This article presents original empirical research carried out in the early total Mandarin language immersion context. The study involves K–5 learners from three early total Mandarin immersion programs whose home language is English. We examined students’ second language (L2) oral proficiency in Mandarin in two ways: (a) a statistical comparative analysis of cross-sectional assessment data for kindergarten, Grade 2, and Grade 5 students and (b) a detailed linguistic complexity analysis comparing immersion students’ speech samples (one per grade level) produced during the assessment interview. Results indicate significant differences in median scores between kindergarten and Grade 2 in all domains; however, no median score differences were found between Grades 2 and 5. An exploratory complexity analysis of three speech samples revealed increasingly higher levels of grammatical complexity across grades. Measures of lexical complexity for the Grade 5 sample, while higher than those in kindergarten, were lower than those of Grade 2. Study findings question the efficacy of existing proficiency assessments at capturing the multidimensionality of oral proficiency in the intermediate and pre-advanced range. They also highlight the important role finely grained complexity measures can play in informing curriculum, instruction, and assessment practices.

中文翻译：

---

评估和探索年轻普通话沉浸式学习者的口语能力

本文介绍了在早期完全普通话沉浸式环境中进行的原始实证研究。该研究涉及来自三个早期全中文沉浸式课程的 K-5 学习者，他们的母语是英语。我们通过两种方式检查学生的第二语言 (L2) 普通话口语能力：(a) 对幼儿园、2 年级和 5 年级学生的横断面评估数据进行统计比较分析，以及 (b) 详细的语言复杂性分析比较评估面试期间制作的沉浸式学生演讲样本（每个年级一个）。结果表明幼儿园和二年级在所有领域的中位数分数存在显着差异；但是，在 2 年级和 5 年级之间没有发现中位数分数差异。对三个语音样本的探索性复杂性分析显示，不同年级的语法复杂性越来越高。5 年级样本的词汇复杂度测量值虽然高于幼儿园，但低于 2 年级。研究结果质疑现有能力评估在捕捉中级和预高级范围内口语能力的多维性方面的有效性。他们还强调了细粒度的复杂性度量在为课程、教学和评估实践提供信息方面可以发挥的重要作用。研究结果质疑现有能力评估在捕捉中级和预高级范围内口语能力的多维性方面的有效性。他们还强调了细粒度的复杂性度量在为课程、教学和评估实践提供信息方面可以发挥的重要作用。研究结果质疑现有能力评估在捕捉中级和预高级范围内口语能力的多维性方面的有效性。他们还强调了细粒度的复杂性度量在为课程、教学和评估实践提供信息方面可以发挥的重要作用。