William J. Mathis: Education reforms: Everything important cannot be measured
Perhaps, the reformers got it wrong.
Attributed to Einstein, "Everything that can be measured is not important and everything important cannot be measured." In focusing on what is easily measured, rather than what is important, we fail to grasp the real problem. To be sure, tests measure reading and math reasonably well and we need to keep tests for that purpose. But that's only one part of education. Schools also teach children to get along with others, prepare young people for citizenship, encourage creativity, teach job and human skills, integrate communities, teach tolerance and co-operation, and generally prepare students to be contributing members of society. These things are not so easily measured.
Even if we limit ourselves to test scores, as a society, we misread them. That is, the low scores are strongly affected by circumstances outside the schools. Children coming from violent, economically challenged and drug addicted homes, as a group, are not going to do as well as their more fortunate classmates. As the family income gap between children has widened, the achievement gap has also widened.
A Stanford professor compared all the school districts in the nation using six different measures of socio-economic well-being and found that a stunning 70 percent of test scores could be predicted by these six factors. When the PARCC tests, which are used to test "college and career readiness" were compared with freshman grade point average, the tests only predicted between one and 16 percent of the GPA. What this means is that the tests do a better job of measuring socio-economic status than measuring schools.
This pattern has been solidly and consistently confirmed by a mountain of research since the famous Coleman report in 1966. It pointed to family and social problems rather than schools. So what did we do? We collected more data. We now have "data dashboards." Countless ads on the web tout this lucrative market and proclaim how people can "drill down," create interactive charts and visuals to provide "deep learning." They display all manner of things such as differences by ethnic group, technical education, graduation rate and a myriad of exotic esoterica. By all means, we need to continue to collect this important data. The problem is that we already know what the dashboard tells us. What it doesn't tell us is the nature of the real problems and how to correct them. First, we must look to those things outside the school that affect school performance. Second, in addition to hard data, we must use on-the-ground observations to see whether we provide legitimate opportunities to all children, whether the school is warm and inviting, and whether the curriculum is up to date and well-delivered.
By concentrating only on the easily measurable, we squeeze the life out of schools. We devalue, de-emphasize and defund things that lead to a better life, better schools and a better civilization.
Finally, it misses the most essential point. Parents want their children to grow and lead productive, happy lives and contribute to society. They want their children to practice civic virtue and have loving relationships. But these things are not easily measured by a test. "Everything that can be measured is not important and everything important cannot be measured."
William J. Mathis is managing director of the National Education Policy Center and vice-chair of the Vermont State Board of Education. The views expressed here are not necessarily those of organizations with which he is affiliated.
TALK TO US
If you'd like to leave a comment (or a tip or a question) about this story with the editors, please email us. We also welcome letters to the editor for publication; you can do that by filling out our letters form and submitting it to the newsroom.