Error May Raise Test Scores in Oxnard

Statewide achievement test scores in the Oxnard Elementary School District may be higher than reported last week because of another scoring error, school officials announced Wednesday.

The discovery comes just one week after a San Antonio-based publisher admitted misclassifying about 300,000 students statewide as not fluent in English, which officials say probably inflated those districts’ scores.

Oxnard, which is a year-round school district, realized a mistake in its Stanford 9 scores after reading an article in The Times about a year-round district in Long Beach where the results were very likely scored incorrectly.

Harcourt Educational Measurement officials believe the scoring problem resulted from not using the correct norm, or average, for the districts.

The national norm, based on a representative sampling of students across the country, was supposed to have been lowered for year-round districts, because the test was given when those students had been in school fewer days than students in traditional schools.

But the norm was very likely not adjusted for either year-round district, possibly lowering the scores, according to Ed Slawski, a senior research scientist for Harcourt.

Slawski said Oxnard and Long Beach are the only two California districts that have reported this concern.

“We don’t have any reason to believe that this is a widespread problem,” he said.

Although Oxnard Assistant Supt. Connie Sharp said she was frustrated by the mistake, she said she was thrilled that her district’s students performed better than the scores reflected. Thirteen of the district’s 19 schools are operated year-round.

“Our principals have been really disappointed, because they really expected our scores to be better than they were before,” Sharp said. “So we’re kind of excited.”

She anticipates scores to rise by about three percentile points. Preliminary scores released by the Ventura County superintendent’s office last week showed Oxnard elementary students had scored below the national average across the board, though the results showed improvement from the previous school year.

For instance, fourth-graders scored at the 28th percentile in reading and the 30th percentile in math, and eighth-graders scored at the 36th percentile in reading and the 34th percentile in math in the 1998-99 school year.

Harcourt will rerun the scores, Slawski said, and then will mail the results to the district in about two weeks.

Simultaneously, the state Department of Education is reevaluating the scores in Long Beach and Oxnard, spokesman Doug Stone said Wednesday. After hearing about the error, Stone said, he checked with districts in Los Angeles, San Jose and San Diego, and did not find any other problems.

“We want to make sure that when we release the data sometime next week that we can stand by it and that it is as complete and accurate as possible,” Stone said.

Harcourt also agreed to reimburse the Oxnard district for the cost of mailing individual score reports to parents, though there wasn’t an estimate of how much money had been spent.

“It’s clear that it’s our error, so we won’t be charging the district anything,” Slawski said.

This is not the first snafu by the publisher, which administered the statewide achievement test to more than 4 million California public school students--including 96,000 in Ventura County--this spring.

First, the tests did not arrive on time. Then the scores were not released on time. They were scheduled to be posted on the Internet June 30, but the education department delayed the release because of the misclassification of hundreds of thousands of students as not being fluent in English.

Stone is worried the mistakes may affect the public’s perception of the test and its results.

“Unfortunately, the problems that have come to light in the past week have to a degree undermined some of the student progress that was reported,” he said.

Times staff writer Richard Lee Colvin contributed to this story.