More Testing Errors

Status: 
Archived
Subject: 
K-12 Testing
Errors and flaws in tests continue to produce harmful consequences for students and their education. Problems have recently afflicted New York and Arizona. And in Chicago, rapidly rising scores suggest both possible flaws in the exams and too much teaching to the test.

 

o New York City families are complaining that delays in test results will prevent their children from attending their favored middle schools in the fall. Officials blamed increased testing volume for the results being delayed until August. Students whose scores from the previous year weren't high enough to gain them entrance to lotteries that determine admission to middle school were hoping to do better this year and use those scores. Now they can't. The scoring delays also meant that decisions on promoting students in grades 2, 5, and 7 were made based on interim pass/fail test results, since the final results will not be available in some cases until the next school year is under way. These delays only compound problems caused by relying so heavily on test scores to make these decisions, adding frustration to the underlying folly of these policies.

 

o When scores for the writing portion of the Arizona state test were released in July, they had changed so drastically from the year before that Superintendent Tom Horne called for an investigation. Horne asked CTB/McGraw-Hill to rescore the tests after an 18-point jump in sixth-grade writing scores and a 20-point decline in scores for third graders. Educators said the results call into question the reliability and utility of such tests. Joe O'Reilly, testing director for Arizona's Mesa Unified School District, said, "We want scores that reflect impact of instruction and student willingness to learn."

 

o In Chicago, scores on the Illinois Standards Achievement Test (ISAT) jumped dramatically, with about 10 percent more students passing each of the math and reading exams at grades 4, 5 and 8. The grade 8 math passing rate jumped from 33 to 66 percent - in large part because the cut-off score was reduced from the 67th to the 38th percentile. While Mayor Richard Daley labeled the gains "historic," critics argued the gains could be "illusory" and a case of score inflation. They pointed out that the test was new and may have been easier; open-ended items apparently were not included in efforts to equate previous tests to this year's exam, which could have led to easier cut-off scores; and students had more time to respond to the questions.

 

Chicago officials also promote teaching to the test. The district used a "benchmark" test administered several times during the year, made by ISAT manufacturer Harcourt, that is intended to inform teachers how students are likely to perform on the full exam. By narrowing the curriculum to what is tested, using the benchmark tests as a guidepost, and possibly through such practices as providing extra assistance to "bubble kids" (those near the passing score), scores can jump dramatically without any real increase in overall student learning.