By Angéline Abadie, Amélie Ravier and Lena Sannier
Naomi P. Friedman published in 2016 an article called, “Research on individual differences in executive functions: implications for the bilingual advantage hypothesis”, which focuses on the executive functions through the bilingual advantage hypothesis.
Executive functions (EF) are defined by Friedman as general purpose control mechanisms that allow people to regulate their thoughts and behaviours. The bilingual advantage hypothesis states that, because bilingual people tend to use more executive functions-like processes when speaking (such as switching from one language to another one depending on the context), it might train them to better perform in non-linguistic tasks that require executive functions processes.
According to the author, there is a problem when testing executive functions processes, because these are high-level control processes which have an influence on lower-level ones. To correctly measure it, these low-level processes must be included in the task. But because of this, the task no longer only tests one executive functions process, and there is also some non-executive functions variance.
Latent variables can be used to reduce this task’s impurity when measuring executive functions processes. Latent variables are hypothetical and/or unobservable variables, which can not be measured directly but can be inferred through the performances at the task they influence.
Another way to measure the executive functions properly is to conduct several tests measuring the same ones, and then assess their consistency. Indeed, some results prove that several functions are involved, although they were obtained on a task that is supposed to measure one particular executive function. Thus, if bilinguals show an advantage on one test (Friedman gives the example of interference control) but not on another one, this may mean that the advantage is related to an effect in the “non-executive functions variance”. Indeed, tasks described as involving one particular executive function might actually involve more general processes.
However, by increasing the number of measures, we increase the number of statistical tests and consequently the possibility of a Type I error, which occurs when a null hypothesis, which is in fact true, is rejected by mistake.
One of the positive aspects is that she lengthily writes about the importance of confounding variables and of tempering the results’ interpretation. Friedman also writes about the sample size and the fact that it should be large enough to detect effect sizes, and provides a base-line number of 800 participants.
A key issue in this article is that Friedman does not define the word “bilingualism”. While she does precise that other research papers have discussed the difficulty of choosing an appropriate group of bilingual people, control group and methods to measure bilingualism, she does not explain what she means by “bilingualism”. Another problem is that she seems to call into question the results of other studies, without a sound argumentation and by suggesting statistical errors.
To conclude, the article offers a series of recommendations for people who are experts in the subject and who wish to carry out an experiment, but not for neophytes or students. Thus, it is hard for us to write a constructive analysis of this article since we do not have the theoretical background to do so.
Key-words:
- Notwithstanding: nonobstant
2. Proficient: compétent
3. Bilingual advantage hypothesis: hypothèse de l’avantage du bilinguisme
4. Task impurity: impureté de la tâche
5. Latent variables: variables latentes
Bibliography:
Friedman N. P. (2016). Research on Individual Differences in Executive Functions: Implications for the Bilingual Advantage Hypothesis. Linguistic approaches to bilingualism, 6(5), 535–548. https://doi.org/10.1075/lab.15041.fri