The Junior Meritocracy in New York Magazine questions the validity of IQ testing of four year olds, which are used to determine entrance to the Gifted and Talented schools (via Joanne Jacobs and virago).
The article had some fascinating information about why testing IQs of 4 years is silly. The numbers are just too fluid.
In 2006, David Lohman, a psychologist at the University of Iowa,
co-authored a paper called “Gifted Today but Not Tomorrow?” in the Journal for the Education of the Gifted,
demonstrating just how labile “giftedness” is. It notes that only 45
percent of the kids who scored 130 or above on the Stanford-Binet would
do so on another, similar IQ test at the same point in time. Combine
this with the instability of 4-year-old IQs, and it becomes pretty
clear that judgments about giftedness should be an ongoing affair,
rather than a fateful determination made at one arbitrary moment in
time. I wrote to Lohman and asked what percentage of 4-year-olds who
scored 130 or above would do so again as 17-year-olds. He answered with
a careful regression analysis: about 25 percent.
There are other problems. Parents coach their kids on the test. The article quotes Nicholas Lemann, who says that IQs are also heavily influenced by socio-economic status. Some kids are better test takers, than others. Some just love performing; others don't give a crap.
Four-year-olds, no matter how smart and delightful they may be, have
obvious limits as test takers. Many, especially boys, can’t sit still
for the full duration of an exam; others can’t stay awake or
concentrate for that long, choosing at some catastrophic point to crawl
under their desks and give up. Nor is the context in which these tests
are administered exactly relaxing for young children. Both IQ tests
require that they sit alone in a room with a tester they probably
haven’t seen before. In the case of the WPPSI, the tester often isn’t
allowed to prompt the children to give more complete answers, even if
it’s clear they’re capable of delivering them (and would score better
if they did). In the case of the OLSAT, the testers can’t even repeat
I'm resisting the urge to tell some anecdotal stories here. Let's just say that we know of a child named Ethan. And Ethan patiently took a test for a few minutes and scored off the charts on that section. Like 99.9th percentile. But let's just after finishing the first section, he decided he was terrified of the tester, he wanted him mom, and he was scared of the change in the routine, so he spent the rest of the test time running around the conference table and turning the lights on and off in the room. Final IQ score was similar to vegetative material.
In Outliers, Gladwell discusses research that finds that having a super, high IQ doesn't really mean all the much. It doesn't predict future success or happiness. One does need a certain base level of intelligence, like 120 or something, but after that, it's all gravy.
The author concludes that we should use the Marshmallow Test to predict future outcomes instead.