Testing the Testers
The SAT has long been a thorn in the side of students, parents, and teachers everywhere. At some point it became the standard for establishing academic potential, and we’ve been forced to deal with it ever since.
It’s almost too easy to loathe the SAT and its administrative body, the College Board: they rake in billions in revenue for providing an assessment that is of debatable value; they have helped create a mindset and industry around the idea of “test prep”; and the College Board has positioned itself as a significant voice in education policy. Worst of all is that, at their heart, they are a secretive entity accountable to no one.
Which makes stories like this all the sweeter.
A smart high school student, Milo Beckman, had a hypothesis about the essay component of the SAT: he thought that longer essays earned higher scores, independent of quality. So he took a poll of his classmates, correlated the length of their essays with their eventual scores, and ran a regression analysis on the data. The results?
Milo says out of 115 samples, longer essays almost always garnered higher scores.
“The probability that such a strong correlation would happen by chance is 10 to the negative 18th. So 00000 …18 zeros and then (an) 18. Which is zero,” he said.
And Milo’s hypothesis seems in line with the opinions of some other prominent SAT critics.
Maybe these important exams are being so closely examined?