Does switching languages make bilingual children better at switching tasks?
Bilingual children who were faster at switching between Turkish and Dutch were also faster at switching between colour and shape rules, yet they were no better at nonverbal switching than monolingual peers.
Source
No Bilingual Benefits Despite Relations Between Language Switching and Task Switching
Study at a glance
- Design
- Cross-sectional — Bilingual vs monolingual children matched on age and nonverbal IQ, compared on a cued colour/shape switching task (MANOVA/MANCOVA with SES and Dutch vocabulary as covariates); bilinguals also did a cued picture-naming language-switching task for correlational analysis
- N
- N=54 · 54 children: 27 Turkish-Dutch bilinguals and 27 Dutch monolinguals; within-group correlations use the 27 bilinguals
- Population
- Children aged 5-8 (mean 7.5) attending Dutch-only elementary schools in the Netherlands
- Outcome
- Task-switching accuracy, response time, switching cost and mixing cost; language-switching naming times and costs
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The study compared 27 Turkish-Dutch bilingual children and 27 Dutch monolingual children, aged 5 to 8 and matched on age and nonverbal IQ, on a computer task in which a cartoon cue told them to sort shapes by colour or by shape, mixing both rules in one block. The bilingual children also named pictures in Dutch or Turkish depending on which of two cartoon 'speakers' appeared, switching languages unpredictably. The researchers compared groups with and without controlling for family socioeconomic status (SES) and Dutch vocabulary, and correlated the two switching tasks within the bilingual group.
What they found
Before controls, bilinguals were if anything worse: they tended toward lower accuracy and larger switching costs and had significantly slower responses and larger mixing costs. Once SES and Dutch vocabulary were controlled, the groups did not differ on any measure. Within the bilingual group, response times in the language-switching and nonverbal switching blocks were correlated for both Dutch and Turkish trials, and these correlations held after controlling for age, nonverbal IQ, SES and vocabulary, but switching and mixing costs on the two tasks were unrelated.
The limits
What it doesn't show
The groups could not be matched on SES or Dutch vocabulary, so these differences had to be handled statistically, and other unmeasured factors such as music lessons or gaming could differ between groups. With 27 children per group, a small advantage could be missed; the authors argue power was adequate based on an earlier study, but their task had fewer trials than that study. The within-bilingual link is a correlation in a small sample and rests on overall response times rather than on switching costs, so it shows shared speed of control but not that language practice trains general control.
Key terms
- Switching cost
- The extra response time on trials where the rule changes compared with trials where it repeats, thought to reflect reactive control.
- Mixing cost
- The extra time on repeat trials in a mixed block compared with single-rule blocks, thought to reflect the proactive effort of keeping two rules ready.
- Bilingual advantage hypothesis
- The claim that managing two languages trains domain-general executive control, so bilinguals outperform monolinguals on control tasks.
- Domain-general cognitive control
- Control processes used across many kinds of tasks, not just language, such as switching rules or suppressing distraction.
- Covariate (MANCOVA)
- A variable, here SES and vocabulary, statistically adjusted for so that group comparisons are not driven by it.
Flashcards
0 of 11 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
After controlling for SES and Dutch vocabulary, how did bilinguals compare with monolinguals on task switching?
Common questions
If language switching and task switching are linked, why no bilingual advantage?
A shared control system can exist without language use training it enough to beat monolinguals. Other factors, such as SES and vocabulary, seem to influence control performance more than language-switching practice does.
Why did bilinguals look worse before the controls?
The bilingual families had lower average SES and the children had smaller Dutch vocabularies; once these were accounted for, the gap disappeared, showing how confounds can create apparent monolingual or bilingual advantages.
Why test children rather than adults?
Monolingual children have had little exposure to a second language yet, and children's control skills are still developing, so there is more room to detect differences.
More on Bilingualism