Is There a Plateau Effect in Test Scores?
July 29, 2009
Naomi Chudowsky and Victor Chudowsky
Center on Education Policy
Move over, Mt. Everest! The Center on Education Policy has found no proof that the so-called "plateau effect" in testing is widespread. Originally used to describe Florida's 1977-1997 test score trajectory, the term "plateau effect" is now commonly used in policy discussions when increases in test scores appear to taper off after larger initial gains. CEP set out to expand on that lone Sunshine State study; they found 16 states with 55 test score proficiency "trend lines" between 1999 and 2008 that met their criteria: The test had to remain substantially unchanged; it had to be given for 6 to 10 years starting in 1999 or later; and the state had to have kept its cut scores unchanged over that period as well. Only 15 of the 55 trajectories (or 28 percent) showed conclusive signs of a plateau for the percentage of students who scored "proficient," while 21 showed steady increases and 19 zig-zagged. The most striking conclusion--appropriately labeled an "informed conjecture"--is that NCLB may have had a lot to do with gains in 20 of the 55 trend lines between 2003 and 2004--the year when NCLB really got rolling. It's important to remember that this report looks at trends for the percent of students reaching "proficient" every year, not average scores for all pupils. So it's still possible that students are, on average, hitting a plateau on the test but there are still more kids getting over the proficiency bar every year. Confused? Dig into it yourself here.