Original URL: http://www.rockymountainnews.com/drmn/election/article/0,1299,DRMN_36_1465272,00.html

Does Proposition 227 work? After analysis, debate still rages

By Holly Yettick, Rocky Mountain News
October 8, 2002

LOS ANGELES - Depending on who's talking, Proposition 227 has been a failure or a triumph. It has changed the lives of immigrant children for the better, or it hasn't affected them at all.

Into this morass of information and misinformation came a ray of light last spring.

It was a 179-page report by the American Institutes of Research and West Ed commissioned by the California legislature. And it was among the most comprehensive and least biased evaluations so far of the 1998 ballot initiative that requires English learners to be taught in English.

But it still fails to lay to rest a debate that pits English-only education against native language instruction.

In the years since Proposition 227 passed, initiative financier Ron Unz has said that students who receive intense English
instruction have made bigger gains on the Stanford 9 standardized exam than students in bilingual education.

But the League of United Latin American Citizens, which opposed the initiative, used results from the same test to suggest that since Proposition 227, the gap between English learners and native speakers has grown.

The report by the American Institutes for Research casts doubt on both claims.

The report argues that Unz's analysis is statistically invalid because he uses national percentile scores that compare
California - where one-quarter of the students are English learners - with a national pool, where less than 2 percent of
students are English learners. 

LULAC also used national percentile ranks to analyze scores. But LULAC compared apples with oranges, the AIR report says. In 2001, LULAC compared English learners with fluent speakers. In 1998, LULAC mistakenly compared English learners with the overall-average scores.

To top it off, a California Department of Education consultant said there are problems with any analysis that looks at pre-2000 Stanford 9 scores for English learners.

"The data from 1999 is extremely terrible," Linda Lownes said.

In addition to debunking conclusions by supporters and opponents, AIR researchers came to conclusions of their own.

They found that the gap between English learners and English speakers has narrowed "slightly" since 1998 on the Stanford 9. This would appear to bolster the initiative's supporters.

But then researchers compared schools that had never adopted bilingual education (67 percent of California schools), schools that changed from bilingual education to teaching mostly in English (15 percent) and schools that preserved bilingual education (9 percent).

They found no evidence that one model worked better than another.