Social interaction, particularly in older adolescents, increasingly involves computer‐mediated communication. Although studies of public computer‐mediated communication are increasingly common, studies of private text messaging remain rare. As approaches for obtaining such data evolve with technological advances, developmental scientists need designs in which to use such approaches that reduce sampling biases in both participants and text messages. In this study (n = 854; 46% male; 22% African American, 60% European American), we examined selection biases in the participant sample (i.e., factors associated with actual participation), procedural biases in the participant sample (i.e., factors related to failed data capture due to technological or procedural issues), and selection biases in the sample of text messages (i.e., based on self‐reported reasons for texting). Findings from our study suggest that studying human interaction directly through analysis of text message data is not only feasible, but also may be successfully undertaken with minimal biases regarding sample selection and text message selection among those who are engaged in research and engaged in text messaging outside of the study context. However, biases may occur depending on the type of platform (iPhone vs. Android) used by participants for texting.

中文翻译：

---

收集大学生的短信：评估一种新颖的方法

社会互动，尤其是在年长的青少年中，越来越多地涉及计算机介导的交流。尽管对公共计算机介导的通信的研究越来越普遍，但对私人文本消息的研究仍然很少。随着获取此类数据的方法随着技术的进步而发展，开发科学家需要设计使用这样的方法，以减少参与者和文本消息中的采样偏差。在这项研究中（ñ = 854; 男性46％；22％的非裔美国人，60％的欧洲裔美国人），我们检查了参与者样本中的选择偏见（即与实际参与相关的因素），参与者样本中的过程偏见（即与由于技术或程序问题导致数据获取失败有关的因素） ），以及短信样本中的选择偏向（即，基于自我报告的短信原因）。我们的研究结果表明，直接通过分析文本消息数据来研究人际互动不仅可行，而且可以成功地进行，而在从事研究和从事外部文本消息收发的人员中，在样本选择和文本消息选择方面的偏差最小。研究背景。但是，根据平台的类型（iPhone与