Reporting Error Inflates Non-Bilingual Gains on Tests

Because thousands of students who are fluent in English mistakenly were classified as not fluent, test scores showing gains for immigrant students in the wake of a California ban on bilingual education were inflated, school officials said Wednesday.

California, home to more immigrants than any other state, has become a national bellwether in the debate over bilingual education, which was banned by state voters last year. The reporting error in a controversial test has only added to the debate.

Though statewide results of the test, the SAT-9, were not revealed until Wednesday, some school districts previously had released their scores.

Because some districts reported very high gains for immigrant students, gains well above those for their English-speaking peers, opponents of bilingual education had touted the results as evidence of the ban’s merits.

Harcourt Educational Measurement, which prepared the report on students’ scores, erroneously categorized some 200,000 students with English proficiency as having only limited proficiency in English, state schools Supt. Delaine Eastin said Wednesday, as statewide results were announced.

“This is a serious glitch, and we hope that Harcourt is going to move apace to get it fixed,” Eastin said.

One suburban San Diego school district in particular had been held up as an ideal test case for the English-immersion approach to teaching immigrant students. Oceanside Unified School District was the state’s only district to drop all non-English instruction, rather than seek waivers to allow some students to continue bilingual education.

On Wednesday, Oceanside school officials released revised scores, which were still being refined at day’s end to see how students learning English fared in comparison to their English-fluent peers.

While the immigrant students’ scores still showed strong improvement, the most spectacular gains fell substantially as a result of the revision–in some cases to a level nearly one-fourth of what previously had been reported.

In the lower grades, scores actually improved after the reporting was corrected. For instance, reading scores for 2nd graders rose 117 percent rather than the 100 percent previously reported.

For 3rd graders the scores rose at higher rates, and for 4th graders the scores remained the same.

For all other grades, rates of improvements fell as a result of the correction. The highest level of improvement in reading, which had been reported for 7th graders, fell from 475 percent to 200 percent.



Comments are closed.