Why Do We Test?
All tests should have reliability and validity. The first is the extent to which results would be likely to be repeated if the tests were run again. In other words, if a student passes a state test for reading, and we can assume that the student would pass the test again tomorrow, and next week, and a month from now, then the test is reliable. The second is a determination of whether the test is a good test. Simply put, if the student can pass the test without being a good reader, the test lacks validity.
Measuring student achievement through state tests is (arguably) valid and reliable. Test questions are developed from state standards that teachers use in their classes. Maybe the cut scores are too high. Maybe they’re too low. In any case, the questions were written to test student learning.
Using the tests to measure teacher quality is not valid. It is not what they were designed to do. It is also not reliable because teachers operate under different conditions. Even within the same school, there are differences between the classes that teachers get. Some teach honors students. Some teach remedial students. Some teach exclusively in classes that students have chosen. Different levels of motivation, preparedness, and home resources all contribute to student achievement.
Using tests designed for one thing to try to tell us something different is inherently flawed. Doing so is a willful commission of poor research methods and bad analysis. Oklahoma isn’t there yet. We have yet to develop rules and procedures for quantifying teacher effectiveness. When we do, I assume it will be messy. Florida messy.
Schools lack confidence in the results of these tests (and the agency that administers them). So does the public. If the scores are too high, they’re too easy. If they’re too low, expectations are unreasonable. If they’re adjusted to compensate for either of these things, we’re just tinkering with facts. Whether the agenda is someone wanting schools to look good or wanting them to look bad, these things can be manipulated at the state level. Applying test results to determine things like third grade retention, high school graduation, teacher effectiveness, and school letter grades.
And now, we don’t even have a testing company for 3-8 grade tests in Oklahoma. By this point in the school year, district test coordinators are usually heavy into planning the test process. With so many assessments online now, this requires managing computer labs and often disrupting classes. This Tulsa World article explains the hardships this will cause schools in more detail.
By the way, a clerical error? That’ why April is going to be complete chaos? I hope the public, legislature, and State Board of Education remember that. In six months, when students are taking more tests than usual, and learning is more disrupted than ever, there will be some complaints. The list of frustrated superintendents is likely to surpass the horde of 300 from earlier this month.
Oklahoma is poised to spend $15 million on testing contracts next year. That’s an insane amount of money to commit to a process with results that are easy to dispute. It’s time for parents, schools, policy makers, and the public to have a long conversation about whether it’s worth the hassle and expense.