Participants included 118 Mandarin-speaking immigrants in the United States (Age of Arrival = 5-27) and a group of 24 English native speakers as the control participants. All participants read and recorded a standard paragraph taken from the Speech Accent Archive website, and their speech production was then rated for the degree of foreign accent by another group of English native speakers. Morphosyntax knowledge was measured by a 112-item grammaticality judgment test, which included 84 sentences from Johnson & Newport (1989) and 28 sentences from White & Genesse (1996). The selected 84 sentences from the Johnson & Newport study assessed ten standard English morphosyntactic structures whereas the other 28 sentences targeted at universal grammar structures.

Current findings failed to support the CPH in the traditional sense with a specific window of time for L2 learning that closes at puberty. The strength of the age effect was found to vary depending on the particular linguistic dimension and structure under study. Specifically, the phonology dimension was found to be more strongly constrained by the age of arrival variable than the morphosyntax dimension. There were also discrepant degrees of the age effect on different structures within each linguistic dimension. Finally, the predictors for the ultimate outcomes differed for the two linguistic dimensions. Age of arrival, language exposure as well as language learning aptitude contributed to the phonology outcome. In contrast, age of arrival was not a significant predictor for the morphosyntax outcome. The ultimate morphosyntactic attainment was predicted by language exposure and language learning aptitude.