by David C. Berliner on September 14,2009
There is a growing movement in the U.S. to abandon high-stakes tests because they don’t work as anticipated and are costly. I agree, but hope that we don’t throw out the need for accountability along with the high-stakes bathwater.
Before No Child Left Behind (NCLB) became law, Audrey Amrein
and I discussed the dangers of high-stakes testing. We found that high-stakes high school exit exams did not improve scores on other tests such as the SAT or NAEP tests, and contributed to higher drop out rates. We also described the corruption that invariably occurs when an indicator of any kind takes on too much value. Both the indicator (test scores, stock prices, return on investment) and those who work with it are frequently corrupted. The 1200+ years of the Chinese civil service exams, and final exams at all three US Military academies are high-stakes examinations. Yet cheating by candidates was common despite the penalties of death and dishonor associated with such cheating. When indictors take on undue value people too often engage in morally questionable or reprehensible activities. States, schools, and teachers act similarly when faced with high-stakes exit exams.
Then came NCLB with mandatory high-stakes tests for all states and schools. With my colleagues Sharon Nichols and Gene Glass we showed that even though scores on state high-stakes tests were going up, scores on the National Assessment of Educational Progress
(NAEP) were not rising as expected. We also found that the pressure within each state for achievement was correlated zero with gains in NAEP reading and mathematics test scores. Thus we negated a basic premise of NCLB, namely, that if pressure were exerted so that teachers and students would work harder, achievement would rise.
Then Sharon Nichols and I, in the book Collateral Damage
, showed why NCLB is not working and why it cannot work. We documented how schools, under pressure to achieve, dump low performing children from the schools; or arrange for absences and suspensions on test days; or move children around from school to school so their scores will not count; or they drill, and drill yet again, on items suspiciously like those that are on the tests; and so forth. We found it hard to blame educators for a little loose record keeping, a little fudging of the data, a little more practice on items close to those that are on the tests, and for designing tests with easier items when their professionalism is undermined, their jobs are at stake, and they are forced to engage in a fruitless attempt to meet unreasonable expectations about student improvement. The Bush administration designed an accountability system perfectly suitable for corrupting the educators of our nation.
We said in our book that NCLB would not work as planned and that one of its terrible side effects would be to narrow the curriculum. We were right. Now, in fall 2009, school accountability systems based on high-stakes have proven to have no or negative effects on the achievement and the attitudes of children, and they have proven costly. Thus there is every reason to call for a moratorium on high-stakes testing in America. That’s what we asked for at the conclusion of our book and the case for a moratorium is even stronger now.
I hope that the Obama administration learns that there are alternative accountability systems that could work and are cheaper to administer. It is time to admit our nation got it wrong and must start over.