Many online learning contexts are characterized by both high levels of student diversity on socio-demographic attributes (e.g., gender, first language) as well as task-related attributes (e.g., prior online-learning experiences, prior degrees). This longitudinal study investigated the relationships of multi-attributional diversity with CSCL processes and outcomes in a cohort of 1525 distance education freshmen randomly allocated to 343 groups over the course of a nine-week CSCL assignment. Group-level path analyses revealed that, if not explicitly managed, higher multi-attributional socio-demographic diversity was negatively related to the groups' structural integration (computed from digital data). Lower structural integration, in turn, was positively related to lower task-related collaboration time among students and, ultimately, a poorer grading of the groups' work by independent tutors. Moderation analyses further indicated that high task-related diversity operated as an amplifier of the negative relationship of high socio-demographic diversity with structural integration pointing to a risk constellation that requires active intervention.

许多在线学习环境的特点是学生在社会人口属性（例如，性别、第一语言）以及与任务相关的属性（例如，之前的在线学习经验、之前的学位）方面的高度多样性。这项纵向研究调查了在为期 9 周的 CSCL 分配过程中，随机分配到 343 个组的 1525 名远程教育新生的队列中，多归因多样性与 CSCL 过程和结果的关系。群体层面的路径分析表明，如果没有明确管理，更高的多属性社会人口多样性与群体的结构整合（从数字数据计算）呈负相关。反过来，较低的结构整合与学生之间较低的任务相关协作时间呈正相关，最终，独立导师对小组作业的评分较差。适度分析进一步表明，与任务相关的高度多样性作为高度社会人口多样性与结构整合的负面关系的放大器，指向需要积极干预的风险星座。