Some schools to be able to wipe out AIMS scores over glitch
The Arizona Republic
Sept. 15, 2006

Pat Kossan

A glitch in scoring caused wild swings in the 2006 AIMS writing scores, and the state will allow some schools to wipe out the results.

State officials said the problems were concentrated in elementary grades and did not involve high schools, where students must pass the reading, writing and math AIMS exam to graduate.

Schools were eagerly awaiting the results of the investigation because the scores are used to help determine state school rankings, from "excelling"
through "failing," to be released in October. Arizona schools chief Tom Horne said any school negatively affected by the writing score problems would be permitted to file an appeal. If state officials agree, the scores would not be used. Horne said the state has not determined how many schools could be involved.

District officials began seeing wide swings in the writing scores as soon as they examined the AIMS results in June. Statewide scores, released in July, showed the same skewed results, with the percent passing in some grades plummeting and, in others, skyrocketing.

Here's what state officials say caused the problem:

Model essays are used to guide the people hired and trained by the state's testing company to score the essays. In 2005, the state used model essays selected by its former testing company. For 2006, the state's new company, CTB/McGraw-Hill, gathered a committee of Arizona teachers at its California scoring center, where they selected the 2006 model essays for all grades, a spokeswoman said.

State officials now say these models were very different from 2005. For example, the model essay used to determine whether sixth-graders passed the essay test was less advanced this year, so the percentage of sixth-graders passing the AIMS writing section skyrocketed 18 points. The model essay used to determine whether third-graders passed was far more advanced, so the percentage of them passing the test plummeted 20 points.

State officials said they were far more involved in selecting model essays used for the high school writing test because students must pass the exam to help earn their diplomas. More teachers selected more model papers to help guide scorers. Even so, the percentage of high school sophomores passing the AIMS writing section fell to 68 percent in 2006 from 74 percent in 2005.
State officials called that a small drop. The state changed the kind of essay high school students were expected to write this year from a persuasive one to a personal narrative, and state officials said that could account for the dip.

Horne said the 2005 and 2006 models should have been better coordinated and the department has taken steps to make sure it doesn't happen again.