The Impact of Bilingual Environments on Language Development in Children with Autism Spectrum Disorders

Abstract

The impact of bilingual exposure on language learning has not been systematically studied in children with Autism Spectrum Disorders. This study compared the social abilities and language levels of children (mean age = 56 months) with ASDs from bilingual (n = 45) and monolingual (n = 30) environments. Bilingually-exposed children were subgrouped based on simultaneous bilingual exposure from infancy (SIM, n = 24) versus sequential post-infancy bilingual exposure (SEQ, n = 21). Despite significantly different amounts of bilingual exposure across all groups (p = <0.001) and significantly stronger social interaction scores in the SIM group compared to the SEQ group on the Vineland Adaptive Behavior Scales-II Interpersonal subdomain (p = 0.025), there were no significant group differences in language level. Bilingually-exposed children with ASDs did not experience additional delays in language development.

Appendices

Appendix 1: Language Environment Interview: Script and Scoring for Language Exposure Questions

The following information was collected for every adult caregiver who regularly provided more than 5 h per week of childcare in the home: (a) place of birth, (b) language(s) spoken at home, (c) fluency in each language spoken at home, (d) languages spoken to the child with an ASD, and (e) reasons for any changes in language(s) spoken to the child. For the fluency assessment (question c), caregivers were categorized as ‘native-level’ speakers only if they reported mastery of the language such that they make no pronunciation, grammar, or vocabulary errors and would be recognized as a native speaker of the language by other native speakers. An estimate of language exposure (e.g., % French vs. English input) was recorded for each 6 month period. Finally, the amount of time in childcare was recorded for each 6 month period: researchers coded parent childcare hours using the following 3 levels based on parental employment outside the home: parent not employed or ‘homemaker’, parent employed 5–15 h per week, or parent employed >15 h per week. For non-parent caregivers, hours spent providing childcare were categorized as 5–15 h per week versus >15 h per week.

The following information was collected for childcare in a group setting or intensive therapy program where multiple individuals provided care for the child: (a) languages spoken to the general group of children, (b) languages spoken directly to the child with an ASD, (c) caregivers’ level of fluency in each language(s) used in the environment (e.g., “Were all the daycare workers fluent in French?”), and (d) reasons for any changes in language(s) spoken to the child. For each 6 month period of care, language exposure estimates were elicited (e.g., % French vs. English input) as well as the amount of time the child spent in that setting (e.g., 5–15 h per week vs. >15 h per week).

Following the interview, data from each caregiver were transferred onto a summary score chart for each 6-month period. An algorithm weighted the exposure estimate data based on how much time each caregiver likely spent with the child in direct one-to-one interaction during each 6 month period. For example, exposure from parents with primary childcare responsibilities (e.g., ‘homemakers’) received the highest weight (9 points) whereas exposure through part-time care in a group daycare setting received the lowest weight (1 point). Points assigned to caregivers who provided bilingual input were divided across the languages based on the estimate of usage provided. For example, from age 36 to 41 months, child A was cared for by:

• 1. his mother, who worked 10 h per week: 100% French input: 6 French points

• 2. his father, who worked 40 h per week: 100% English input: 4 English points

• 3. part-time (15 h per week) daycare: 75% French, 25% English input, 1.5 French points, 0.5 English points

Points were added for each language (French: 6 + 1.5 = 7.5 points; English: 4 + 0.5 = 4.5 points) and then divided by the total childcare points (12) for that 6 month period to generate a ratio of 62.5% French (7.5/12) and 37.5% English (4.5/12) exposure. All 6 month exposure estimates were subsequently averaged into yearly subtotaled ratios and a summary Lifetime Ratio (LR). The language that received the highest percentage of lifetime exposure was designated the child’s dominant exposure language.

Appendix 2: Questions from the ADI-R: Scoring and Modifications

The ADI-R is a semi-structured interview (Le Couteur et al. 2003). Trained interviewers ask a series of questions to elicit caregiver descriptions of target behaviors; the interviewer categorizes each behavior using the coding schema for each specific question. For complete question and coding information, please consult the ADI-R manual. The following overview summarizes the lead questions and the coding synopses including codes that were merged due to low cell numbers to facilitate statistical analyses.

Overall level of language (ADI-R #30):

Lead interview question: “How much speech does the subject have now?”

Codes of ‘0’ (functional use of phrases) were contrasted against merged codes of ‘1’ (no phrase use but speech used on a daily basis) and ‘2’ (fewer than 5 words total).

Comprehension of simple language (ADI-R #29):

Lead interview question: “How much language do you think s/he understands if you don’t gesture?”

Codes of ‘0’ (phrase-level comprehension; able to perform novel requests) and ‘1’ (phrase-level comprehension; able to perform familiar requests) were merged, as were scores of ‘2’ (word-level comprehension; >50 words understood) and ‘3’ (word-level comprehension; <50 words); three scores of ‘4’ (‘little or no word comprehension’) were excluded from the analyses.

Questions pertaining to pointing (ADI-R #42) were divided for the interview into separate items for pointing initiation and response:

Initiation of pointing to express interest (variable ‘42a’):

Lead interview question: Does your child ever spontaneously point at things around him/her?

Codes of ‘0’ (spontaneous pointing to express interest in distant objects) were contrasted against merged codes of ‘1’ (limited use of pointing) and ‘2’ (little to no pointing).

Response to pointing to express interest (variable ‘42b’)

Lead interview question: When you point at an item that interests you, does your child turn to look at what you are pointing at?

Codes of ‘0’ (spontaneous response to pointing) were contrasted against merged codes of ‘1’ (responds to points when cued) and ‘2’ (little or no response).

Attention to voice (ADI-R #46)

Lead interview questions: If you come into a room and start talking to your child without calling her/his name, what does s/he do?

Codes were analyzed in this study as follows: ‘0’ (usually looks up when spoken to), ‘1’ (inconsistent or brief response when spoken to; may respond to a firm voice), and merged codes ‘2’ (usually does not look up without strong cues) and ‘3’ (rarely responds).

Age of first single words used meaningfully (ADI-R #9)

Lead question: How old was s/he when he first used words meaningfully, apart from ‘mama’ and ‘dada’?

Single word use was specified as repeated and consistent communicative use of a word other than terms for parents. Age in months when the milestone was reached was elicited.

Age of first phrases used meaningfully (ADI-R #10)

Lead question: How old was s/he when s/he first said something that involved putting words together meaningfully (i.e., using two word phrases)?

Phrase use was specified as a two word phrase including a verb; echoed phrases are excluded. Age in months when the milestone was reached was elicited.