All forms of bilingual education are more effective than all forms of English-only instruction. FALSE. It is essential to look beyond labels and assess overall design, quality of staff and materials, and effective implementation. Cummins considers especially weak those early-exit transitional programs that provide primary-language instruction with some oral English until grades 2 or 3, then drop students into all-English classes taught by mainstream teachers unprepared to support bilingual students’ academic growth. He states that he would not hesitate to choose a monolingual program where the entire school was striving to partner with parents and community, build on students’ personal and cultural experiences, and promote critical literacy, over a bilingual program where there was no commitment to these goals.

The more time children are exposed to English, the more English they will learn. FALSE. Intuitively, this seems true. But there is no simple, linear relationship between amount of exposure to a second language and amount learned. What matters is not just quantity of time, but also the degree of engagement in learning: Students learn a second language through comprehensible input that they can connect to prior knowledge. Also, students learn best when "instruction is chunked into meaningful units, spread over larger periods, and when format is varied" (Gandara, 1997). Moreover, Hakuta (1998) notes the "time-on-task" theory of learning in general is no longer considered viable by scientists of learning: "The question of learning is not how much [time-on-task], but when and in what sequence."

It’s always best to use the child’s native language when introducing reading instruction. FALSE. According to Cummins, circumstances and resources should dictate whether to teach reading first in the child’s native language, English, or both simultaneously¹. Moreover, the NRC Committee on Preventing Reading Difficulties in Young Children recommends that the student’s level of oral English, along with available resources, be used to determine the language of initial reading instruction (see "Teaching Reading to English Language Learners").

Limited English proficient students learn English faster in English-only programs. FALSE. English learners are not randomly or equally assigned to different programs: Those entering bilingual education programs tend to have had less schooling, are from poorer families, and attend higher-poverty schools than those in all-English programs. Uncontrolled comparisons show all-English program students on average learn English faster than students in bilingual education programs, but these differences disappear when background factors are controlled for. When factors such as initial proficiency in English and the native language, prior schooling, and socioeconomic status are controlled for, students acquire English at similar rates regardless of program.

Canadian-French immersion proves that structured English immersion works. FALSE. The former is a successful, fully bilingual model which supports biliteracy and aims to develop language-majority students’ abilities in the minority language. Also, researchers note that these students are still far behind the native French-speaking comparison group after 2 years of monolingual L2 instruction, but catch up fully after 5 years -- three years after instruction in their native (English) language is introduced. (Lambert and Tucker, 1972, Cummins, 1998) As such, they argue this model actually provides better support for bilingual education approaches.

English language arts instruction should be delayed for several grades until students’ literacy in their primary language is established. FALSE. All well-designed bilingual programs have English language development, including literacy, built into the overall plan across grades. Some bilingual models develop primary language literacy first and delay English language arts instruction until oral English fluency is developed sufficiently either to phase in English reading instruction (e.g., "90/10" dual immersion; maintenance) or to transition students to English reading (e.g., late-exit transitional). Other models introduce English language arts instruction much sooner (e.g., those without bilingualism as a goal, such as early-exit transitional, or bilingual immersion). (Brisk, 1998, Cummins, 1998)

Young children learn second languages easily, and the younger the child, the more skilled he or she will be in acquiring a second language. FALSE. The impression that children learn languages faster than adults arises because a young child does not have to learn as much as an adult to achieve competence in communicating. However, research does not support these beliefs, particularly in learning more abstract, academic language skills. Other than in pronunciation, younger children are often at a disadvantage compared with older children and adults in learning second languages quickly and effectively because they don’t have access to prior knowledge, memory techniques and other learning strategies and cognitive skills. (McLaughlin, 1992; August and Hakuta, 1997).

Bilingual education in and of itself will elevate student achievement. FALSE. Native language use is an important but insufficient ingredient in promoting language-minority students’ academic success. Many elements of effective schooling must converge to foster success for language-minority students. (See August and Hakuta’s effective schools’ and classrooms’ attributes, in "English Language Acquisition and Academic Success: What Do We Know?" above; also see Brisk’s lists of quality school, curricular, and instructional characteristics, in briefing binder appendix.)

¹ Cummins and others note that because Spanish has a higher phoneme/grapheme correspondence (how you say it is how you write it) than English, it may be easier to introduce reading in Spanish first, if circumstances allow.