Age of acquisition ratings score better on criterion validity than frequency trajectory or ratings “corrected” for frequency

ABSTRACT

From the very first studies on the effect of age of acquisition (AoA) on word processing, researchers have validated their AoA ratings by correlating them with other, more objective indices of the age at which children know words. Still, the ratings have been questioned, and alternative measures have been proposed. Two of these are differences in word frequency between language directed at young children and language directed at older children (frequency trajectory) and AoA ratings corrected for word frequency. Surprisingly, the criterion validity of these alternative measures has never been established, partly because one of the validation criteria (the age at which children are able to name pictures) has been questioned. In the present study, four databases are used that aimed to establish the order of English word acquisition, going from the very first words learned to words taught in secondary education. The criteria for word knowledge included word production, multiple-choice questions about the meaning of the words, and teacher judgments about when words should be taught in the school curriculum. For all databases, the frequency trajectory correlated substantially less with the criterion than AoA ratings. For all but one, the same was true for AoA ratings “corrected” for other variables. On the basis of these findings, researchers should be cautious interpreting null effects with the “improved” variables as evidence against a genuine AoA effect.

KEYWORDS:

• Age of acquisition
• Word recognition
• Word frequency

Notes

¹To these variables, Brysbaert et al. (2016) added a fifth: word prevalence (how many people know the word in a vocabulary test). However, the impact of this variable so far has only been attested in the Dutch language.

²With thanks to Philip S. Dale, who kindly provided the data.

³The results are very similar when AoA ratings are corrected for word frequency only.