|
A mom wrote me to tell me her son had scored 398 on one of the 5th grade
History SOL tests, 2 points below the cut off score, and she wanted to know
if I thought it would be okay to lie to him and tell him he passed. Her son
was one of the many youngsters who truly has performance anxiety when faced
with SOL tests. I know the state scoffs at the notion of young children
worrying about the tests but this mom and many others like her can assure
you that it is real and it is painful to watch one's children endure. Her
biggest concern was that her son really works hard to do well in school and
might take the failing score to heart and become discouraged. He worked so
hard this year and really did seem to finally be finding his niche in
school. A failing score, this mom worried, might knock him back off course. Well, I responded, do you think the score really describes your son's performance? Are you sure he has learned what he needs to know? Have you checked his work and asked him to explain his answers? Have you talked to his teacher and asked her professional opinion on your son's performance? Have you looked at his grades, his other test scores, and his schoolwork over time? She had done all those things and was confident that all was well with her son. So she asked me what I thought. Should she lie to her kid? I wanted to tell her to explain to her son that test scores are not really precise. If you took that test several times, you wouldn't get the exact same score every time; scores are expected to fluctuate around the true score. In fact, test experts identify a range within which your scores would be expected to fall if you took the test again and again. That range is called the standard error of measurement and each and every test has this. In layman's term, the standard error of measurement is that scoring range that is simply too close to call. Every SOL test's passing score has an identified numerical range which is supposed to be included and explained as part of your children's score reports, just like the plus-or- minus margin of error that is reported in opinion polls in your newspaper. The test her son failed has a plus or minus of about 20 (I couldn't be real precise because the state is several years behind on technical reporting of the tests). His score of 398 is well within the passing range for this test. I wanted her to explain to her son that SOL test scores also get it wrong by misclassifying 1 in 10 kids as failing when their true score is passing or as passing when their true score would have been failing. Misclassification is inevitable and stems from measurement error which is inherent in all tests. According the state's data, in the first three years of SOL testing, about half a million test scores misclassified our children. The tests are fallible; the scores and classifications based on them are fallible. But then I thought about that. It's too hard for the politicians to understand and accept that a single test score can't really do or be what we are asking them to do and be in Virginia. They prefer the "simplistic more-than-400 is pass and less-than-400 is fail." Simplistic and wrong. If it's too hard for those learned folks, what chance would an 11 year-old have to understand? And as for the education decision-makers in Virginia, they ignore what they know. They know scores are not precise. They know test experts warn against the way we are using test scores. They know because they publish the statistics that quantify these things and the accompanying narratives that explain them. They know and lie to the public when they tell us "the SOLs are working." I told her to tell her son that he was doing just fine. And to have a great sixth grade year. She's not the liar. It's the state that is lying to him. And to us. Mickey VanDerwerker |