Built to Fail
Today's high-stakes tests are designed to create failure. In 2014, the failure rate in Florida (score of 2 or less) was 43% for third-graders and 45% for 10th graders in the Reading exam. Similarly, we are consistently failing greater than 40% of our test takers in Mathematics, Science, and Writing. You can find these documents and more here. On top of that, there is no trend of improvement shown in the FCAT 2.0 data from 2011 to 2014. Is it plausible that our schools are that bad? Year-to-year SAT score comparisons do not indicate any change in reading ability over time, so were our schools always that bad? And why hasn't the testing improved our children's education as promised?
The simpler explanation is that the tests are bad, and not the schools. The established standards are inappropriate for grade level in the early grades, but even more significantly, the tests used are built to trick children.
Questions on these exams are deceptive, and frequently obtusely worded. Please see the Sample Tests page for examples! Consider the age of the children taking these exams. In many cases, also consider the computer skill and interpretation required. It should also be noted that a failing student requires re-testing which generates more money for the testing corporation, as they are paid per child. There is therefore a financial incentive to create tests that fail students.
Florida did not field test the 2015 Florida State Assessment exams, instead relying on field testing in Utah. In the state of Utah, approximately 60% of students failed the exams. (Does this mean that their schools are also that bad?) Due to differing socioeconomic conditions, Florida children are predicted to have an even-higher rate of failure. The Florida Department of Education (DOE) excuses these failure rates, and says the test results will be acceptable due to the application of a curve applied, for this year only, that ensures the same failure rate as last year. Ironically, the notion of passing only those children who meet the bar set by any sort of "standard" is lost when a known percentage of students are prescribed to fail.
Please also consider the demoralized mental state of a third grader coming out of such a test, regardless of his or her landing place on the curve. What must it feel like at that age to get more questions wrong than right? Teachers and parents report statements of 8-year olds who feel like "failures" and children who come home in tears.
The application of the curve described above is for this test year. However, the tests are constructed in a way such that, in every other year, a predetermined, prescribed percentage of children will fail. Note that, in future years, each question on every exam will be selected with a known, field-tested probability of success or failure. It is therefore possible to predetermine an overall rate of success or failure on the exam. Again, failing a known, predictable percentage of children is NOT the same as measuring them against a "standard", unless the standard means the top 80% of test-takers. For a complete description of the "psychometric" process the Florida DOE uses to select items, please see Sections 4.3 and 4.4 of the Test Development and Construction document on the Resources page.
Also, if the probability of passing is allowed to change from year to year, e.g. the items selected are not of equal difficulty, then other questions arise. Items could be selected based on field-tested probability data that are able to show improvement where improvement is wanted, failure where failure is wanted, all at the discretion of Florida DOE. As no one is allowed to see the tests before or after, it would be easy to design them to serve political or monetary gain.
Ultimately, these tests create a generation who hates school, becomes disengaged in the classroom, and gives up. The tests also fail their objective of assessing the amount of learning that takes place in a schoolyear.