Numbers Game:
Challenging the Fallacies about Proposition 227

by James Crawford 

Five years after California voters passed Proposition 227, dismantling most bilingual education in the state, what has been the impact on student achievement? No one can say with certainty. That’s because no controlled scientific studies have yet focused on this question.

Nevertheless, raw test scores are routinely invoked by those who hope to vindicate the move toward English immersion programs. Such claims are unscientific at best. But that has not stopped them from being circulated by journalists – who seem to find these numbers irresistible.

In 2000, for example, the New York Times cited “striking rates” of improvement for English language learners on a standardized achievement test. It hailed the results as “a tentative affirmation of the vision of Ron K. Unz,” the sponsor of Proposition 227.

What the Times neglected to note was that year-to-year gains were striking for all groups of California students – rich and poor, white and minority, English-proficient and limited-English-proficient – as teachers became more familiar with the test and more adept at teaching to it.

The newspaper also ignored a relevant study by Professor Kenji Hakuta and colleagues at Stanford University. This analysis determined that patterns of achievement were virtually identical in schools that had retained bilingual education under the new law, those that had eliminated it, and those that had never offered it.

Soon the Los Angeles Times showcased still more encouraging news about Proposition 227. On a new California English Language Development Test (CELDT), it reported, “students in immersion programs were nearly three times as likely to score in the advanced or early advanced categories as students in bilingual programs.”

This sounded rather conclusive. Once again, however, a crucial bit of context was missing: the English immersion students were three times more likely than bilingual students to start out as advanced or early advanced in English. So the test results said nothing about the relative outcomes of the two programs.

Recently, the release of a second year of CELDT scores set off yet another rush to judgment. As reported by numerous media outlets, the percentage of English learners reaching the “advanced” and “early advanced” levels nearly tripled between 2001 and 2002 – from 11 percent to 32 percent. California’s top education official called the results “very exciting for our state.” The San Francisco Chronicle described them as “measurable evidence” that Proposition 227 “seems to be working.” 

Quite impressive, except for one small detail. This was an apples-and-oranges comparison. When tested in 2002, the students had received an additional year of English instruction. No wonder this same group scored higher than they did in 2001!

Even so, their gains were hardly cause for celebration. Only 11 percent moved up from beginning English – the lowest of five levels – while just 7 percent reached the highest category. Meanwhile, more than two-thirds of the students scored below advanced or early advanced after at least one year – and in most cases, several years – in California schools (see table below).

California English Language Development Test (CELDT)
results for 862,004 students
who took the test in 2 consecutive years
(percentage scoring at 5 levels of English proficiency)



41 30  19
21   8

   Source: California Department of Education
This is a far cry from what Proposition 227 promised. In the most effective sound-bite of the campaign, Ron Unz condemned bilingual education for its “95 percent annual failure rate” in teaching English. Thereupon the voters adopted his proposal to mandate all-English immersion programs “not normally intended to exceed one year.”

So it would now seem fair to ask: what is the annual rate of English acquisition under Proposition 227? According to the California Department of Education, just 7.8 percent of English language learners were “redesignated” as fully English proficient in 2002. That compares with 7.0 percent in 1998, the year before the initiative took effect. In other words, the rate is virtually unchanged.

Yet, despite the current focus on “standards and accountability,” this is one set of numbers that U.S. news media have chosen to overlook.

This article first appeared in The Bilingual Family Newsletter 20, no. 2 (Summer 2003), published by Multlingual Matters, Frankfurt Lodge, Clevedon Hall, Victoria Road, Clevedon England BS21 7HH.

COPYRIGHT NOTICE: Copyright © 2003 by James Crawford. All rights reserved. Except where otherwise marked, material from this web site may not be republished in any form and for any purpose – including course use, electronic reserves, and Internet postings – except by permission of the author at this email address or via PayPal links on this site.

SPECIAL NOTE TO STUDENTS: No permission is required to quote from or paraphrase articles from this site in term papers, dissertations, or other course work not intended for publication. Of course, appropriate bibliographical credit should be given to avoid plagiarism. For further information, see my permissions FAQ.

SPECIAL NOTE TO INSTRUCTORS: No permission is required to direct students to this page or to link to it from another web site. Please note, however, that reproducing multiple copies of copyrighted material without permission is a copyright infringement that could make individuals and their institutions liable to legal action.