After the first round of standardized testing using a new progress scale, Juneau School District students turned out seemingly satisfying scores, relative to state results.
Superintendent Mark Miller said the Alaska Measures of Progress (AMP) preliminary test results, released Wednesday, show English language arts and mathematics scores for every grade level are at or above the state average. Between half to two-thirds of the students partially met standards.
AMP testing, created by the Kansas-based Achievement and Assessment Institute, replaced the Standards Based Assessment. Schools first administered it in Spring 2015 to about 70,300 third through 10th graders in Alaska.
A statewide projection in August estimated only one-third of Alaska’s students were rated proficient in math and English.
While the preliminary report Miller now has shows better performance for students at the local level, he said teachers and administrators won’t rely heavily on it to direct them in instruction.
“On the other hand this is information that the media uses because it’s fairly easy to understand,” Miller said. “It’s the only test taken by all the students statewide. From a public perspective, a lot of public opinion is formed on the basis of this test.”
What that “information” is exactly, as far as a precise number of students who met standards and a break down of school-by-school analysis, is still on hold.
The test, which is not a pass-fail exam, assesses each students’ personal growth and skill level. Results do not affect grades.
Education Commissioner Mike Hanley said it’s that extra data analysis of ability, such as reading for information compared to reading for literature, that helps teachers understand where students need greater help or could face challenges.
Unfortunately, those results were not properly translated.
“(The results) are not inaccurate, but they’re not what we’ve asked for,” Hanley said.
The group of educators statewide who set proficiency levels were surprised by results in the reading portion. Reading abilities for literature versus information, or fiction versus non-fiction, were unexpectedly similar, Hanley said. It was expected one area would differ significantly from the other.
This was the red flag that the assessment institute might have used an inappropriate scale.
“We continued to find issues and glitches that we weren’t happy about,” Hanley said, adding that he understands there were upwards of 160,000 results to process, but that it was “still not an excuse.”
The education department uses $25 million of allocated funds to pay for the assessment institute’s services as part of a five-year contract.
This questionable scaling issue wasn’t the only matter the commissioner and superintendents were concerned about.
Miller said, considering costs, he’s not sure why test results were late. Students took the electronically administered tests March 30 and May 1. Instantly, the assessment institute had its cyber-hands on them, but it had to wait for an approved grading scale from the state school board.
With data and an approved grading scale in hand by Oct. 9, results were expected within a 48-hour window. That wasn’t the case.
“I called the (director) of the organization about the delivery and the quality and explained how this is not OK,” Hanley said. “It’s added anxiety nobody needs.”
Marianne Perie, the project director with the assessment institute, said people were justifiably upset and apologies were made.
“It was a bad judgement on our part to not recognize (the results) weren’t going to be as helpful,” Perie said, further clarifying that she wouldn’t call anything the institute did an “error” or “wrong,” just not perfect.
The assessment institute only handles general state assessments for one other state, Kansas, and Perie said hindsight being 20/20, the institute should have looked at similar first-year glitches.
However, Perie said Alaska has problems that are unique to this state. Student mobility was more frequent than expected, with at least 16 students in the testing window starting exams at one school and finishing at another. Then there were absence-related overrides that should have been made, but some students’ scores weren’t counted because they missed just one day in the testing window.
“We need to talk about that, this isn’t good,” Perie said. “We can’t let it take this long when everything is ready to go.”
Perie said the scores will be processed using a more accurate scale and new results with greater attention to sub-scores should be available Tuesday for superintendents statewide.
Until then, Miller said the results as they stand aren’t enough for him to comment on the district’s successes or shortcomings.
“I need to look at individual sub-scores and subsets of sub-scores and reports by schools,” Miller said.
• Contact reporter Paula Ann Solis at 523-2272 or at paula.solis@juneauempire.com.
