Friday, March 13, 2015

Can the "Opt Out" Movement Succeed?

There's a movement growing across the nation. It's called "Opt Out," and it means, refusal to subject oneself or one's children to the rampant standardized testing that has gripped public schools. The tentacles of the accountability testing movement have reached into every quarter of America's public schools. And the audience for this information is composed of politicians attempting to bust unions, taxpayers hoping to replace high-salary teachers with low-salary teachers, and Realtors dodging red-lining laws while steering clients to the "best schools." Those urging parents and students to refuse to be tested cite the illegitimacy of these motives and the increasing amount of time for learning that is being given over to assessing learning.

At present, the Opt Out movement is small — a few thousand students in Colorado, several hundred in New Mexico, and smatterings of ad hoc parent groups in the East. Some might view these small numbers as no threat to the accountability assessment industry. But the threat is more serious than it appears. Politicians and others want to rank schools and school districts according to their test score averages. Or they want to compare teachers according to their test score gains (Value Added Measurement) and pressure the low scorers or worse. It only takes a modest amount of Opting Out to thwart these uses of the test data. If 10% of the parents at the school say "No" to the standardized test, how do the statisticians adjust or correct for those missing data? Which 10% opted out? The highest scorers? The lowest? A scattering of high and low scorers? And would any statistical sleight of hand to correct for "missing data" stand up in court against a teacher who was fired or a school that was taken over by the state for a "turn around"? I don't think so.

Gene V Glass
Arizona State University
National Education Policy Center
University of Colorado Boulder


The opinions expressed here are those of the authors and do not represent the official position of the National Education Policy Center, Arizona State University, nor the University of Colorado Boulder.

Tuesday, March 3, 2015

Management by Pinheads

Management by numbers proves to be management by pinheads. Nothing exposes so clearly the naïveté and lack of understanding of managers and boards as when they start believing that complex human organizations can be managed by numbers: quantitative goals and numerical quotas; and in the case of educating children, drop-out rates, attendance statistics, promotion percentages, or test scores.

Beverly Hall died yesterday (March 2, 2015) after having been diagnosed recently with stage 4 breast cancer. Beverly Hall had a distinguished career as an education administrator that led her to one of the most prestigious positions in the nation: Superintendent of the Atlanta, Georgia public schools. Her service there was recognized with several honors: National Superintendent of the Year in 2009 and, later, the Distinguished Contributions to Service award of the American Educational Research Association. Hall was celebrated far and wide as a new breed of administrator. In her own words, she was a “data driven” manager. Decisions often rested heavily on the test scores of the classes and schools of the people who worked under her. Atlanta schools were the jewel in the crown of Management by Numbers … until it all came crashing down.

Hall was ultimately indicted for overseeing the scheme that inflated test scores of thousands of students in an attempt to create the fiction that she had “turned around” a failing school system. The criminal case against Hall will obviously be dropped. To be sure, Hall was at least guilty of some pretty shady activities, and she put her subordinates in circumstances where equally shady dealings were required to maintain one’s employment (a la Michelle Rhee and the Washington D.C. school system). But equally guilty is an ethos permeating organizations of many different kinds that holds that the management of complex dealings with human beings can be reduced to simple arithmetic.

Take my local school district, for example: Scottsdale (AZ) Unified School District. Recent actions (February 10, 2015) by the school governing board have resulted in a system of quantitative goals and bonuses for the district superintendent. Among these goals are 1) maintaining a 90% graduation rate, 2) remaining in the top 10% of the state’s districts on whatever the state standardized test will be, 3) reducing the teacher turn-over rate by half a percentage point, and 4) ensuring that 85% of the district students meet the state benchmarks in reading and math. Exactly what the superintendent’s bonus will be if these goals are met has not been disclosed to the media.

Anyone who knows the first thing about how schools run will recognize immediately that numerical goals like the ones favored by the Scottsdale Unified school board can be “gamed.” Counsel a few students away from some of the more challenging courses, put out a few broad hints about grading scales in those tough science courses that are flunking too many kids, or give a little extra time on those state bench marking tests, and voila the goals are met and the bonus is in hand.

It’s not just in Atlanta where "gaming" management-by-numbers systems is common. It has happened over and over in education at all levels. In one infamous incident in the 1980s, school principals in Houston were caught erasing and correcting answer sheets in their office in order to receive hefty bonuses for achieving annual growth targets. But one might have expected the Scottsdale Unified governing board to know better. After all, Phoenix is where in 2014 the head of the Veterans Administration hospital was caught falsifying wait list data in order to meet numerical goals and receive a large monetary bonus. The VA head was fired; it was a scandal that attracted national attention.

In the social sciences there is something known as Campbell’s Law — attributed to a former colleague of mine, the social psychologist Donald T. Campbell. It goes as follows: “"The more any quantitative social indicator … is used for social decision-making, the more subject it will be to corruption pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor."








Gene V Glass
Arizona State University
National Education Policy Center
University of Colorado Boulder


The opinions expressed here are those of the authors and do not represent the official position of the National Education Policy Center, Arizona State University, nor the University of Colorado Boulder.