The purpose of this study was to evaluate expressive language sampling (ELS) as a procedure for generating spoken language outcome measures for treatment research in Down syndrome (DS). We addressed (a) feasibility, (b) practice effects across two short-term administrations, (c) test-retest reliability across two short-term administrations, (d) convergent and discriminant construct validity, and (e) considered comparisons across the conversation and narration contexts. Participants were 107 individuals with DS between 6 and 23 years of age who presented with intellectual disability (IQ < 70). The utility of ELS procedures designed to collect samples of spoken language in conversation and narration were evaluated separately. Variables of talkativeness, vocabulary, syntax, utterance planning, and articulation quality, derived from transcripts segmented into C-units (i.e., an independent clause and its modifiers), were considered. A 4-week interval was used to assess practice effects and test-retest reliability. Standardized direct assessments and informant report measures were collected to evaluate construct validity of the ELS variables. Low rates of noncompliance were observed; youth who were under 12 years of age, had phrase-level speech or less, and had a 4-year-old developmental level or less were at particular risk for experiencing difficulty completing the ELS procedures. Minimal practice effects and strong test-retest reliability across the 4-week test-retest interval was observed. The vocabulary, syntax, and speech intelligibility variables demonstrated strong convergent and discriminant validity. Although significant correlations were found between the variables derived from both the conversation and narration contexts, some differences were noted. The ELS procedures considered were feasible and yielded variables with adequate psychometric properties for most individuals with DS between 6 and 23 years old. That said, studies of outcome measures appropriate for individuals with DS with more limited spoken language skills are needed. Context differences were observed in ELS variables suggest that comprehensive evaluation of expressive language is likely best obtained when utilizing both contexts.

中文翻译：

---

唐氏综合症治疗研究的口语结果测量：可行性、实践效果、重测信度和表达性语言抽样产生的变量的结构效度

本研究的目的是评估表达性语言抽样 (ELS) 作为一种为唐氏综合征 (DS) 治疗研究生成口语结果测量的程序。我们讨论了（a）可行性，（b）两次短期管理的实践效果，（c）两次短期管理的重测信度，（d）收敛和判别结构有效性，以及（e）考虑了跨对话和叙述语境。参与者是 107 名 6 至 23 岁的患有智力障碍（智商 < 70）的 DS 个体。分别评估了旨在收集对话和叙述中口语样本的 ELS 程序的效用。健谈性、词汇、句法、话语规划和发音质量的变量，衍生自被分割成 C 单元的转录本（即，一个独立的子句及其修饰语），被考虑在内。使用 4 周的时间间隔来评估练习效果和重测信度。收集标准化的直接评估和知情人报告措施以评估 ELS 变量的结构有效性。观察到不合规率低；年龄在 12 岁以下、具有短语水平或更低水平的言语、以及 4 岁或更低发育水平的青少年在完成 ELS 程序时遇到困难的风险特别高。观察到在 4 周的重测间隔内练习效果最小，重测信度高。词汇、句法和语音可懂度变量表现出很强的收敛效度和判别效度。尽管在源自对话和叙述语境的变量之间发现了显着的相关性，但也注意到了一些差异。所考虑的 ELS 程序是可行的，并且为大多数 6 至 23 岁的 DS 个体产生了具有足够心理测量特性的变量。也就是说，需要研究适合口语技能有限的 DS 患者的结果测量方法。在 ELS 变量中观察到的语境差异表明，在使用两种语境时可能最好地获得对表达性语言的综合评估。所考虑的 ELS 程序是可行的，并且为大多数 6 至 23 岁的 DS 个体产生了具有足够心理测量特性的变量。也就是说，需要研究适合口语技能有限的 DS 患者的结果测量方法。在 ELS 变量中观察到的语境差异表明，在使用两种语境时可能最好地获得对表达性语言的综合评估。所考虑的 ELS 程序是可行的，并且为大多数 6 至 23 岁的 DS 个体产生了具有足够心理测量特性的变量。也就是说，需要研究适合口语技能有限的 DS 患者的结果测量方法。在 ELS 变量中观察到的语境差异表明，在使用两种语境时可能最好地获得对表达性语言的综合评估。