Purpose Childhood stuttering is common but is often outgrown. Children whose stuttering persists experience significant life impacts, calling for a better understanding of what factors may underlie eventual recovery. In previous research, language ability has been shown to differentiate children who stutter (CWS) from children who do not stutter, yet there is an active debate in the field regarding what, if any, language measures may mark eventual recovery versus persistence. In this study, we examined whether growth in productive language performance may better predict the probability of recovery compared to static profiles taken from a single time point. Method Productive syntax and vocabulary diversity growth rates were calculated for 50 CWS using random coefficient models. Logistic regression models were then used to determine whether growth rates uniquely predict likelihood of recovery, as well as if these rates were predictive over and above currently identified correlates of stuttering onset and recovery. Results Different linguistic profiles emerged between children who went on to recover versus those who persisted. Children who had steeper productive syntactic growth, but not vocabulary diversity growth, were more likely to recover by study end. Moreover, this effect held after controlling for initial language ability at study onset as well as demographic covariates. Conclusions Results are discussed in terms of how growth estimates can be incorporated in recommendations for fostering productive language skills among CWS. The need for additional research on language in early stuttering and recovery is suggested.

中文翻译：

---

生产性语言的增长区分儿童口吃持久性和恢复的初步证据。

目的 儿童口吃很常见，但通常已经过时了。口吃持续存在的儿童会经历重大的生活影响，需要更好地了解哪些因素可能导致最终康复。在先前的研究中，语言能力已被证明可以区分口吃儿童 (CWS) 和不口吃儿童，但该领域存在着激烈的争论，即语言测量可能标志着最终恢复还是坚持。在这项研究中，我们检查了与从单个时间点获取的静态配置文件相比，生产性语言性能的增长是否可以更好地预测恢复的概率。方法 使用随机系数模型计算 50 个 CWS 的有效句法和词汇多样性增长率。然后使用逻辑回归模型来确定增长率是否唯一地预测恢复的可能性，以及这些增长率是否可以预测超过目前确定的口吃发作和恢复的相关性。结果 继续康复的儿童与坚持康复的儿童之间出现了不同的语言特征。生产性句法增长较快但词汇多样性增长较快的儿童更有可能在学习结束时恢复。此外，在控制了研究开始时的初始语言能力以及人口统计协变量之后，这种效应仍然存在。结论 讨论了如何将增长估算纳入培养 CWS 生产性语言技能的建议中的结果。