Exploring the effects of word features on French immersion children's ability to deconstruct morphologically complex words

Abstract

The present study investigated factors influencing the ability to decompose multimorphemic words in French in non-francophone children educated in French. In particular, we focused on the effects of two word features: English-French cognate status and base frequency. We also examined the effect of child language background (English first language (EL1) versus English second language (ELL)) on performance. In two related studies, children in grades 1 to 3 completed a translation task requiring them to match morphologically complex words in French and English. Target words were manipulated with respect to cognate status and base frequency. Overall, performance was found to improve over time and to be influenced by cognate status and word frequency. Across all grades, EL1 and ELL children were comparable on task performance. Taken together, these results suggest that French immersion students' ability to deconstruct words and extract morphemes in French is influenced by the presence of cognates, as well as base frequency.

Keywords:

• Morphological awareness
• Cognate
• Word frequency

This work was supported by the Social Sciences and Humanities Research Council of Canada [grant number 494811].

This work was supported by the Social Sciences and Humanities Research Council of Canada [grant number 494811].

Notes

¹ As part of the larger project in which the children were involved, they completed the Peabody Picture Vocabulary Test, Fourth Edition, Form A (PPVT-IV A) (Dunn & Dunn, 2007) as a measure of their English vocabulary. Independent sample t-tests (equal variances not assumed) revealed a significant difference between the EL1 and the ELL group, t(71.35) = 2.11, p = .038.

² The mean base-word frequency is the overall word frequency (i.e., F) of the base word for grades 1 to 5 computed from 1.9 million words taken from 54 French elementary school readers (Lété, Sprenger-Charolles, & Colé, 2004).

³ As part of the larger project in which the children were involved, they completed the PPVT-IV A (Dunn & Dunn, 2007) as a measure of their English vocabulary. Independent sample t-tests revealed significant differences between the EL1 and the ELL group in their vocabulary knowledge at grade 2, t(58) = 2.73, p = .008; at grade 3, the group difference approached significance, t(58) = 1.89, p = .064.

⁴ Levene's test of equality of error variances revealed that for the high-frequency cognate items in grade 3, the EL1 and ELL groups differed significantly in their error variances. Mann-Whitney U test was therefore used in the post-hoc group comparison for these items. Results were consistent with the parametric test in indicating that there was no significant group difference.