Academic Performance Index

Six Flaws in the Academic
Performance Index (API)

FLAW 1: THE FALLACY OF GROWTH

The API uses a weird ruler to measure growth. It only measures movement across four statistical boundary lines. And it's similar to a first-down play in football. If you don't move across a quintile boundary, you don't get credit for progress.

FLAW 2: THE API SHOULD MAKE EXPLICIT UNAVOIDABLE MEASUREMENT ERRORS, AND REVEAL THE CONFIDENCE INTERVALS ASSUMED.

When a school's year-to-year progress is measured, unavoidable measurement errors should be accounted for. Chance alone will account for some variation in the annual change. The question is how much variation and with what degree of confidence; or how imprecise can the API's change factor really be? The error margins and confidence intervals were only revealed to the public in July 2002, and legislators are aghast at this surprise. (See Orange County Register features for more on this controversy.)

FLAW 3: THE MORE DIVERSE A SCHOOL'S STUDENTS, THE HARDER IT IS FOR THE SCHOOL TO MEET ITS GROWTH TARGETS

If a school's students are demographically similar -- all of one ethnicity, all from lower income families, or all English-proficient -- the school is more likely to reach its API growth target. If a school has three subgroups, they all have to meet their targets in order for the school to be considered successful in meeting its growth targets. The more subgroups, the harder it gets.

FLAW 4: TEST SCORES ALONE WERE NEVER THE IDEAL DETERMINANTS

The API was supposed to combine "input" factors like student and teacher attendance. (For high schools, graduation rates were supposed to be included.) None of those factors were factored in ultimately because the guardians of the API wisely noted important inconsistencies in the ways that these data are collected by districts.

FLAW 5: STUDENTS TAKE THE TEST WHO CAN'T READ THE TEST

Students who are not yet capable of reading, writing, or understanding English when spoken, indeed took the SAT-9. If they were in their school district for one year or more, their scores were factored into the school's index. This means that students who score low due to low English comprehension, and students who score low due to their low level grasp of content knowledge and skills, are lumped together.

FLAW 6: READING AND MATH DIFFERENCES ARE DISGUISED
The API intentionally conceals differences between subject areas (math and reading, for example). If a school's students show higher scores in math and lower scores in reading, those meaningful differences disappear when they're combined into a single API factor.

BACK TO MAIN API PAGE