Starman or Sterrenman? Methodological challenges in measuring attitudes towards English in Flemish elementary school children

This study aspires to complement research in three disciplines: (1) developmental sociolinguistics, aiming to chart the way in which the social meaning of language variants and varieties is acquired (see Smith et al. 2013, Nardy et al. 2013); (2) contact linguistics, aiming to discover how young children evaluate the presence or absence of English loanwords in Belgian Dutch (see Zenner et al. 2015, and see Matras 2009 on the importance of prestige for lexical borrowing); (3) psycholinguistics, aiming to overcome methodological challenges when conducting attitudinal research in children (see De Vogelaer & Toye 2017).
Specifically, we report on a within-subject experiment conducted with 204 Belgian Dutch-speaking elementary school children from three age groups (7-, 9- and 11-year olds, balanced for gender). In the experiment, the children are presented with a new cartoon in two guises: (1) a Dutch-only guise; (2) a guise with English alternatives for 15 Dutch nouns in the script, pretested for interchangeability (see Fragment 1 for an extract).
Fragment 1:
Hallo [guys/jongens]! Ik ben [Star Man/Sterrenman]. Samen met mijn [sidekick/hulpje] [Planet Kid/Planeetkind] strijd ik tegen [Dark Moon/Duistermaan] en z’n [gang/bende]. Verdorie, daar is ie! [Brains/Brein], onze [inventor/uitvinder], steekt onze [backpack/rugzak] nog snel vol met [gadgets/snufjes] zodat we aan de [battle/strijd] kunnen beginnen. Klaar? Daar gaan we!
Following the tradition of matched guise experiments (Lambert 1968), the children are asked to evaluate the two superheroes after listening to both fragments (balanced across respondents for order of presentation). Specifically, the children rate the two heroes on child-proof versions of the traditional language attitude scales solidarity and prestige, and for the newer scale ‘dynamism’ (see Grondelaers & Van Hout 2013). Additionally, children are asked to motivate their general preference for either guise in an open answer field. Next, children are presented with a picture naming task to verify their receptive vocabulary knowledge of the words presented in the script (compare De Wilde et al. 2017). Finally, the parents of the children were offered a questionnaire on the children’s exposure to English in the family home (see also Zenner & Van de Mieroop forthcoming).
Overall, the preference for the English hero increases as children grow older. Additionally, in line with the findings of De Vogelaer & Toye (2017), factor analyses and IRT-analyses reveal how the dimensions shaping children’s sensitivity to the prestige of English become more structured and nuanced as children grow older. Finally, our results show that the degree of language awareness, mediated through the receptive knowledge of English, also strongly influences children’s preference for the English or Dutch guise.