The suggestion is made that bilingual education at that time was nearly all Spanish ("they might hear English only 30 minutes a day"). According to one study of bilingual programs in California before 227, by the time the children were in grade 3, 75% of their subject matter instruction was in English, and it was 90% by grade 5 (Mitchell, Destino, and Karan, 1997).

The suggestion was made that bilingual education in California had typically utilized concurrent translation (which of course conflicts with the previous accusation of little use of English): One teacher interviewed said that immersion students learn English better: "If they know they're going to hear it in their own language, they don't listen as carefully" (p. 6). The model in used in California was the gradual exit model (Krashen and Biber, 1988), which does not use concurrent translation at all.

Interestingly, Lexington describes with enthusiastic approval the technique of front-loading lessons with information provided in the first language, the idea of using the first language to supply background information that makes input in English more comprehensible: "Before teaching a new concept, an educator might quickly summarize it in the primary language" (p. 10).

This is precisely the core idea underlying successful bilingual education, including the gradual exit model that was used in California. The difference is that Lexington prefers the use of this concept lesson by lesson, while bilingual programs typically provide this background knowledge over an extended period of time. But the underlying concept is the same.

The report asserts that Open Court has been shown to be successful, and that intensive phonics has been shown to be superior to whole language (p. 7,8). Torrance either ignores or is unaware of publications arguing that this is not true (Garan, 2002; Krashen, 2002, Coles, 2003).

The Lexington Institute is free to disagree with the findings of other writers and researchers, but they are not free to ignore them.

2