The other day, I posted on Philissa Cramer’s report on the Quinnipiac poll, and the finding that Joel Klein’s approval rating had fallen from 44% in February to 37% in March. I said that Philissa’s graph was misleading, and that the decline in Klein’s approval among registered voters in New York City was not statistically significant.
I was wrong.
Thanks to Leonie Haimson for obliging me to go back to my calculations and find the mistake I made. (For the two people who might care, it involved substituting the margin of error, rather than the standard error, for the percentages into the formula estimating the standard error of the difference between the two percentages.)
So no, Leonie, you’re not a statistical idiot. I get to wear that hat today.
The more substantive interpretation is that support for how Joel Klein is handling his job as Chancellor really did seem to fall from February to March. We’ll have to wait until the next poll to see if this decline persists.