Standardized Tests Actually Inflate Scores

That is exactly what Mike Antonucci claims, but without supporting data. Of course, he has no supporting data because most politicians rail against standardizd "fill in the bubbles" testing without thinking about the implication if they went with another form. Writes Antonucci:

Every one of the eight presidential candidates to address the National Education Association Representative Assembly made a derogatory reference to standardized tests. At least five - possibly more - railed at "fill in the bubble" tests instead of more comprehensive assessments.

There is no question about it. A "fill-in-the-bubble" test is an insufficient method of assessing a student's knowledge and skills. But to whose advantage is that? One would think that the Educational Testing Service alone benefits from standardized tests.

NEA and the presidential candidates assume that alternative forms of assessment would demonstrate students, teachers and schools are performing better than standardized tests would indicate. That assumption is a statement of faith and not of empiricism.

Any question on a fill-in-the-bubble test provides all the data necessary to come up with the correct answer. Students are then supplied with four or five possible responses. By their very nature, standardized tests inflate the scores of students on the low end of the scale. The only students who score lower than 20 percent on a fill-in-the-bubble test are victims of bad luck, since entirely random responses should raise you at least that high.

Just the appearance of a correct answer printed on a test booklet should increase scores across-the-board. Some percentage of students who cannot correctly answer the question "Who was President of the United States during the Civil War?" with no further prompting, will no doubt choose the correct answer when it is placed next to George Washington, Woodrow Wilson and Bill Clinton. There is good reason to believe that scores would plummet if tests were "fill in the blank" instead of "fill in the bubble."

The logic of that last paragraph is undeniable. Unless a test has a "guessing" deduction, that is, you receive a greater penalty for a wrong answer than for leaving a question blank, a random guess on a four distractor multiple choice exam has a 25 percent chance to be a correct choice. Given that most test have obviously wrong distractors, a moderately skilled test taker can improve their chances. Of course a guessing deduction type test will not sit well with the educational "elite."

Antonucci goes on to talk about the benefit of standardized testing when it comes to scoring:

So why would educators and their political allies criticize measurements that cast them in a better light than the alternatives? It seems silly to put it this way, but it's because standardized tests are standardized. A machine checks the answers. The correct answer in Utah is the correct answer in Alabama. Not much interpretation is needed for parents, journalists and the public. There is little avenue to appeal to subjectivity.

All alternative assessments require an evaluator. Evaluators are not, and cannot be, standardized. They may be influenced, properly or improperly, by whether they taught the student, the student's performance on previous tests, the student's background and circumstances, the evaluator's view of the curriculum, the test, and the specific question being asked, and hundreds of other factors.

But here is another issue--money.

A machine is expensive, but you only need one or two machines to grade tens of thousands of tests in a single day. Program the machine, put a stack of tests into it and let it run. You don't need particularly skilled labor to do it. The machine doesn't care if it is 3:00am or 3:00pm, it doesn't care if it must work 24/7 for three weeks. A machine doesn't make mistakes because it is fatigued or because their spouse yelled at them this morning or their son or daughter is hurt, sick or something.

Human evaluators have all of those distractions, plus they want to get paid. Human evaluators would be astromonically more expensive than anyone could even predict. Ask a teacher how long they spend grading tests for 150 students. Now imagine having to do that for tens of thousands of students.

Evaluation is always a touchy subject and unfortunately, it must be done. The schools demand it, the parents demand it and the taxpayers demand it.