Academic Performance Index

API Fallacy No. 1:
False Measures of Growth

WHY A FIRST DOWN IS A BETTER MEASURE OF PROGRESS THAN THE API

The API's ground rules are not as fair or as smart as football's. Both use threshholds to offer rewards, but in very different ways:

In football, you get four downs to move the ball ten yards. The referee measures that distance from your own starting point. If you start on your own 21 yard line, you get a first down when you move the ball past the 31 yard line.

The API has the equivalent of "yard lines" — four threshholds called quintile boundaries — but it uses them in a radically different way. When a school has a student whose improving score, in the course of two years of testing, crosses a boundary, the school wins credit toward a better API score. If enough students cross any of the boundaries, their school shows an improved API. If not enough students cross the boundaries — regardless of the quantitative improvement in their scores — their school does not show an improved API.

So if you're a student (A, at right) with a reading score at the 21st national percentile rank (or NPR), and you do a lot better the following year, advancing 18 points to the 39th NPR, your school gets no credit whatsoever for your progress, since you didn't cross the 40th NPR. Student B, however, in posting a modest two-point gain from the 39th to the 41st NPR, gets a big "win" for the school. It's an inaccurate way of measuring student progress. And it's little wonder that many principals mainly focus on helping students whose scores are within a few NPR points of the 20th, 40th, 60th, or 80th NPR.

If football worked the way the API worked, you'd only get a first down when you crossed the 20, 40, 60 or 80 yard line. A team starting at the 21st yard line would have to move the ball to the 41st yard line to get a first down: a 20 yard gain, a much more daunting task than advancing two yards, from the 39th to the 41st yard line, which, using the API system, would award them the same first down.

The inequality works in reverse, too. If the student's score on the math portion of the SAT-9 was at the 39th percentile in fourth grade, and at the 21st percentile in the fifth grade, the school's API is unaffected. If, however, the student's score drops by just two points from 61 to 59, it contributes to a drop in the school's API score.

A school can actually have its API jump dramatically, yet show a net decline in its students' SAT-9 scores. And a school's API can decline, when its students' SAT-9 scores have improved. The simple, stupid secret: growth only counts when students cross a quintile boundary year-to-year.

BACK TO "FLAWS" PAGE | BACK TO MAIN API PAGE