Language Assessment of Bilingual Children

Information about bilinguals in U.S.

• Bilinguals not “two monolinguals in one” (Grosjean, 1989)

• Bilinguals use amalgamated system• Forward & back transfer of strategies (Hernandez, Bates & Avila, 1994)

• Bilingualism changes over lifespan (Kohnert, Bates, & Hernandez, 1999)

Primary Language Impairment

• Primary language impairment (PLI) is a language disorder that delays the mastery of language skills in children who have no hearing loss or other developmental delays.

Bilingual Language Impairment

• Primary language impairment (PLI) is a language disorder that delays the mastery of language skills in children who have no hearing loss or other developmental delays

• Not an effect of second language acquisition• Affects BOTH languages

Challenges in Assessment

True Language Impairment

vs.

Outcomes of Divided Input

Challenges in Assessment

• Second language learning looks like PLI• First language loss looks like PLI• Language Dominance

• Age of first English exposure• Percent of current exposure

• Test domain (by language experience)

Bilingualism

• HIGHLY VARIABLE at individual level

Translation

• Linguistic Equivalence• Functional Equivalence• Cultural Equivalence• Psychometric Equivalence

From: Peña, 2007

Linguistic Equivalence

• Translation focusing on using the same words in similar ways• Ensures similarity in task presentation

• Problem• Same words ≠ same FUNCTION & MEANING

Same words ≠ same FUNCTION & MEANING

What’s the goal of:

• Where’s the ______.

• Show me the ______.

Will translation meet the goal?

Functional Equivalence

• “Where’s the” 6.97 times/million

• “En donde esta” 1 time/million

• “Show” 609 times/million (“show me” 8.02)

• “Enseña” 1 time/million

• “Mira” 439 times/million

• “Busca” 45 times/million

Functional Equivalence

• Goal is to be able to test same concept or behavior• Need to be able to elicit target

Meaning

• Cultural meaning• Words are not culture/experience neutral

Bread

Pan

Meaning

• Goal is to elicit behavior that means the same culturally

• Example: Temperament questionnaire

Practical Considerations

• Language of home

• Language of school

• Demands across contexts

Psychometric Equivalence

• Item difficulty• Based on number of participants who got item correct

• Item discrimination• Difference in correct performance between clinical and non-clinical

group

• Goal equivalent tests

Development of a Bilingual Test

Four Domains: • Semantics• Morphosyntax• Pragmatics• Phonology

Development of a Bilingual Test

• Test blueprint• Focus on markers

• Iterative approach Large item set

Smaller item set

Final

Semantics

• Children with PLI:• Weak semantic representations & semantic depth

(McGregor et al., 2002; Sheng, Peña, Bedore, & Fiestas, 2012)• Require more exposures to learn new word

(Gray, 2003, 2005)

• BUT• Vocabulary knowledge often WNL (low normal)

Semantics

Similarities and differences

Linguistic concepts

Characteristics of properties

Categorization

Word associations

Function of objects

Morphosyntax

Spanish• Direct object clitics• Articles• Subjunctive

(Bedore & Leonard, 2002; Gutiérrez-Clellen, Restrepo, & Simon-Cereijido, 2006; Jacobson & Schwartz, 2005)

English• Tense marking (-ing, -ed)• Passives• Plurals

(Conti-Ramsden, 2003; Dale & Cole, 1991; Leonard, 1998; Rice, 1996)

• Difficulty with inflectional morphology (Leonard, 2000)

• Difficulties differ depending on the language

Morphosyntax

Articles

Prepositions

Periphrastic future

Present Progressive

Preterite

Subjunctive verbs

Clitics

Other-- SR

Spanish

Possessives

Third Person Singular

Regular and Irregular Past

Plural NounsPresent and Past Aux +

ingCopula

Prepositions

Negatives

Passives

Other-- SR

English

What language should be tested?

• L1• L2• Both

• How should the two languages be combined?

Language Performance

• Associated with• age of first L2 exposure• current input/output

(Bedore, et al., 2012)

Language Dominance

Patterns

Bilingual English Spanish Assessment (BESA)• Best (higher) in each language for:

morphosyntax + semantics

Morpho Eng X XMorpho Spn X X

Sem Eng X X

Sem Spn X X

Procedure

• Children tested in Spanish, English or both• Ability testing: language samples, parent/teacher questionnaire, clinical concern• Bilingual input-output based on parent/teacher questionnaire

• Bilingual English Spanish Assessment– BESA Semantics

Discriminant Analysis

• Determined cuts via FME and FMS• Applied same cut-scores to BSE group

• Single language• Both languages

Age Sensitivity(LI as LI)

Specificity (TD as TD)

+ likelihood ratio

- likelihoodratio

English Semantics 4 81.0% 73.1% 3.01 0.26 5 90.0% 63.3% 2.45 0.16 6 52.6% 85.7% 4.09 0.48 Spanish Semantics 4 70.0% 83.3% 4.19 0.36 5 93.8% 77.8% 4.25 0.08 6 87.5% 86.8% 6.63 0.14

5 10 15 20 25 30 35 40 450

5

10

15

20

25

30

35

40

45

LINL

Age Sensitivity(LI as LI)

Specificity (TD as TD)

+ likelihood ratio

- likelihoodratio

English Semantics 4 81.0% 73.1% 3.01 0.26 5 90.0% 63.3% 2.45 0.16 6 52.6% 85.7% 4.09 0.48 Spanish Semantics 4 70.0% 83.3% 4.19 0.36 5 93.8% 77.8% 4.25 0.08 6 87.5% 86.8% 6.63 0.14

English & Spanish Semantics 4 85.7% 100.0% infinite 0.12 5 90.9% 84.0% 5.68 0.11 6 91.7% 85.7% 6.41 0.10

Discrimination

Bilingual English Spanish Assessment (BESA)

Sensitivity SpecificityBest language morphosyntax + Best language semantics

90.5% 87.5%

Discussion

For ID of PLI in bilinguals consider:• Structure of language

• Markers of PLI particular to language

• Level of exposure– L1 vs. L2• Performance by language and domain• Performance in best language

Practical Considerations

• Language of home• Language of school• Demands across contexts