The Utility of Estimating Bilingualism and Native Language Literacy in the U.S. Immigrant Heritage Language Population UCLA June 18, 2012 Terrence G. Wiley.

The Utility of Estimating Bilingualism and Native Language Literacy

in the U.S. Immigrant Heritage Language Population

UCLA June 18, 2012

Terrence G. WileyCenter for Applied Linguistics &

Arizona State University

Overview

• This presentation elaborates on the utility of the American Community Survey (ACS) for studying the language and literacy characteristics of immigrant groups in the United States of America, and also compares the ACS to the U.S. Census.• Neither the Census nor ACS questionnaire are structured to capture the language and literacy skills of immigrant communities in as far as these surveys only collect information about respondents’ English oral language abilities and household data on languages other than English.

Overview Continued

• Direct, self-reported, and surrogate measures of literacy are discussed, with a proposal to use education level as surrogate for literacy. • Select HL subpopulations from the ACS raw microdata, to measure respondents’ bilingualism, native language literacy, and biliteracy. • When such new variables are used in analysis of HL immigrant communities, a more complex multilingual picture emerges than is presented normally in Census and ACS data products available to the public.• But first, a few background considerations about popular myths and assumptions that affect popular attitudes, scholarly assumptions, and the national instruments that are available.

Examples of Concerns about Literacy in the United States

• Literacy skills are declining• College students can’t write• A blue-ribbon report by finds numerous problems in adult

learning, particularly among speakers of languages other than English

• An adult literacy survey finds 90 million are deficient in literacy skills

Feel free to add to the deficiency list…

Why is there a perpetual literacy gap?

Expectations for Literacy

Perceived Literacy Gap

Actual Levels of Literacy over Time

Popular Myths about Language and Literacy in the United States

• The U.S. is an Anglophone Country• Illiteracy in the U.S. is high because many do not speak English• Illiteracy in the U.S. is high because immigrants don’t want to

study English• English is threatened by the sheer number of immigrant

languages• English literacy is the only literacy worth noting in the U.S.• English is a global language, therefore “Americans” don’t need

to learn other languages• The best way to promote literacy is through English immersion

(Wiley, 2005)

Historically, for some groups there has been a decline based on the prescription and imposition of English Only Policies:Consider the case of Cherokee:

• 1822 Sequoyah develops a writing system for Cherokee (Lepore, 2002)

• 1833 Three-in-five Cherokee were literate in their native language and one-in-five in English

• By 1852 Cherokee had better schools than most neighboring states• The overall Cherokee literacy rate reaches 90% during the 1850s• By 1906 the imposition of English-only policies carried out by the

U.S. government had devastating consequences on Cherokee literacy and biliteracy. “The loss of schools spelled the end of the widespread bilingual literacy that has distinguished Cherokees in the nineteenth century (Weinberg, 1995, p. 222).

• Hawaii educational polices yielded a similar fate (Wiley, 1998, 2005)

The Hegemony of Monolingualist PerspectivesWhen politicians, policy makers, teachers, and even scholars, and census designers, assume that nation-states and societies are normally or optimally monolingual, bilingualism/biliteracy and multilingualism/multiliteracies are seen as aberrant, with negative consequences, or just rendered as invisible.Many years ago, Bhatia (1983) noted four dominant assumptions about so-called monolingual societies:

1. Monolingualism is assumed to have a feeding affect with literacy

2. Multilingualism is assumed to have a bleeding affect3. Multilingual societies are assumed to have greater

communication problems4. The linguistic situation in so-called monolingual societies is

assumed to be too obvious to warrant further serious consideration

Monolingualists’ Strategic Confusion /Obfuscation

Question: To encourage democratic participation, given a large Spanish-speaking population, shouldn’t we have bilingual ballots?

Answer: There are numerous languages spoken in the U.S.. Therefore, we need a common language.

Statements of partial fact: E.g., 65% of those who speak other languages in the U.S. don’t speak English very well.

Scholarly Reductionism on Language ShiftOnly shift counts—not acquisition

There is almost no in-migration into language groups from the English language group. We are not here referring to the number of people from English language backgrounds who learn a minority language. Rather, when we speak of linguistic migration into a language groups, we require that a person of English language origin adopt the minority language as his principle language of use. This is a rather stringent test…. What is important to understand, however, is that in terms of this definition, there is virtually no linguistic in-migration into minority language groups. A high degree of bilingualism in a minority language does not constitute linguistic in-migration [italics added]. (Veltman, Language Shift, 1983, p. 12-13. (See Wiley, 2005 for discussion).

What are the implications of the ideological primacy placed on English for the retention of heritage and community languages in the United States, even for major languages such as Spanish?

To what extent are schools keeping up?

To better inform policy and practice ACTFL notes that we need to know:

• How many K-12 students in the nation are studying foreign [and heritage languages].

• We need to have such data at the state, county, and district and school levels.

• We need to know how many students are studying each language.

• We need such data for each grade level at the school level.

• We need to know how long students have been studying each language.

(ACTFL, Are Students Prepared for a Global Society, 2010)

What we do know from the U.S. Census:

Back to the data problem: How do we assess literacy and biliteracy in the national population?

Historically, there are three major approaches to measuring literacy:

1. Self-reported measures

2. Surrogate measures

3. Direct measures

Self-Reported Measures

• In 1850 the U.S. Census, and subsequent censuses merely asked a simple yes/no question

• Gradually the emphasis shifted to a focus on how well one could read and write

• By the 1930s– the rate for whites was 97%– for immigrants, it was 90%

(Wiley, 2005)

Self-Reported Measures, Continued

• Concerns about inflating abilities:– A tendency for some to inflate– Skills achieved may not be retained (Wiley, 2005)

• Concerns about under-valuating one’s abilities – Those with lesser education may confuse

education with general knowledge/ability (Wiley, 1988, 2005)

There has been some data that suggests self-assessments can be useful (McAuthur, 1993.

Self-Reported Literacy in the Census

• In 1850 the U.S. Census, and subsequent censuses merely asked a simple yes/no question

• Gradually the emphasis shifted to a focus on how well one could read and write

• By the 1930s– the rate for whites was 97%– for immigrants, it was 90%

(see Wiley, 2005)

Again, why is there a perpetual literacy gap?

Expectations for Literacy

Perceived Literacy Gap

Actual Levels of Literacy over Time

Surrogate Measures

• Surrogate measures substitute years-of-schooling for evidence of literacy– 4 years was accepted by the U.S. army in 1940– 5 years in 1947– 6 years in 1952– Some now argue for 12 years

• Surrogate measure don’t offer clear evidence of mastery, but can be useful in the analysis of large data sets, such as the U.S. Census

(Wiley, 2005)

Direct Measures

Involve tests of simulated functional literacy.Examples include:• Adult Proficiency Level (APL 1971), which was

competency-based• The National Adult Literacy Survey (1992), which

focused on prose, document, and quantitative literacy

Draw Backs of Direct Measures

Tests of simulated functional literacy may lack ecological validity.

Who but the person or group involved can really describe what “effective functioning in one’s own cultural group” [an APL definition] rally means? Whose needs are served by the generalized statistics about the population?

(Hunter & Harman, 1979).

Using the U.S. Census and ACS for Language and Literacy Data

• As noted, there is a tendency to equate literacy with English literacy and, thereby, a failure to acknowledge literacy in languages other than English.

• This omission tends to inflate perceptions of a literacy “crisis.” Millions of people in the United States are literate in languages other than English; they use other languages as resources, but their abilities are often ignored.

• Thus, analyses of U.S. Census data (Macías, 1988, 1993, 1994, 2000; Macías & Spencer, 1984) and other nationwide surveys such as the ACS can provide valuable information that can help correct common misperceptions about literacy and language diversity.

• As Macías (1990) has noted, however, there are three patterns of literacy among U.S. language minority groups:

(de Klerk & Wiley, 2008)

One more problem: What do we mean by “literacy”?(1) Native language literacy, which is literacy in one’s

mother tongue; (2) Second language literacy (e.g. English), which

implies no literacy in one’s native language; (3) Biliteracy, or literacy in two languages (e.g., in

one’s native language and in English). (4) Nonliteracy (i.e., no literacy in any language) is also

a possibility for those who speak a language that has no written form.

The term illiteracy carries the negative connotation that one is not educated.

(Wiley, 2005; Venezky, Wagner, & Ciliberti, 1990, see also de Klerk & Wiley, 2008).

Utilizing What We’ve Got to Inform Us about Literacy and Language Diversity in the Multilingual U.S.

• The U.S. Census v. The American Community Survey(see de Klerk & Wiley, 2008, Table 1)

• What we’ve got versus what we need.• Being more creative with what we’ve got.• Some examples

Total Hispanic-origin (Age 5+) by Schooling

Total Hispanic-origin (Age 5+) by English Proficiency

Total Hispanic-origin (Age 5+) by Education Levels

Selected Hispanic Foreign-born Hispanic Groups by Literacy

Selected Hispanic Foreign-born Hispanic Groups by Numbers Literate

Mexican Puerto Rican Cuban Dominican Central Am South Am0

1000000

2000000

3000000

4000000

5000000

6000000

Non-LitLiterate

Selected Hispanic Foreign-born by Percent Literate

Mexican Puerto Rican Cuban Dominican Central Am South Am0

10

20

30

40

50

60

70

80

90

100

Non-LitLiterate

Next Steps

• Subgroup Analysis of the Hispanic Population focusing on the relationship between English proficiency, English literacy, Spanish literacy, and employment.

• Results to be presented at UCLA in September for the UCLA Civil Rights Project.

In closing, it is useful to consider that:

“The development of a schooled literacy that grew up around the development of Western patterns of schooling gradually privileged one kind of literacy. Literacy was not exclusively tied to just one sense of grammatical correctness, as the term grammar school suggests to us. Rather the original sense of a common literate discourse was based on a notional of social democracy in the making—a community discourse available to all… We need to continually remind ourselves that other voices need to be heard and not disenfranchised by a single view of a correct language of literacy.”

(Cook-Gumperz & Keller, 1993)