IQ tests for admission to kindergarten

They've gone too far again.

This is a New York Magazine article by Jennifer Senior about the admissions process to New York kindergartens, in which "almost every prestigious private elementary school in the city requires that prospective kindergartners take" an intelligence test. Senior gives examples of one requiring a high 98th percentile IQ for incoming kindergarteners. Admission to that preschool and completion through the third grade means automatic admission to the related prestigious high school. So it is not unexpected (though it is unreasonable) for a parent to think that this test may determine their 4-year-old's  entire academic success.
 
I believe there is a set of mental characteristics that are usefully described as general intelligence. I believe that some but not all of the common sense idea of high intelligence is captured in an IQ test. I believe we should test infants, toddlers, and preschoolers for all sorts of mental and physical characteristics. But to give an IQ test to 4-year-olds for admission to kindergarten is ludicrous.
 
How do I hate this? Let me count the ways.

• IQ is a ratio: it is the kid's test score kid divided by the average score of all kids that age. But little kids don't progress in a steady manner; they grow in fits and starts, and when one trait stops and starts varies a lot from kid to kid. A difference in development of 3 or 4 months in two 4-year-olds makes a huge difference in their computed IQs compared to some allegedly average 4-year-old development. So a kid who would be denied admission today might knock the top of the scale in 6 months. Or vice versa.
• Partly for that reason, IQ is not reliable for kids under about 8. IQ at 4 is not a very good predictor of IQ at 12.
• Even when they work, IQ tests only measure part of general intelligence.
• Even so, cognitive intelligence, which I value very highly, shouldn't be the only criterion for admission to any school. In general, you have to be smart enough, but above a couple of standard deviations in IQ, there is no sense making IQ differences an important part of school admissions. At that level, other personality factors matter more for school (or career, or marital) success than IQ.
• It's advancing a pernicious system. It's pushing the idea of  high-stakes testing further and further down. I predict SAT-like study courses for 3-year-olds with obsessive parents, and we will end up teaching them to bubble Scantron sheets instead of play with blocks.