Was Jump in State Test Scores a Fluke? Students Held Back Not Tested in '04
Ohanian Comment: Note the slimy use of reform. Note the way kids are manipulated and maligned so that scores look better in the reform plan.
Buoyed last month by an apparent bounce in the state's standardized test scores, elected officials rushed to put themselves in the way of reflected glory.
Gov. Ruth Ann Minner released the test results in reading, writing and math at a State Board of Education meeting and in a written statement called the student improvement "remarkable."
U.S. Rep. Mike Castle, R-Del., issued his own statement, attributing the success to No Child Left Behind, the school accountability law he helped shepherd through Congress. And Democratic U.S. Sen. Tom Carper's press secretary made calls to remind people that, as governor, Carper reformed Delaware's public schools.
But what looked like rising test scores could be, in part, a fluke, a one-time bounce that next year could burst like a bubble and this year could be obscuring academic decline in some student groups. The inflation may be related to a state law that took effect two years ago to hold back students with low test scores, meaning this year's testing pool may have contained more higher-achieving students than in past years.
Nonetheless, both Minner and Education Secretary Valerie Woodruff said the improvement in test results is genuine.
The anomaly was so obvious in the 10th-grade scores of one school district, New Castle County Vo-Tech, that officials posted a disclaimer on the internal Web site in its three high schools. "What we said was we have no bragging rights, make no mistake," spokeswoman Kathy K. Demarest said. "It's important for the community to have all the information."
District superintendent Steven Godowsky said at first glance, 10th-grade test scores were startling: "Look at Howard. I wish we could do that over there." State test results show the percentage of 10th-graders meeting the reading standard at Howard High School of Technology this year soared to 80 percent from 56 percent last year. The percentage meeting the math standard jumped from 23 percent to 60 percent.
But behind the immediate test scores are layers of data that, peeled away, reveal an apples-to-oranges problem for Delaware.
"Since this year's 10th grade testing group was limited to a higher-performing student [level 2 or above], we could presume to expect higher overall test scores," the vo-tech district said in its memo. "This will be a one-time-only reduction in the number of students to whom the 10th grade DSTP tests are administered."
Godowsky and others said next year, when hundreds of kids held back in 2002 reach 10th grade and take those state tests, overall test results will be lower. "The only way that wouldn't happen is if those kids drop out of school," said Robert Smith, Milford superintendent.
At first blush, statewide scores went up this year. The percentage of 10th-graders meeting the math standard is 53 percent compared with last year's 45 percent; 71 percent met the reading standard, up from 66 percent last year. But how much of the increase is due to better student performance and how much to a stronger testing pool?
Woodruff insisted the rise in test scores this year, in all subjects, at all grades tested, is real, regardless of the testing pool.
Minner issued a written statement that said, "despite the fact that some students were retained, this year is no different than the last seven in that there has been a constant and continual improvement in test scores."
But Castle, a strong supporter of the No Child Left Behind accountability law and a former governor who called for improving Delaware schools, said this week, "Rest assured that as one who follows education progress in the state, I think we have to dig into this."
The federal education law rates every school in the country on how well students score on standardized tests. Under the law, testing data is reported by student subgroups based on such factors as race and income.
No other superintendents issued disclaimers about test scores, as Godowsky did. But the state's data shows the difference in this year's and last year's testing pools.
Compared with last year, 16 percent fewer special education students took the 10th-grade tests in March 2004, 12 percent fewer black students did and 10 percent fewer low-income students did. In Delaware and nationwide, these groups have scored less well on standardized tests than white, middle-class non-special education students.
Educators at the district level said it is important to know to what degree this year's test results may be inflated by the state's retention policy.
"We're still looking at our data to try to understand what it's telling us," said Andrew Hegedus, organization development director in the Christina schools. "Part of ... going through the analysis includes asking questions like this."
State retention data shows that, due to low test scores, in the spring of 2002 Christina held back 11.5 percent of its eighth-graders, compared with 5 percent the previous year. Otherwise, those students would have taken the 10th-grade test this spring. In 2002, the retention rate for eighth grade was 11.5 percent compared with 3.8 percent in 2001.
To help those who failed the reading or math test, most districts set up what they call readiness or transition academies for them in high schools.
"If they're retained in the eighth-grade," said Colonial Superintendent George Meney, "they're going to take some of the courses again they were not successful in; they're going to get some individual help in some areas.
"So, if you had a kid who did well in mathematics but didn't do well in language arts, then, their program would be more focused in one area. But the case is that kids who are not successful in one area are not successful in the other either."
Often, academy students are still not doing well even with a second year of eighth-grade work. State test results this year show in the transition academy at Newark High School, no repeat eighth-graders passed the math test. In Red Clay's Dickinson High School, 96 percent of those in the transition academy for math failed the test again.
Results at fifth and eighth grades may be skewed too. Starting in 2002, students in third and fifth grades had to repeat those grades if they scored poorly in reading. So, the weakest students from 2002 did not advance to this year's testing pool of fifth- and eighth-graders.
It is more difficult, however, to sort out whether those testing pools were stronger or weaker this year, in part because the state Department of Education refused to release the same detailed testing data it has given The News Journal in the past.
Without that data, it's impossible to determine how many students were retaking the fifth- and eighth-grade tests after going to summer school and repeating a grade. At those grades, repeat test takers are included in the test results, state officials said.
At the 10th grade, test results reflect only the scores of first-time test-takers, officials said. At that level students aren't forced to repeat a grade.
State Board of Education member Claibourne Smith said he's determined to sort out this year's results. One of two African-American board members, he's paid attention to the gap in scores between minority and low-income students and white and upper-income students. He said this year's 10th-grade data shows nearly a quarter of students tested fell behind relative to where they tested in eighth grade.
Woodruff insisted the same test data - which matches individual students' 10th-grade test results with their eighth-grade results - shows Delaware 10th-graders improved this year.
But Smith said students with the lowest scores on the eighth-grade math test wouldn't be taking the 10th-grade math test this year because they weren't promoted. Yet, some 20 percent of this year's 10th-grade math test takers scored ones. "I'm almost sure that there are kids who are falling backward," Smith said.
The News Journal
INDEX OF NCLB OUTRAGES