I've been writing this blog for a long time - nearly four years. You'd think that
after all of the bad math I've written about, I must have reached the point where
I wouldn't be surprised at the sheer innumeracy of most people - even most supposedly
educated people. But alas for me, I'm a hopeless idealist. I just never quite
manage to absorb how clueless the average person is.
Today in the New York Times, there's an editorial which talks about
the difficulties faced by the children of immigrants. In the course of
their argument, they describe what they claim is the difference between
the academic performance of native-born versus immigrant children:
Whereas native-born children's language skills follow a bell
curve, immigrants' children were crowded in the lower ranks: More than
three-quarters of the sample scored below the 85th percentile in English
Scoring in the 85th percentile on a test means that you did better on that
test than 85 percent of the people who took it. So for the population as a
whole, 85% of the people who took it scored below the 85th percentile -
by definition. So, if the immigrant population were perfectly matched
with the population as a whole, then you'd expect more than 3/4s the
score below the 85th percentile.
As they reported it, the most reasonable conclusion would be that on the
whole, immigrant children do better than native-born children! The
population of test takers consists of native-born children and immigrant
children. (There's no third option - if you're going to school here, either
you were born here, or you weren't.) If 3/4s of immigrant children are scoring
85th percentile or below, then that means that more than 85% of
the non-immigrant children are scoring below 85th percentile.
I have no idea where they're getting their data. Nor do I have any idea of
what they thought they were saying. But what they actually said is a
mind-boggling stupid thing, and I can't imagine how anyone who had the most
cursory understanding of what it actually meant would miss the fact that
the statistic doesn't in any way, shape, or form support the statement it's
The people who write the editorials for the New York Times don't even
know what percentiles mean. It's appalling. It's worse that appalling - it's
an absolute disgrace.