Tuesday, February 17, 2015

Making High-Stakes Tests Unusable for High Stakes Purposes

Thanks to numerous constituencies insisting on the end of high stakes testing, our chances of bringing an end to the madness are growing by the day.  Now if we only can get Ravitch on board, rather than continuing to support Alexander's Testing Lite . . .

In an effort to shore up conservative support to keep his presidential aspirations alive, Bobby Jindal in Louisiana is supporting students' right to not fill in any answers on the PARCC tests.  Here is a good synopsis of why this is a good idea and the implications if parents and students participate in opting out:
. . . there's an inherent logic in the request that a student's "sit and stare" zero not be averaged into classroom and student scores. Look at the two parts of the issue.

First, a parent and child's decision not to fill in the bubbles on a standardized test is part of a valued tradition of passive resistance. A school should not have the right to force any student to read the test questions and bubble in the answers if the student chooses not to.

Second, the rationale behind using standardized tests as an evaluative tool for teachers, schools and districts is that the tests are a valid reflection of student achievement, and averaging the student scores produces a valid picture of the group's achievement. It's a questionable assumption, but assuming it's true, including a student's zero score into the total would wipe out any vestige of validity the average score might have. Clearly, that zero doesn't reflect the achievement level of that student, and its inclusion would dramatically, and unfairly, lower the average score. The more students who take a zero, the lower the score would go.

For a district or the state to insist that a "sit and stare" zero be added to other scores makes a mockery of the testing process. The motive would be punitive, not educational: "If you can't make your students obey our order that every student takes the test, we'll punish you by artificially lowering your score."

There's one more important wrinkle here. If students are allowed to refuse to take the test, that also harms the validity of the results, even if their zeros aren't counted. Assuming average test scores actually have some validity, that only holds true if all, or nearly all, the students take the test. The average scores possibly would be valid if the test were given to a random sample of students, but students choosing to opt out has nothing to do with randomization. That means, if enough students opt out, the average scores and their use as a tool to evaluate teachers, schools and districts are rendered meaningless. That's a real problem for the powers-that-be who want to use the test results. But that's the strength of a good protest movement. A group of people who lack official authority to change a law or regulation can throw a monkey wrench into the works through the right kind of organized protest.

Bobby Jindal was for the Common Core standards before he was against them. I suspect his recent stand against the standards is all about his presidential ambitions. Nonetheless, he's correct when it comes to scoring the tests. If students purposely take a zero on the state test, their scores have to be thrown out.

1 comment:

  1. Anonymous8:57 PM

    Maybe somebody should do a study on which parents are opting their kids out and the correlation with SES.

    Abigail Shure

    ReplyDelete