Structured Immersion Falls Short of Expectations:
An Analysis of Clark (1999)

by Stephen Krashen
University of Southern California
and
Jeff McQuillan
Arizona State University

In a recent report, Clark (1999) presents data on oral language development in several California districts that use a sheltered English immersion approach for their limited English proficient children. The data are presented in terms of Terrell's proposed stages of oral language development (Terrell, 1982; Krashen and Terrell, 1983 [neither cited in Clark, 1999]):

• preproduction,
• early production,
• speech emergence, and
• intermediate fluency.

Clark does not provide explicit descriptions of the criteria districts used for classification in these categories, but Terrell describes them as follows:

• Preproduction: A "silent period" in which language acquirers can only respond nonverbally.
• Early production: Acquirers are able to respond using single words and short fixed phrases.
• Speech emergence: Students are able to produce more complex but often grammatically incorrect output.
• Intermediate fluency: Students are able to produce more complex discourse, which still will contain errors.

Clark also include an "advanced fluency" stage, but does not provide a description.

According to Clark (1999), attainment of the intermediate fluency stage is necessary before students are able to "modified, grade appropriate content" (termed Specially Designed Academic Instruction). This is not the same thing as mainstream: To begin to do mainstream academic work, children need to attain the advanced fluency level; full mainstream requires reaching an even higher level, fluent English proficient (pp. 10-11).

Clark presents data on oral language development in three districts in terms of the number of students at each level at one point in time, compared to the number of students at each level at a second, later point in time.

A problem with this approach is that it does not tell us how far a student went when he or she moved up from a level. For example, for Orange, 559 students were at preproduction in November of 1998. Of these, 413 moved up and 146 did not. But we can not tell how far the 413 moved. We can, however, conclude the following from Clark's data:

1. Many students stayed at the preproduction stage after one-half to one full academic year: Percent of students still at preproduction Orange: after one academic year: 26% Atwater: after one-half academic year: 29% Delano: after one-half academic year: 51%

2. After one year, a substantial number of students have not reached intermediate fluency, the minimum level thought to be necessary to understand "modified, grade-appropriate" special subject matter teaching.

The following table, based on Clark's data, shows that many students who were NOT ready for modified subject matter teaching at the start of the year were still not ready at the end of the period studied. In Orange, for example. after one year, 77% of those who were not at the intermediate fluency at the start of the year were still not there at the end of the year, even though most were beyond the beginning level at the start of the year. Seen another way, only 23% moved to intermediate fluency in one year, and most of those below that level at the start of the year were beyond the beginner range (at the start of the year, 65% of Orange LEP's were beyond the beginner level but not yet at intermediate fluency; early production = 792, speech emergence = 1116).

Percentage of LEP children NOT ready for sheltered subject matter teaching at start of the year who are still NOT ready:

"not ready" = has not attained intermediate fluency or higher
now = Orange - after one year; Atwater, Delano - after one-half year

Clark's data thus show us how hopeless the Proposition 227 requirement is. Proposition 227 requires LEP students to go from ground zero to mainstream level (advanced fluency or fluent English proficient) in one year. Even with a substantial head start, most students do not even make it to the much more modest level of being able to handle modified subject matter instruction after one year. (Note that in the Delano district, not one child moved to advanced fluency after one half year of English immersion; in Atwater, only 12 did, about 1.5% of the LEP population.)

Finally, one can conclude absolutely nothing from Clark's data on the efficacy of immersion compared to bilingual education: No comparison is made with oral language development under bilingual education.

To summarize, Clark's data only show us:

1. A substantial number of students appear to be making no measurable progress in oral English.
2. Overall progress in three districts is hopelessly short of the requirements of Proposition 227.

In addition, no comparison with oral language development in bilingual programs is presented. Thus, the data do not demonstrate the superiority of structured immersion.

References
Clark, K. 1998. Preliminary evaluation report: Orange Unified School District alternative education plan for LEP students. Stockton, CA: Clark Consulting Group.

Clark, K. 1999. From primary language instruction to English immersion: How five California districts made the switch. Washington, DC: READ (Institute for Research in English Acquisition and Development).

Krashen, S. and Terrell, T. 1983. The natural approach: Language acquisition in the classroom. Hayward, CA: Alemany Press.

Terrell, T. 1982. The natural approach to language teaching: An update. Modern Language Journal 66 (2): 121-32.