Re: Fw: Bilingual classes better than English-only approach
- Here is a URL for the 23 Jan. 2004 JHU press release on this item:
The full report "Effective Reading Programs for English Language
Learners: A Best-Evidence Synthesis" is available at:
A critique by Stephen Krashen is appended below (there is apparently
no public archive of the list it appeared on, multied-l, for me to
insert a link to). I repost it to this list as I think that the
original study and his comments may be useful in disaggregating
elements of learning to read in bilingual classrooms. Approaching
this as someone from outside of the field I'm aware there probably is
a lot already that does that.
If it doesn't already exist, though, it would I think also be helpful
to have a layman's summary of the elements of literacy (in this case
at the primary school level) discussions for use in diverse settings
worldwide where bilingual education is either not well understood or
----- Original Message -----
From: Stephen Krashen
Sent: Saturday, January 31, 2004 8:01 AM
Subject: Recent Slavin and Cheung
Some comments on Slavin and Cheung
Slavin and Cheung's recent report "Effective reading programs for
English language learners" is a survey of the professional literature
that concludes that bilingual education is effective, that bilingual
students are better off learning to read in both languages at the
same time, that systematic phonics will help English learners learn
to read, and that direct vocabulary instruction is helpful.
In my view, their data does not support (or disconfirm) the last
1. Paired bilingual approach to reading
Slavin and Cheung conclude that "many of the studies with the
strongest positive effects for English language learners used
a 'paired bilingual approach,' in which children were taught reading
in both English and the native language at different times each day
from the beginning of their schooling" (p. 19). They list only two
studies in which the paired bilingual approach is compared to first
teaching reading in the first language: The El Paso study (Gersten
and Woodward, 1995) and the McAllen study (Pena-Hughes and Solis,
1980). The problem is that there were a lot of other differences
between the two programs in each of these studies: Paired reading was
only one of them.
The El Paso bilingual immersion program had used whole language and
sheltered subject matter teaching, and provided L1 instruction were
it counted the most, in areas that were the most cognitively
demanding. The "regular" bilingual program was phonics-based (see
below) and did not have sheltered classes. The McAllen "paired
reading" program was compared to a terrible bilingual program that
featured concurrent translation and inconsistent bilingual
It is doubtful that paired bilingual reading was the crucial element
in these studies, and these are the only direct comparisons of paired
versus sequential reading instruction. The research does not help us
decide between these options.
2. Slavin and Cheung conclude that programs emphasizing systematic
phonics are best for English learners, a result they say parallels
the research results for native speakers.
The evidence supporting systematic phonics for native speakers is
faulty, and has been criticized several times. None of these
criticisms is mentioned in Slavin and Cheung. They do not consider,
for example, Elaine Garan's analysis of the National Reading Panel
report, in which she shows that studies claiming to show the
superiority of intensive phonics only show a superiority for tests in
which children read words in isolation. The impact of intensive
phonics on reading comprehension tests given after grade 1 is
Slavin and Cheung present several sets of studies that, they claim,
show that systematic intensive phonics is effective for second
language acquirers. One set consists of studies of Success for All.
There are two problems with their conclusions: Success for All
consists of much more than phonics. It also includes real reading, a
more likely cause, in my opinion, of reading success. In addition,
one must ask what Success for All was compared to. All we know is
that it was compared to something. Nearly all of the studies of
Success for All presented by Slavin and Cheung are unpublished.
Another set of studies consists of comparisons of Direct Instruction
with "regular" instruction. In these studies, children who experience
Direct Instruction in reading do very well on tests of reading words
in isolation, but do not do nearly as well on tests that involve
actual texts. There is, in other words, a "large discrepancy between
decoding skills . . . and reading comprehension scores . . . ."
(Direct Instruction advocates Wes Becker and Russell Gersten, in an
article published originally in 1982 in the American Educational
Research Journal and reprinted in the Journal of Direct Instruction
in 2001). This pattern of high scores on decoding tests and lower
scores on reading tests also appears quite a bit in the Success for
All research literature.)
Also, Direct Instruction has only been compared to other skill-based
approaches. A number of studies show that students in programs that
emphasize free voluntary reading outperform those in traditional
skill-based instruction on tests of reading comprehension if the free
reading program is allowed to run for a sufficient length of time (an
academic year). Readers do at least as well as traditionally taught
students in shorter-term programs. Direct Instruction has never been
compared to these kinds of "book flood"
Slavin and Cheung regard Stuart's Jolly Phonics study as evidence for
the superiority of phonics. This study was actually a comparison
between BIg Books and Jolly Phonics for kindergarten children, most
of whom spoke English as a second language. The latter was a phonic
awareness/phonics program. Big Books involved the use of large size
books, but it is not clear that much real reading was included in
this program, or that there was a great deal of focus on meaning.
Here is the description of the advice given to teachers of the Big
"Teachers were asked to spend time on word level work, that is, to
emphasize words and letters, by drawing children's attention to
written words in the text, and talking about the letters in words.
Work with letters should involve introduction to their names and
sounds, and children should be encouraged to notice and learn words
and letters in the classroom environment. Activities to foster word
and letter learning were discussed: the teachers were already using
many imaginative and fun activities to these ends, such as having a
collection of handbags, each with a different letter, containing
small interesting objects whose names began with the sound of the
There was very little additional description of the Big Books program.
Jolly Phonics children were indeed better on tests of phonemic
awareness and reading words in isolation and did somewhat better on a
test of reading comprehension. As usual the advantage for tests of
decoding was much larger than the effect on the test of reading. But
this is apparently a comparison of more versus less skills, not more
versus less reading.
3. Slavin and Cheung found only two studies of vocabulary
instruction, and concluded that "direct teaching of English
vocabulary can help the reading performance of ELLs, " (p. 41), even
though in one study after two years experimental students did not
make significant gains on the Peabody Vocabulary test. Slavin and
Cheung do not mention the extensive research showing supporting the
hypothesis that vocabulary is acquired by reading (Krashen, 1993).
To summarize: Slavin and Cheung's conclusions in favor of paired
reading and systematic phonics are based on studies in which many
other factors were present, and in which comparison group treatments
are unclear. Their conclusions on the direct teaching of vocabulary
are based on two studies, one of which produced one important
negative result. In my opinion, their review provides no evidence
supporting paired reading, systematic phonics or the direct teaching
of vocabulary, nor does it provide counterevidence. Others have
reviewed the professional literature and have concluded that the
phonics instruction is highly limited, and that vocabulary emerges as
a result of reading. These reviews are not mentioned.
References not available in Slavin and Cheung:
Garen, E. 2002. Resisting Reading Mandates. Heinemann.
Krashen, S. 1993. The Power of Reading. Libraries Unlimited.
Slavin and Cheung's paper: "Effective Reading Programs for English
Language Learners: A Best-Evidence Synthesis:"