posted by Joe Polk at 5:19 PM /
Seven local public high schools have been honored as among the best in the nation in a Newsweek magazine rating. The magazine identified more than 1,200 high schools as among the top 5 percent nationwide on the basis of a ratio that emphasizes certain advanced courses.
The local schools and their rankings are: Quaker Valley, 592; Mt. Lebanon, 879; Hampton, 912; Upper St. Clair, 913; Peters, 918; Allderdice in Pittsburgh, 1,183; and Fox Chapel Area, 1,253.
The formula divides the number of Advanced Placement, International Baccalaureate and Cambridge tests taken by all students at the school in 2006 by the number of graduating seniors.
Link: www.postgazette.com/pg/07144/788718-100.stm
The local schools and their rankings are: Quaker Valley, 592; Mt. Lebanon, 879; Hampton, 912; Upper St. Clair, 913; Peters, 918; Allderdice in Pittsburgh, 1,183; and Fox Chapel Area, 1,253.
The formula divides the number of Advanced Placement, International Baccalaureate and Cambridge tests taken by all students at the school in 2006 by the number of graduating seniors.
Link: www.postgazette.com/pg/07144/788718-100.stm
Labels: high school, school district
6 Comments:
I think this is a dumb and limited way to rate high schools. What does this really tell us?
It's worse than idiotic. My colleague Ed Felten, who teaches computer science at Princetion, explains (as he does every year) why his fictional "Monkey High" deserves to be ranked number one according to this method.
In case you're wondering: a year ago, Mt. Lebanon was ranked #988.
If Mt Lebanon wasn't ranked high, people would be screaming that we pay too much tax for the education we get - of course since we are ranked high, people have to find something else to complain about, so they pick apart the ranking system.
If you read Ed Felten's commentary (and the comments that follow it), you'll see that no one is picking apart rankings because they're looking to complain about their schools being ranked highly. The problem is that the method simply tells you nothing, and because it tells you nothing, the rankings are affirmatively misleading. In fact, one of the more egregious weaknesses of the methodology is that the authors deliberately exclude many of highest-performing high schools in the country!
To really see what's wrong with the rankings, click through the Felton's link to his original 2005 "Monkey High" post, and read down to the analysis of Oxnard High in California.
America's obsession with school rankings comes from the fact that they're self-fullfilling prophesies: At the college and grad-school level, the highly ranked schools and departments attract the best faculty and the best students, which increases their rankings more, etc.
A similar thing happens at a local level with geographically based school districts: Districts with high metrics (e.g. test scores) attract more people, increasing the market value of their property, so that only relatively wealthy people can afford to live there. This biases the population toward college-educated, high-achieving homeowners, and biases the student population toward college-bound, high-achieving kids, raising the reputation of the school, by most measures.
I would argue that, if Newsweek continues publishing these rankings, the schools in the rankings that do poorly by other measures will improve.
The problem with statistics isn't so much how easily they can be manipulated to tell a story, intentionally or otherwise. (Though that is a problem, as Mr. Clemens famously noted.)
The problem with statistics is how much tractability they gain in the mind of the layperson when they are a) widely disseminated and b) psuedo-scientificky enough to seem well-reasoned and therefore sound.
Everyone knows the 90% unused brain capacity thing, the 51% of people in the U.S. are female thing, and the 66.7% of all statistics being made up thing. :)
My point is, yes, Newsweek sucks at the Maths. But if well-meaning (if un-mathy themselves) Newsweek subscribers move to Lebo as a result, well, I won't complain.
p.s. - Did you know that 40% of all sick days taken by workers in the U.S. are on Mondays and Fridays? Psh! Slackers looking for three day weekends, I say!
Post a Comment
<< Home