that not all gifted kindergartners end up at Harvard, the operating assumption has been that these screening tests do predict which kids will be the best at reading, writing, and math in the second and third grades. But they turn out to not even do that.
To give you a hint of the scale of the problem—if you picked 100 kindergartners as the “gifted,” i.e. the smartest, by third grade only 27 of them would still deserve that categorization. You would have wrongly locked out 73 other deserving students.
Most schools don’t realize how poorly the tests predict a child’s elementary school academics. The few with concerns have tried to come up with other ways to test for giftedness—everything from asking a kid to draw a picture to rating a child’s emotional empathy or behavior. However, scholars’ analyses have shown that each of these alternatives turns out to be even less effective than the intelligence tests.
The issue isn’t which test is used, or what the test tests. The problem is that young kids’ brains just aren’t done yet.