Here’s a fun party trick.

Look at your hands. Put your hands flat up against each other: Now see if each finger is exactly the same length as its match on the other hand. How about your ears? Are they the same shape and size, down to the nearest hundredth millimeter?

You probably don’t think this has anything to do with IQ. But strangely—amazingly—it sort of does. According to many studies from the University of Edinburgh and the University of New Mexico, body symmetry is a predictor of your IQ. It's not a perfect one-to-one correlation, of course. If a perfect relationship equals 1.00, consider that the correlation between body symmetry and IQ is 0.39. That probably sounds at once surprisingly strong and completely ridiculous, as predictors go.

What this means is that many kids who turned out to be very good students were still fidgety and misbehaving at age 5, while many of the kids who were well-behaved at age 5 didn’t turn into such good students.

But keep that .39 correlation in mind, because we’re going to use it as a benchmark.

In our book NurtureShock, we wrote about how millions of young kids’ fates are decided by IQ tests administered by private schools and gifted programs. Yet those early IQ tests are also far from perfect at predicting which kids will actually excel on achievement tests down the road—whether it’s 3rd grade, 8th grade, or the SAT. In fact, they’re so far from perfect it’s laughable. The correlation is only 0.40.

In other words, early IQ predicts academic success only 1/100 ^th better than matching ears predict IQ.

The reason for this is no mystery—brain development is uneven. Fully one-third of the top-performing 3rd graders scored below average prior to kindergarten.

Since our book's publication in September, a number of other reporters have begun tackling the issue. First, Susan Dominus did a piece for The New York Times, as did Laura Vanderkam for her Gifted Exchange blog. Psychology Today picked up on the theme. More recently, Jennifer Senior did a piece for New York magazine on testing kids for admission to New York City's elite schools. The consensus is clear: Early IQ tests aren’t reliable enough to be the basis for such momentous decisions over children’s lives.

However, there’s been a fatal flaw in where the logic has gone from there. Some reporters and bloggers have been advocating that schools should use behavioral measures and social skill ratings instead of IQ tests, or to supplement IQ tests. The idea—which has been bought into, wholesale—is that young kids’ behavior is the missing piece in the admission equation.

Schools are fairly mysterious about their exact testing methods, but word leaks out. Admissions officials scrutinize kids' behavior during on-campus play groups. Others regularly send spies into preschools. And preschool teachers are routinely sent paperwork to formally rate applicants’ attention span, ability to obey instructions, and self-control.

Sandy Eiges, a Los Angeles-based school consultant, explains that the schools are looking for kids' ability to work with other children and focus. "Do they dash around the room, not really doing anything? Do they seem curious, inquisitive, hyper?"

According to Eiges, schools absolutely believe they can predict kids' future academic success through these short observations.

Amanda Uhry, founder of Manhattan Private School Advisors, agrees—and says that New York City's elite schools go even further. "Behavior is incredibly important. Compliance and staying on task." Uhry herself doesn't believe those qualities predict school success, but she thinks they are determinative in school admissions.

Uhry reports that schools ask kids to hop on one foot and perform other tasks. These aren't a test of the children's motor skills, but a test of the child's willingness to follow instructions. Kids are asked to list the months of the year; it's a back-door assessment of kids' ability to complete a task without their minds wandering in the middle. Other activities that are considered telling include seeing how kids handle criticism when drawing a circle, and if they can resist playing with a cool, distracting toy that’s nearby when they’ve got an academic task in front of them, such as penning the alphabet.

While it may be normal for a 3 year old to fidget, resist, and even occasionally take another child's toy, such normal behavior doesn't cut it for the New York kindergarten admissions officer. According to Uhry, a child really could be a genius, but he will never get into an elite school if he exhibits any such weakness. Instead, schools demand an unrealistic amount of obedience at admissions.

The argument behind it all seems lock-solid. If a kindergartner can’t focus, he won’t learn. If he doesn’t learn, he’ll fall behind. Therefore, by simple logic, each year this gap compounds itself—the kids who behave well and follow instructions will excel.

Unfortunately, this logic totally ignores common sense. Because a child’s behavior is regulated by the brain, as well. And those neural networks governing behavior are no more fully formed in a 5 year old than the intelligence networks. So the kids who can focus better at age 5 are not the same kids who can focus best in 2nd grade.

In 2007, Dr. Greg Duncan published a massive analysis of 34,000 children, with no less than 11 other prominent co-authors. The scholars had expected to find that behavioral ratings were a stronger predictor of academic success through elementary school. So they combed through the data from six long-term population studies—four from the United States, one from Canada, and one from the United Kingdom.

Prior to kindergarten, the children participating all took some variety of intelligence test or achievement test. As well, mothers and teachers rated their social skills, attention skills, and behaviors—sometimes during preschool, sometimes in kindergarten.

The scholars sought out data on every aspect of temperament and behavior we recognize can affect performance in school—acting out, anxiety, aggression, lack of interpersonal skills, hyperactivity, lack of focus, et cetera.

“It took us three years to do this analysis, as the pattern slowly emerged,” recalled Duncan.

On the whole, IQ tests confirmed the strength of correlations that had been seen in other research: Combining math and reading together, early IQ had at best a 0.40 correlation with achievement in 3rd and 5th grade.

Attention ratings didn’t beat that—not even close. They only showed a 0.20 correlation with later achievement. And the real surprise was how poorly the behavior ratings predicted school success—that correlation topped out at 0.08.

“That is what surprised me the most,” confirmed Duncan.

What this means is that many kids who turned out to be very good students were still fidgety and misbehaving at age 5, while many of the kids who were well-behaved at age 5 didn’t turn into such good students.

Duncan's work was so surprising, in fact, that a special session was dedicated to reviewing his work during last year's biennial meeting of the Society for Research in Child Development.

University of Virginia professor David Grissmer was among the presenters at that session. He announced to his colleagues that he had repeatedly set out to prove how Duncan had got it wrong, using some of the very same long-term population studies that Duncan's team had used. Instead, Grissmer shamefacedly admitted that it turned out Duncan’s analysis was right. Behavior is just not the reliable prognosticator we imagine it to be. (And Duncan is now president of the Society for Research in Child Development.)

It’s tempting to imagine one could start with the 0.40 correlation of IQ tests, add the 0.20 correlation of attention-skill ratings, and top it off with a social-skills measure to jack the total up near a 0.70 correlation. But that’s not how it works. The various measures end up identifying the same well-behaved, precocious preschoolers, missing the children who blossom a year or two later. Used in combination—as they are by every private school around—behavioral screening is just as likely to subtract from the overall accuracy as add to it.

The rush to judge kids on their good behavior has been accelerated by the notion of emotional intelligence. In the 1995 publication of his book, Emotional Intelligence, Daniel Goleman theorized that EI could be more determinative of a child's academic success than the child's IQ. That idea is so popular that few have noticed it’s not supported by the data—the theory has proven to be a fallacy. In a recent meta-analysis of 57 studies, scholars at Florida International University determined that the correlation between emotional intelligence and school success tops out at 0.10. That was for teens; for 5 year olds, it’s even more problematic.

University of Delaware professor Carroll Izard has suggested that those looking for emotionally intelligent kids may have it exactly backward. In one test of emotional knowledge, Izard's team asked kids how someone would feel if his best friend moved away. The more verbal a child is, the more she’s able to score high on these tests—but verbal ability is also what drives early cognitive intelligence. So rather than triumphantly arguing that emotional intelligence supplants cognitive ability, Izzard's work suggests that early cognitive ability may increase emotional functioning.

In the United Kingdom, an extensive socio-emotional training program was implemented in 80 elementary schools; it had no impact on the kids’ self esteem or social skills, and it slightly drove the academic results down, not up.

The problem with judging kids on their personality is that at every age, different personality traits seem to matter. The traits that school administrators are looking for in those staged play groups—obedience, compliance—are only linked to kids' earliest school success. In kindergarten, agreeableness is indeed an asset—and the extroverts are the good students. However, researchers have found that, by 2nd or 3rd grade, extroversion is only half as important. By 6th grade, extraversion isn't an asset. Instead, it has an increasingly negative impact. By 8th grade, the best students are conscientious and often introverted. Agreeableness is no longer a factor in kids' success.

So if early IQ scores don’t predict academic success very well, and behavioral games do so even worse, then how are all the private schools and gifted programs to fill their ranks?

Schools should start by admitting the truth—stop describing admissions selection as “an art form” and be honest that it’s mostly random, save for an occasional bona fide prodigy who blows the top off these tests.

There’s a handful of private schools who don’t guarantee their lower school families entrance into 3rd grade. They weed out a few of the lowest performers to make sure those kids aren’t slowing the class down. And there’s a few public school gifted programs—such as in Seattle—that make the students retest every three years to be sure they still belong.

These steps are a dramatic improvement. Schools that screen kids a little later—even just two or three years later—might actually deserve the reputation they so seek.

Plus: Check out Book Beast, for more news on hot titles and authors and excerpts from the latest books.

Po Bronson and Ashley Merryman's New York Magazine articles on the science of children won the magazine journalism award from the American Association for the Advancement of Science, as well as the Clarion Award from the Association for Women in Communications. Their articles for Time Magazine won the award for outstanding journalism from the Council on Contemporary Families. Bronson has authored five books, including the #1 New York Times bestseller What Should I Do With My Life?